Showing 51–100 of 223 results for author: Law, M

Multi-outcome trials with a generalised number of efficacious outcomes

Authors: Martin Law, Michael J. Grayling, Adrian P. Mander

Abstract: Existing multi-outcome designs focus almost entirely on evaluating whether all outcomes show evidence of efficacy or whether at least one outcome shows evidence of efficacy. While a small number of authors have provided multi-outcome designs that evaluate when a general number of outcomes show promise, these designs have been single-stage in nature only. We therefore propose two designs, of group-… ▽ More Existing multi-outcome designs focus almost entirely on evaluating whether all outcomes show evidence of efficacy or whether at least one outcome shows evidence of efficacy. While a small number of authors have provided multi-outcome designs that evaluate when a general number of outcomes show promise, these designs have been single-stage in nature only. We therefore propose two designs, of group-sequential and drop the loser form, that provide this design characteristic in a multi-stage setting. Previous such multi-outcome multi-stage designs have allowed only for a maximum of two outcomes; our designs thus also extend previous related proposals by permitting any number of outcomes. △ Less

Submitted 18 December, 2020; originally announced December 2020.

Comments: 38 pages, 11 figures, 4 tables. Submitted for the degree of Doctor of Philosophy at University of Cambridge

NSL: Hybrid Interpretable Learning From Noisy Raw Data

Authors: Daniel Cunnington, Alessandra Russo, Mark Law, Jorge Lobo, Lance Kaplan

Abstract: Inductive Logic Programming (ILP) systems learn generalised, interpretable rules in a data-efficient manner utilising existing background knowledge. However, current ILP systems require training examples to be specified in a structured logical format. Neural networks learn from unstructured data, although their learned models may be difficult to interpret and are vulnerable to data perturbations a… ▽ More Inductive Logic Programming (ILP) systems learn generalised, interpretable rules in a data-efficient manner utilising existing background knowledge. However, current ILP systems require training examples to be specified in a structured logical format. Neural networks learn from unstructured data, although their learned models may be difficult to interpret and are vulnerable to data perturbations at run-time. This paper introduces a hybrid neural-symbolic learning framework, called NSL, that learns interpretable rules from labelled unstructured data. NSL combines pre-trained neural networks for feature extraction with FastLAS, a state-of-the-art ILP system for rule learning under the answer set semantics. Features extracted by the neural components define the structured context of labelled examples and the confidence of the neural predictions determines the level of noise of the examples. Using the scoring function of FastLAS, NSL searches for short, interpretable rules that generalise over such noisy examples. We evaluate our framework on propositional and first-order classification tasks using the MNIST dataset as raw data. Specifically, we demonstrate that NSL is able to learn robust rules from perturbed MNIST data and achieve comparable or superior accuracy when compared to neural network and random forest baselines whilst being more general and interpretable. △ Less

Submitted 25 June, 2021; v1 submitted 9 December, 2020; originally announced December 2020.

Comments: This article has been replaced with arXiv:2106.13103

Revisiting the HD 21749 Planetary System with Stellar Activity Modeling

Authors: Tianjun Gan, Sharon Xuesong Wang, Johanna K. Teske, Shude Mao, Ward S. Howard, Nicholas M. Law, Natasha E. Batalha, Andrew Vanderburg, Diana Dragomir, Chelsea X. Huang, Fabo Feng, R. Paul Butler, Jeffrey D. Crane, Stephen A. Shectman, Yuri Beletsky, Avi Shporer, Benjamin T. Montet, Jennifer A. Burt, Adina D. Feinstein, Erin Flowers, Sangeetha Nandakumar, Mauro Barbieri, Hank Corbett, Jeffrey K. Ratzloff, Nathan Galliher , et al. (4 additional authors not shown)

Abstract: HD 21749 is a bright ($V=8.1$ mag) K dwarf at 16 pc known to host an inner terrestrial planet HD 21749c as well as an outer sub-Neptune HD 21749b, both delivered by TESS. Follow-up spectroscopic observations measured the mass of HD 21749b to be $22.7\pm2.2\ M_{\oplus}$ with a density of $7.0^{+1.6}_{-1.3}$ g~cm$^{-3}$, making it one of the densest sub-Neptunes. However, the mass measurement was su… ▽ More HD 21749 is a bright ($V=8.1$ mag) K dwarf at 16 pc known to host an inner terrestrial planet HD 21749c as well as an outer sub-Neptune HD 21749b, both delivered by TESS. Follow-up spectroscopic observations measured the mass of HD 21749b to be $22.7\pm2.2\ M_{\oplus}$ with a density of $7.0^{+1.6}_{-1.3}$ g~cm$^{-3}$, making it one of the densest sub-Neptunes. However, the mass measurement was suspected to be influenced by stellar rotation. Here we present new high-cadence PFS RV data to disentangle the stellar activity signal from the planetary signal. We find that HD 21749 has a similar rotational timescale as the planet's orbital period, and the amplitude of the planetary orbital RV signal is estimated to be similar to that of the stellar activity signal. We perform Gaussian Process (GP) regression on the photometry and RVs from HARPS and PFS to model the stellar activity signal. Our new models reveal that HD 21749b has a radius of $2.86\pm0.20\ R_{\oplus}$, an orbital period of $35.6133\pm0.0005$ d with a mass of $M_{b}=20.0\pm2.7\ M_{\oplus}$ and a density of $4.8^{+2.0}_{-1.4}$ g~cm$^{-3}$ on an eccentric orbit with $e=0.16\pm0.06$, which is consistent with the most recent values published for this system. HD 21749c has an orbital period of $7.7902\pm0.0006$ d, a radius of $1.13\pm0.10\ R_{\oplus}$, and a 3$σ$ mass upper limit of $3.5\ M_{\oplus}$. Our Monte Carlo simulations confirm that without properly taking stellar activity signals into account, the mass measurement of HD 21749b is likely to arrive at a significantly underestimated error bar. △ Less

Submitted 9 December, 2020; originally announced December 2020.

Comments: 20 pages, 16 figures, accepted to MNRAS

Self-Supervised Real-to-Sim Scene Generation

Authors: Aayush Prakash, Shoubhik Debnath, Jean-Francois Lafleche, Eric Cameracci, Gavriel State, Stan Birchfield, Marc T. Law

Abstract: Synthetic data is emerging as a promising solution to the scalability issue of supervised deep learning, especially when real data are difficult to acquire or hard to annotate. Synthetic data generation, however, can itself be prohibitively expensive when domain experts have to manually and painstakingly oversee the process. Moreover, neural networks trained on synthetic data often do not perform… ▽ More Synthetic data is emerging as a promising solution to the scalability issue of supervised deep learning, especially when real data are difficult to acquire or hard to annotate. Synthetic data generation, however, can itself be prohibitively expensive when domain experts have to manually and painstakingly oversee the process. Moreover, neural networks trained on synthetic data often do not perform well on real data because of the domain gap. To solve these challenges, we propose Sim2SG, a self-supervised automatic scene generation technique for matching the distribution of real data. Importantly, Sim2SG does not require supervision from the real-world dataset, thus making it applicable in situations for which such annotations are difficult to obtain. Sim2SG is designed to bridge both the content and appearance gaps, by matching the content of real data, and by matching the features in the source and target domains. We select scene graph (SG) generation as the downstream task, due to the limited availability of labeled datasets. Experiments demonstrate significant improvements over leading baselines in reducing the domain gap both qualitatively and quantitatively, on several synthetic datasets as well as the real-world KITTI dataset. △ Less

Submitted 18 August, 2021; v1 submitted 29 November, 2020; originally announced November 2020.

Comments: accepted at ICCV 2021. Project page: https://research.nvidia.com/publication/2021-08_Sim2SG

Orbital Foregrounds for Ultra-Short Duration Transients

Authors: Hank Corbett, Nicholas M. Law, Alan Vasquez Soto, Ward S. Howard, Amy Glazier, Ramses Gonzalez, Jeffrey K. Ratzloff, Nathan Galliher, Octavi Fors, Robert Quimby

Abstract: Reflections from objects in Earth orbit can produce sub-second, star-like optical flashes similar to astrophysical transients. Reflections have historically caused false alarms for transient surveys, but the population has not been systematically studied. We report event rates for these orbital flashes using the Evryscope Fast Transient Engine, a low-latency transient detection pipeline for the Ev… ▽ More Reflections from objects in Earth orbit can produce sub-second, star-like optical flashes similar to astrophysical transients. Reflections have historically caused false alarms for transient surveys, but the population has not been systematically studied. We report event rates for these orbital flashes using the Evryscope Fast Transient Engine, a low-latency transient detection pipeline for the Evryscopes. We select single-epoch detections likely caused by Earth satellites and model the event rate as a function of both magnitude and sky position. We measure a rate of $1800^{+600}_{-280}$ sky$^{-1}$ hour$^{-1}$, peaking at $m_g = 13.0$, for flashes morphologically degenerate with real astrophysical signals in surveys like the Evryscopes. Of these, $340^{+150}_{-85}$ sky$^{-1}$ hour$^{-1}$ are bright enough to be visible to the naked eye in typical suburban skies with a visual limiting magnitude of $V\approx4$. These measurements place the event rate of orbital flashes orders of magnitude higher than the combined rate of public alerts from all active all-sky fast-timescale transient searches, including neutrino, gravitational-wave, gamma-ray, and radio observatories. Short-timescale orbital flashes form a dominating foreground for un-triggered searches for fast transients in low-resolution, wide-angle surveys. However, events like fast radio bursts (FRBs) with arcminute-scale localization have a low probability ($\sim10^{-5}$) of coincidence with an orbital flash, allowing optical surveys to place constraints on their potential optical counterparts in single images. Upcoming satellite internet constellations, like SpaceX Starlink, are unlikely to contribute significantly to the population of orbital flashes in normal operations. △ Less

Submitted 4 November, 2020; originally announced November 2020.

Comments: 8 pages, 4 figures

Journal ref: ApJL, 903, L27 (2020)

The Hot Hand and Its Effect on the NBA

Authors: Brian McNair, Eric Margolin, Michael Law, Ya'acov Ritov

Abstract: This paper aims to revisit and expand upon previous work on the "hot hand" phenomenon in basketball, specifically in the NBA. Using larger, modern data sets, we test streakiness of shooting patterns and the presence of hot hand behavior in free throw shooting, while going further by examining league-wide hot hand trends and the changes in individual player behavior. Additionally, we perform simula… ▽ More This paper aims to revisit and expand upon previous work on the "hot hand" phenomenon in basketball, specifically in the NBA. Using larger, modern data sets, we test streakiness of shooting patterns and the presence of hot hand behavior in free throw shooting, while going further by examining league-wide hot hand trends and the changes in individual player behavior. Additionally, we perform simulations in order to assess their power. While we find no evidence of the hot hand in game-play and only weak evidence in free throw trials, we find that some NBA players exhibit behavioral changes based on the outcome of their previous shot. △ Less

Submitted 29 October, 2020; originally announced October 2020.

From core collapse to superluminous: The rates of massive stellar explosions from the Palomar Transient Factory

Authors: C. Frohmaier, C. R. Angus, M. Vincenzi, M. Sullivan, M. Smith, P. E. Nugent, S. B. Cenko, A. Gal-Yam, S. R. Kulkarni, N. M. Law, R. M. Quimby

Abstract: We present measurements of the local core collapse supernova (SN) rate using SN discoveries from the Palomar Transient Factory (PTF). We use a Monte Carlo simulation of hundreds of millions of SN light curve realizations coupled with the detailed PTF survey detection efficiencies to forward-model the SN rates in PTF. Using a sample of 86 core collapse SNe, including 26 stripped-envelope SNe (SESNe… ▽ More We present measurements of the local core collapse supernova (SN) rate using SN discoveries from the Palomar Transient Factory (PTF). We use a Monte Carlo simulation of hundreds of millions of SN light curve realizations coupled with the detailed PTF survey detection efficiencies to forward-model the SN rates in PTF. Using a sample of 86 core collapse SNe, including 26 stripped-envelope SNe (SESNe), we show that the overall core collapse SN volumetric rate is $r^\mathrm{CC}_v=9.10_{-1.27}^{+1.56}\times10^{-5}\,\text{SNe yr}^{-1}\,\text{Mpc}^{-3}\, h_{70}^{3}$ at $ \langle z \rangle = 0.028$, and the SESN volumetric rate is $r^\mathrm{SE}_v=2.41_{-0.64}^{+0.81}\times10^{-5}\, \text{SNe yr}^{-1}\,\text{Mpc}^{-3}\, h_{70}^{3}$. We further measure a volumetric rate for hydrogen-free superluminous SNe (SLSNe-I) using 8 events at $z{\le}0.2$ of $r^\mathrm{SLSN-I}_v=35_{-13}^{+25}\, \text{SNe yr}^{-1}\text{Gpc}^{-3}\, h_{70}^{3}$, which represents the most precise SLSN-I rate measurement to date. Using a simple cosmic star-formation history to adjust these volumetric rate measurements to the same redshift, we measure a local ratio of SLSN-I to SESN of $\sim1/810^{+1500}_{-94}$, and of SLSN-I to all CCSN types of $\sim 1/3500^{+2800}_{-720}$. However, using host galaxy stellar mass as a proxy for metallicity, we also show that this ratio is strongly metallicity dependent: in low-mass ($\mathrm{log} M_{*} < 9.5 \mathrm{M}_\odot$) galaxies, which are the only environments that host SLSN-I in our sample, we measure a SLSN-I to SESN fraction of $1/300^{+380}_{-170}$ and $1/1700^{+1800}_{-720}$ for all CCSN. We further investigate the SN rates a function of host galaxy stellar mass and show that the specific rates of all core collapse SNe decrease with increasing stellar mass. △ Less

Submitted 28 October, 2020; originally announced October 2020.

Comments: 17 pages, 10 figures, submitted to MNRAS

When Do Stalled Stars Resume Spinning Down? Advancing Gyrochronology with Ruprecht 147

Authors: Jason Lee Curtis, Marcel A. Agüeros, Sean P. Matt, Kevin R. Covey, Stephanie T. Douglas, Ruth Angus, Steven H. Saar, Ann Marie Cody, Andrew Vanderburg, Nicholas M. Law, Adam L. Kraus, David W. Latham, Christoph Baranec, Reed Riddle, Carl Ziegler, Mikkel N. Lund, Guillermo Torres, Søren Meibom, Victor Silva Aguirre, Jason T. Wright

Abstract: Recent measurements of rotation periods ($P_\text{rot}$) in the benchmark open clusters Praesepe (670 Myr), NGC 6811 (1 Gyr), and NGC 752 (1.4 Gyr) demonstrate that, after converging onto a tight sequence of slowly rotating stars in mass$-$period space, stars temporarily stop spinning down. These data also show that the duration of this epoch of stalled spin-down increases toward lower masses. To… ▽ More Recent measurements of rotation periods ($P_\text{rot}$) in the benchmark open clusters Praesepe (670 Myr), NGC 6811 (1 Gyr), and NGC 752 (1.4 Gyr) demonstrate that, after converging onto a tight sequence of slowly rotating stars in mass$-$period space, stars temporarily stop spinning down. These data also show that the duration of this epoch of stalled spin-down increases toward lower masses. To determine when stalled stars resume spinning down, we use data from the $K2$ mission and the Palomar Transient Factory to measure $P_\text{rot}$ for 58 dwarf members of the 2.7-Gyr-old cluster Ruprecht 147, 39 of which satisfy our criteria designed to remove short-period or near-equal-mass binaries. Combined with the $Kepler$ $P_\text{rot}$ data for the approximately coeval cluster NGC 6819 (30 stars with $M_\star > 0.85$ M$_\odot$), our new measurements more than double the number of $\approx$2.5 Gyr benchmark rotators and extend this sample down to $\approx$0.55 M$_\odot$. The slowly rotating sequence for this joint sample appears relatively flat (22 $\pm$ 2 days) compared to sequences for younger clusters. This sequence also intersects the $Kepler$ intermediate period gap, demonstrating that this gap was not created by a lull in star formation. We calculate the time at which stars resume spinning down, and find that 0.55 M$_\odot$ stars remain stalled for at least 1.3 Gyr. To accurately age-date low-mass stars in the field, gyrochronology formulae must be modified to account for this stalling timescale. Empirically tuning a core$-$envelope coupling model with open cluster data can account for most of the apparent stalling effect. However, alternative explanations, e.g., a temporary reduction in the magnetic braking torque, cannot yet be ruled out. △ Less

Submitted 5 October, 2020; originally announced October 2020.

Comments: 51 pages and 21 figures. Machine-readable tables for Ruprecht 147 and the Benchmark Clusters catalogs are included in arxiv source

EvryFlare III: Temperature Evolution and Habitability Impacts of Dozens of Superflares Observed Simultaneously by Evryscope and TESS

Authors: Ward S. Howard, Hank Corbett, Nicholas M. Law, Jeffrey K. Ratzloff, Nathan Galliher, Amy L. Glazier, Ramses Gonzalez, Alan Vasquez Soto, Octavi Fors, Daniel del Ser, Joshua Haislip

Abstract: Superflares may provide the dominant source of biologically relevant UV radiation to rocky habitable zone M-dwarf planets (M-Earths), altering planetary atmospheres and conditions for surface life. The combined line and continuum flare emission has usually been approximated by a 9000 K blackbody. If superflares are hotter, then the UV emission may be 10X higher than predicted from the optical. How… ▽ More Superflares may provide the dominant source of biologically relevant UV radiation to rocky habitable zone M-dwarf planets (M-Earths), altering planetary atmospheres and conditions for surface life. The combined line and continuum flare emission has usually been approximated by a 9000 K blackbody. If superflares are hotter, then the UV emission may be 10X higher than predicted from the optical. However, it is unknown for how long M-dwarf superflares reach temperatures above 9000 K. Only a handful of M-dwarf superflares have been recorded with multi-wavelength high-cadence observations. We double the total number of events in the literature using simultaneous Evryscope and TESS observations to provide the first systematic exploration of the temperature evolution of M-dwarf superflares. We also increase the number of superflaring M-dwarfs with published time-resolved blackbody evolution by ~10X. We measure temperatures at 2 min cadence for 42 superflares from 27 K5-M5 dwarfs. We find superflare peak temperatures (defined as the mean of temperatures corresponding to flare FWHM) increase with flare energy and impulse. We find the amount of time flares emit at temperatures above 14,000 K depends on energy. We discover 43% of the flares emit above 14,000 K, 23% emit above 20,000 K and 5% emit above 30,000 K. The largest and hottest flare briefly reached 42,000 K. Some do not reach 14,000 K. During superflares, we estimate M-Earths orbiting <200 Myr stars typically receive a top-of-atmosphere UV-C flux of ~120 W m^-2 and up to 10^3 W m^-2, 100-1000X the time-averaged XUV flux from Proxima Cen. △ Less

Submitted 1 October, 2020; originally announced October 2020.

Comments: 14 pages, 9 figures, 2 tables. Accepted for publication to the Astrophysical Journal. In press

EVR-CB-004: An Inflated Hot Subdwarf O star + Unseen WD Companion in a Compact Binary Discovered with the Evryscope

Authors: Jeffrey K. Ratzloff, Thomas Kupfer, Brad N. Barlow, David Schneider, Thomas R. Marsh, Ulrich Heber, Kyle A. Corcoran, Evan Bauer, Steven Hammerich, Henry T. Corbett, Amy Glazier, Ward S. Howard, Nicholas M. Law

Abstract: We present the discovery of EVR-CB-004, a close binary with a remnant stellar core and an unseen white dwarf companion. The analysis in this work reveals the primary is potentially an inflated hot subdwarf (sdO) and more likely is a rarer post-blue horizontal branch (post-BHB) star. Post-BHBs are the short-lived shell-burning final stage of a blue horizontal star or hot subdwarf before transitioni… ▽ More We present the discovery of EVR-CB-004, a close binary with a remnant stellar core and an unseen white dwarf companion. The analysis in this work reveals the primary is potentially an inflated hot subdwarf (sdO) and more likely is a rarer post-blue horizontal branch (post-BHB) star. Post-BHBs are the short-lived shell-burning final stage of a blue horizontal star or hot subdwarf before transitioning to a WD. This object was discovered using Evryscope photometric data in a southern-all-sky hot subdwarf variability survey. The photometric light curve for EVR-CB-004 shows multi-component variability from ellipsoidal deformation of the primary and from Doppler boosting as well as gravitational limb darkening. EVR-CB-004 is one of just a handful of known systems, and has a long period (6.08426 hours) and large amplitude ellipsoidal modulation (16.0 $\%$ change in brightness from maximum to minimum) for these extremely close binary systems, while the properties of the primary make it a truly unique system. EVR-CB-004 also shows a peculiar low-amplitude (less than $1\%$) sinusoidal light curve variation with a period that is a 1/3 resonance of the binary period. We tentatively identify this additional variation source as a tidally-induced resonant pulsation, and we suggest followup observations that could verify this interpretation. From the evolutionary state of the system, its components, and its mass fraction, EVR-CB-004 is a strong merger candidate to form a single high-mass ($\approx1.2M_{\odot}$) WD. EVR-CB-004 offers a glimpse into a brief phase of a remnant core evolution and secondary variation, not seen before in a compact binary. △ Less

Submitted 13 September, 2020; originally announced September 2020.

Comments: Accepted to ApJ September 7, 2020

Induction and Exploitation of Subgoal Automata for Reinforcement Learning

Authors: Daniel Furelos-Blanco, Mark Law, Anders Jonsson, Krysia Broda, Alessandra Russo

Abstract: In this paper we present ISA, an approach for learning and exploiting subgoals in episodic reinforcement learning (RL) tasks. ISA interleaves reinforcement learning with the induction of a subgoal automaton, an automaton whose edges are labeled by the task's subgoals expressed as propositional logic formulas over a set of high-level events. A subgoal automaton also consists of two special states:… ▽ More In this paper we present ISA, an approach for learning and exploiting subgoals in episodic reinforcement learning (RL) tasks. ISA interleaves reinforcement learning with the induction of a subgoal automaton, an automaton whose edges are labeled by the task's subgoals expressed as propositional logic formulas over a set of high-level events. A subgoal automaton also consists of two special states: a state indicating the successful completion of the task, and a state indicating that the task has finished without succeeding. A state-of-the-art inductive logic programming system is used to learn a subgoal automaton that covers the traces of high-level events observed by the RL agent. When the currently exploited automaton does not correctly recognize a trace, the automaton learner induces a new automaton that covers that trace. The interleaving process guarantees the induction of automata with the minimum number of states, and applies a symmetry breaking mechanism to shrink the search space whilst remaining complete. We evaluate ISA in several gridworld and continuous state space problems using different RL algorithms that leverage the automaton structures. We provide an in-depth empirical analysis of the automaton learning performance in terms of the traces, the symmetry breaking and specific restrictions imposed on the final learnable automaton. For each class of RL problem, we show that the learned automata can be successfully exploited to learn policies that reach the goal, achieving an average reward comparable to the case where automata are not learned but handcrafted and given beforehand. △ Less

Submitted 16 March, 2021; v1 submitted 8 September, 2020; originally announced September 2020.

Comments: Published in the Journal of Artificial Intelligence Research (JAIR)

Journal ref: Journal of Artificial Intelligence Research, 70, 1031-1116 (2021)

Describing transport in defected nanoparticle solids using a new, hierarchical, simulation tool, TRIDENS

Authors: Chase Hansen, Davis Unruh, Miguel Alba, Caroline Qian, Alex Abelson, Matt Law, Gergely T. Zimanyi

Abstract: The efficiency of nanoparticle (NP) solar cells has grown impressively in recent years, exceeding 16%. However, the carrier mobility in NP solar cells, and in other optoelectronic applications remains low, thus critically limiting their performance. Therefore, carrier transport in NP solids needs to be better understood to further improve the overall efficiency of NP solar cell technology. However… ▽ More The efficiency of nanoparticle (NP) solar cells has grown impressively in recent years, exceeding 16%. However, the carrier mobility in NP solar cells, and in other optoelectronic applications remains low, thus critically limiting their performance. Therefore, carrier transport in NP solids needs to be better understood to further improve the overall efficiency of NP solar cell technology. However, it is technically challenging to simulate experimental scale samples, as physical processes from atomic to mesoscopic scales all crucially impact transport. To rise to this challenge, here we report the development of TRIDENS: the Transport in Defected Nanoparticle Solids Simulator, that adds three more hierarchical layers to our previously developed HINTS code for nanoparticle solar cells. In TRIDENS, we first introduced planar defects, such as twin planes and grain boundaries into individual NP SLs that comprised the order of 10^3 NPs. Then we used HINTS to simulate the transport across tens of thousands of defected NP SLs, and constructed the distribution of the NP SL mobilities with planar defects. Second, the defected NP SLs were assembled into a resistor network with more than 10^4 NP SLs, thus representing about 10^7 individual NPs. Finally, the TRIDENS results were analyzed by finite size scaling to explore whether the percolation transition, separating the phase where the low mobility defected NP SLs percolate, from the phase where the high mobility undefected NP SLs percolate drives a low-mobility-to-high-mobility transport crossover that can be extrapolated to macroscopic length scales. For the theoretical description, we adapted the Efros-Shklovskii bimodal mobility distribution percolation model. We demonstrated that the ES bimodal theory's two-variable scaling function is an effective tool to quantitatively characterize this low-mobility-to-high-mobility transport crossover. △ Less

Submitted 9 March, 2021; v1 submitted 2 September, 2020; originally announced September 2020.

Comments: 11 pages, 5 figures

Evryscope-South Survey of Upper- and Pre-main Sequence Solar Neighborhood Stars

Authors: Nathan W. Galliher, Jeffrey K. Ratzloff, Henry Corbett, Nicholas M. Law, Ward S. Howard, Amy L. Glazier, Alan Vasquez Soto, Ramses Gonzalez

Abstract: Using photometric data collected by Evryscope-South, we search for nearby young variable systems on the upper-main sequence (UMS) and pre-main sequence (PMS). The Evryscopes are all-sky high-cadence telescope arrays operating in the Northern and Southern hemispheres. We base our search on a Gaia-selected catalog of young neighborhood upper- and pre-main sequence stars which were chosen through bot… ▽ More Using photometric data collected by Evryscope-South, we search for nearby young variable systems on the upper-main sequence (UMS) and pre-main sequence (PMS). The Evryscopes are all-sky high-cadence telescope arrays operating in the Northern and Southern hemispheres. We base our search on a Gaia-selected catalog of young neighborhood upper- and pre-main sequence stars which were chosen through both astrometric and photometric criteria. We analyze 44,971 Evryscope-South light curves in search of variability. We recover 615 variables, with 378 previously known, and 237 new discoveries including 84 young eclipsing binary (EB) candidates. We discover a new highly eccentric binary system and recover a further four previously known systems, with periods ranging from 299 to 674 hr. We find 158 long-period (>50 hr) candidate EB systems, 9 from the PMS and 149 from the UMS, which will allow constraints on the mass-radius-age relation. These long-period EBs include a 179.3 hr PMS system and a 867.8 hr system from the UMS. For PMS variable candidates we estimate system ages, which range from 1 to 23 Myr for non-EBs and from 2 to 17 Myr for EBs. Other non-EB discoveries that show intrinsic variability will allow relationships between stellar rotation rates, ages, activity, and mass to be characterized. △ Less

Submitted 28 April, 2021; v1 submitted 17 August, 2020; originally announced August 2020.

Comments: Final version published in PASP, typos corrected, 16 pages, 15 figures, 4 tables, 2 supplementary tables

Segmentation of Pulmonary Opacification in Chest CT Scans of COVID-19 Patients

Authors: Keegan Lensink, Issam Laradji, Marco Law, Paolo Emilio Barbano, Savvas Nicolaou, William Parker, Eldad Haber

Abstract: The Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) has rapidly spread into a global pandemic. A form of pneumonia, presenting as opacities with in a patient's lungs, is the most common presentation associated with this virus, and great attention has gone into how these changes relate to patient morbidity and mortality. In this work we provide open source models for the segmentation o… ▽ More The Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) has rapidly spread into a global pandemic. A form of pneumonia, presenting as opacities with in a patient's lungs, is the most common presentation associated with this virus, and great attention has gone into how these changes relate to patient morbidity and mortality. In this work we provide open source models for the segmentation of patterns of pulmonary opacification on chest Computed Tomography (CT) scans which have been correlated with various stages and severities of infection. We have collected 663 chest CT scans of COVID-19 patients from healthcare centers around the world, and created pixel wise segmentation labels for nearly 25,000 slices that segment 6 different patterns of pulmonary opacification. We provide open source implementations and pre-trained weights for multiple segmentation models trained on our dataset. Our best model achieves an opacity Intersection-Over-Union score of 0.76 on our test set, demonstrates successful domain adaptation, and predicts the volume of opacification within 1.7\% of expert radiologists. Additionally, we present an analysis of the inter-observer variability inherent to this task, and propose methods for appropriate probabilistic approaches. △ Less

Submitted 8 July, 2020; v1 submitted 7 July, 2020; originally announced July 2020.

Comments: 9 pages, 5 figures. Fix typo in delimiter between author names in arXiv metadata

A Weakly Supervised Consistency-based Learning Method for COVID-19 Segmentation in CT Images

Authors: Issam Laradji, Pau Rodriguez, Oscar Mañas, Keegan Lensink, Marco Law, Lironne Kurzman, William Parker, David Vazquez, Derek Nowrouzezahrai

Abstract: Coronavirus Disease 2019 (COVID-19) has spread aggressively across the world causing an existential health crisis. Thus, having a system that automatically detects COVID-19 in tomography (CT) images can assist in quantifying the severity of the illness. Unfortunately, labelling chest CT scans requires significant domain expertise, time, and effort. We address these labelling challenges by only req… ▽ More Coronavirus Disease 2019 (COVID-19) has spread aggressively across the world causing an existential health crisis. Thus, having a system that automatically detects COVID-19 in tomography (CT) images can assist in quantifying the severity of the illness. Unfortunately, labelling chest CT scans requires significant domain expertise, time, and effort. We address these labelling challenges by only requiring point annotations, a single pixel for each infected region on a CT image. This labeling scheme allows annotators to label a pixel in a likely infected region, only taking 1-3 seconds, as opposed to 10-15 seconds to segment a region. Conventionally, segmentation models train on point-level annotations using the cross-entropy loss function on these labels. However, these models often suffer from low precision. Thus, we propose a consistency-based (CB) loss function that encourages the output predictions to be consistent with spatial transformations of the input images. The experiments on 3 open-source COVID-19 datasets show that this loss function yields significant improvement over conventional point-level loss functions and almost matches the performance of models trained with full supervision with much less human effort. Code is available at: \url{https://github.com/IssamLaradji/covid19_weak_supervision}. △ Less

Submitted 7 July, 2020; v1 submitted 4 July, 2020; originally announced July 2020.

Ultrahyperbolic Representation Learning

Authors: Marc T. Law, Jos Stam

Abstract: In machine learning, data is usually represented in a (flat) Euclidean space where distances between points are along straight lines. Researchers have recently considered more exotic (non-Euclidean) Riemannian manifolds such as hyperbolic space which is well suited for tree-like data. In this paper, we propose a representation living on a pseudo-Riemannian manifold of constant nonzero curvature. I… ▽ More In machine learning, data is usually represented in a (flat) Euclidean space where distances between points are along straight lines. Researchers have recently considered more exotic (non-Euclidean) Riemannian manifolds such as hyperbolic space which is well suited for tree-like data. In this paper, we propose a representation living on a pseudo-Riemannian manifold of constant nonzero curvature. It is a generalization of hyperbolic and spherical geometries where the nondegenerate metric tensor need not be positive definite. We provide the necessary learning tools in this geometry and extend gradient-based optimization techniques. More specifically, we provide closed-form expressions for distances via geodesics and define a descent direction to minimize some objective function. Our novel framework is applied to graph representations. △ Less

Submitted 10 January, 2021; v1 submitted 30 June, 2020; originally announced July 2020.

Comments: NeurIPS 2020

Evryscope and K2 Constraints on TRAPPIST-1 Superflare Occurrence and Planetary Habitability

Authors: Amy L. Glazier, Ward S. Howard, Hank Corbett, Nicholas M. Law, Jeffrey K. Ratzloff, Octavi Fors, Daniel del Ser

Abstract: The nearby ultracool dwarf TRAPPIST-1 possesses several Earth-sized terrestrial planets, three of which have equilibrium temperatures that may support liquid surface water, making it a compelling target for exoplanet characterization. TRAPPIST-1 is an active star with frequent flaring, with implications for the habitability of its planets. Superflares (stellar flares whose energy exceeds 10^33 erg… ▽ More The nearby ultracool dwarf TRAPPIST-1 possesses several Earth-sized terrestrial planets, three of which have equilibrium temperatures that may support liquid surface water, making it a compelling target for exoplanet characterization. TRAPPIST-1 is an active star with frequent flaring, with implications for the habitability of its planets. Superflares (stellar flares whose energy exceeds 10^33 erg) can completely destroy the atmospheres of a cool star's planets, allowing ultraviolet radiation and high-energy particles to bombard their surfaces. However, ultracool dwarfs emit little ultraviolet flux when quiescent, raising the possibility of frequent flares being necessary for prebiotic chemistry that requires ultraviolet light. We combine Evryscope and Kepler observations to characterize the high-energy flare rate of TRAPPIST-1. The Evryscope is an array of 22 small telescopes imaging the entire Southern sky in g' every two minutes. Evryscope observations, spanning 170 nights over 2 years, complement the 80-day continuous short-cadence K2 observations by sampling TRAPPIST-1's long-term flare activity. We update TRAPPIST-1's superflare rate, finding a cumulative rate of 4.2 (+1.9 -0.2) superflares per year. We calculate the flare rate necessary to deplete ozone in the habitable-zone planets' atmospheres, and find that TRAPPIST-1's flare rate is insufficient to deplete ozone if present on its planets. In addition, we calculate the flare rate needed to provide enough ultraviolet flux to power prebiotic chemistry. We find TRAPPIST-1's flare rate is likely insufficient to catalyze some of the Earthlike chemical pathways thought to lead to RNA synthesis, and flux due to flares in the biologically relevant UV-B band is orders of magnitude less for any TRAPPIST-1 planet than has been experienced by Earth at any time in its history. △ Less

Submitted 17 August, 2020; v1 submitted 25 June, 2020; originally announced June 2020.

Comments: 12 pages, 9 figures. Accepted to The Astrophysical Journal, in press

Multi-wavelength Photometry and Progenitor Analysis of the Nova V906 Car

Authors: Jerrick Wee, Nadejda Blagorodnova, Bryan Edward Penprase, Jett Pierce Facey, Taiga Morioka, Hank Corbett, Brad N. Barlow, Thomas Kupfer, Nicholas M. Law, Jeffrey K. Ratzloff, Ward S. Howard, Ramses Gonzalez Chavez, Amy Glazier, Alan Vasquez Soto, Takashi Horiuchi

Abstract: We present optical and infrared photometry of the classical nova V906 Car, also known as Nova Car 2018 and ASASSN-18fv, discovered by ASASS-SN survey on 16.32 March 2018 UT (MJD 58193.0). The nova reached its maximum on MJD 58222.56 at $V_{\rm{max}} = 5.84 \pm 0.09$ mag and had decline times of $t_{2,V} = 26.2 $ d and $t_{3,V} = 33.0 $ d. The data from Evryscope shows that the nova had already bri… ▽ More We present optical and infrared photometry of the classical nova V906 Car, also known as Nova Car 2018 and ASASSN-18fv, discovered by ASASS-SN survey on 16.32 March 2018 UT (MJD 58193.0). The nova reached its maximum on MJD 58222.56 at $V_{\rm{max}} = 5.84 \pm 0.09$ mag and had decline times of $t_{2,V} = 26.2 $ d and $t_{3,V} = 33.0 $ d. The data from Evryscope shows that the nova had already brightened to $g'\simeq 13$\,mag five days before discovery, as compared to its quiescent magnitude of $g=$20.13$\pm$0.03. The extinction towards the nova, as derived from high resolution spectroscopy, shows an estimate consistent with foreground extinction to the Carina Nebula of $A_V = 1.11_{-0.39}^{+0.54}$. The light curve resembles a rare C (cusp) class nova with a steep decline slope of $α=-3.94$ post cusp flare. From the lightcurve decline rate, we estimate the mass of white dwarf to be $M_{WD}$ = $ < 0.8$M\textsubscript{\(\odot\)}, consistent with $M_{WD}=0.71^{+0.23}_{-0.19}$ derived from modelling the accretion disk of the system in quiescence. The donor star is likely a K-M dwarf of 0.23-0.43\,\Msun, which is being heated by its companion. △ Less

Submitted 25 June, 2020; originally announced June 2020.

Comments: 13 pages, 10 figures, 7 tables. Accepted to ApJ

The ILASP system for Inductive Learning of Answer Set Programs

Authors: Mark Law, Alessandra Russo, Krysia Broda

Abstract: The goal of Inductive Logic Programming (ILP) is to learn a program that explains a set of examples in the context of some pre-existing background knowledge. Until recently, most research on ILP targeted learning Prolog programs. Our own ILASP system instead learns Answer Set Programs, including normal rules, choice rules and hard and weak constraints. Learning such expressive programs widens the… ▽ More The goal of Inductive Logic Programming (ILP) is to learn a program that explains a set of examples in the context of some pre-existing background knowledge. Until recently, most research on ILP targeted learning Prolog programs. Our own ILASP system instead learns Answer Set Programs, including normal rules, choice rules and hard and weak constraints. Learning such expressive programs widens the applicability of ILP considerably; for example, enabling preference learning, learning common-sense knowledge, including defaults and exceptions, and learning non-deterministic theories. In this paper, we first give a general overview of ILASP's learning framework and its capabilities. This is followed by a comprehensive summary of the evolution of the ILASP system, presenting the strengths and weaknesses of each version, with a particular emphasis on scalability. △ Less

Submitted 2 May, 2020; originally announced May 2020.

Comments: Submitted to the ALP newsletter

TESS Hunt for Young and Maturing Exoplanets (THYME) II: A 17 Myr Old Transiting Hot Jupiter in the Sco-Cen Association

Authors: Aaron C. Rizzuto, Elisabeth R. Newton, Andrew W. Mann, Benjamin M. Tofflemire, Andrew Vanderburg, Adam L. Kraus, Mackenna L. Wood, Samuel N. Quinn, George Zhou, Pa Chia Thao, Nicholas M. Law, Carl Ziegler, Cesar Briceno

Abstract: We present the discovery of a transiting hot Jupiter orbiting HIP 67522 ($T_{eff}\sim5650$ K; $M_* \sim 1.2 M_{\odot}$) in the 10-20 Myr old Sco-Cen OB association. We identified the transits in the TESS data using our custom notch-filter planet search pipeline, and characterize the system with additional photometry from Spitzer, spectroscopy from SOAR/Goodman, SALT/HRS, LCOGT/NRES, and SMARTS/CHI… ▽ More We present the discovery of a transiting hot Jupiter orbiting HIP 67522 ($T_{eff}\sim5650$ K; $M_* \sim 1.2 M_{\odot}$) in the 10-20 Myr old Sco-Cen OB association. We identified the transits in the TESS data using our custom notch-filter planet search pipeline, and characterize the system with additional photometry from Spitzer, spectroscopy from SOAR/Goodman, SALT/HRS, LCOGT/NRES, and SMARTS/CHIRON, and speckle imaging from SOAR/HRCam. We model the photometry as a periodic Gaussian process with transits to account for stellar variability, and find an orbital period of 6.9596$^{+0.000016}_{-0.000015}$ days and radius of 10.02$^{+0.54}_{-0.53}$ R$_\oplus$. We also identify a single transit of an additional candidate planet with radius 8.01$^{+0.75}_{-0.71}$ R$_\oplus$ that has an orbital period of $\gtrsim23$ days. The validated planet HIP 67522 b is currently the youngest transiting hot Jupiter discovered and is an ideal candidate for transmission spectroscopy and radial velocity follow-up studies, while also demonstrating that some young giant planets either form in situ at small orbital radii, or else migrate promptly from formation sites farther out in the disk. △ Less

Submitted 5 May, 2020; v1 submitted 30 April, 2020; originally announced May 2020.

Comments: Submitted to AAS Journals, 19 pages, 10 figures, 6 tables

Robo-AO M Dwarf Multiplicity Survey: Catalog

Authors: Claire Lamman, Christoph Baranec, Zachory K. Berta-Thompson, Nicholas M. Law, Jessica Schonhut-Stasik, Carl Ziegler, Maissa Salama, Rebecca Jensen-Clem, Dmitry A. Duev, Reed Riddle, Shrinivas R. Kulkarni, Jennifer G. Winters, Jonathan M. Irwin

Abstract: We analyze observations from Robo-AO's field M dwarf survey taken on the 2.1m Kitt Peak telescope and perform a multiplicity comparison with Gaia DR2. Through its laser-guided, automated system, the Robo-AO instrument has yielded the largest adaptive optics M dwarf multiplicity survey to date. After develo** an interface to visually identify and locate stellar companions, we selected eleven lows… ▽ More We analyze observations from Robo-AO's field M dwarf survey taken on the 2.1m Kitt Peak telescope and perform a multiplicity comparison with Gaia DR2. Through its laser-guided, automated system, the Robo-AO instrument has yielded the largest adaptive optics M dwarf multiplicity survey to date. After develo** an interface to visually identify and locate stellar companions, we selected eleven lowsignificance Robo-AO detections for follow-up on the Keck II telescope using NIRC2. In the Robo-AO survey we find 553 candidate companions within 4" around 534 stars out of 5566 unique targets, most of which are new discoveries. Using a position cross match with DR2 on all targets, we assess the binary recoverability of Gaia DR2 and compare the properties of multiples resolved by both Robo-AO and Gaia. The catalog of nearby M dwarf systems and their basic properties presented here can assist other surveys which observe these stars, such as the NASA TESS mission. △ Less

Submitted 16 January, 2020; originally announced January 2020.

Comments: 16 pages, 9 figures. To be published in The Astronomical Journal

The Robotilter: An Automated Lens / CCD Alignment System for the Evryscope

Authors: Jeffrey K. Ratzloff, Nicholas M. Law, Henry T. Corbett, Octavi Fors, Daniel del Ser

Abstract: Camera lenses are increasingly used in wide-field astronomical surveys due to their high performance, wide field-of-view (FOV) unreachable from traditional telescope optics, and modest cost. The machining and assembly tolerances for commercially available optical systems cause a slight misalignment (tilt) between the lens and CCD, resulting in PSF degradation. We have built an automated alignment… ▽ More Camera lenses are increasingly used in wide-field astronomical surveys due to their high performance, wide field-of-view (FOV) unreachable from traditional telescope optics, and modest cost. The machining and assembly tolerances for commercially available optical systems cause a slight misalignment (tilt) between the lens and CCD, resulting in PSF degradation. We have built an automated alignment system (Robotilters) to solve this challenge, optimizing 4 degrees of freedom - 2 tilt axes, a separation axis (the distance between the CCD and lens), and the lens focus (the built-in focus of the lens by turning the lens focusing ring which moves the optical elements relative to one another) in a compact and low-cost package. The Robotilters remove tilt and optimize focus at the sub 10 micron level, are completely automated, take 2 hours to run, and remain stable for multiple years once aligned. The Robotilters were built for the Evryscope telescope (a 780 MPix 22-camera array with an 8150 sq.deg. field of view and continuous 2-minute cadence) designed to detect short timescale events across extremely large sky areas simultaneously. Variance in quality across the image field, especially the corners and edges compared to the center, is a significant challenge in wide-field astronomical surveys like the Evryscope. The individual star PSFs (which typically extend only a few pixels) are highly susceptible to slight increases in optical aberrations in this situation. The Robotilter solution resulted in a limiting magnitude improvement of .5 mag in the center of the image and 1.0 mag in the corners for typical Evryscope cameras, with less distorted and smaller PSFs (half the extent in the corners and edges in many cases). In this paper we describe the Robotilter mechanical and software design, camera alignment results, long term stability, and image improvement. △ Less

Submitted 3 January, 2020; originally announced January 2020.

Comments: Accepted to JATIS, January 2020

Hot Subdwarf All Southern Sky Fast Transit Survey with the Evryscope

Authors: Jeffrey K. Ratzloff, Brad N. Barlow, Peter Nemeth, Henry T. Corbett, Stephen Walser, Nathan W. Galliher, Amy Glazier, Ward S. Howard, Nicholas M. Law

Abstract: We have conducted a survey of candidate hot subdwarf stars in the southern sky searching for fast transits, eclipses, and sinusoidal like variability in the Evryscope light curves. The survey aims to detect transit signals from Neptune size planets to gas-giants, and eclipses from M-dwarfs and brown dwarfs. The other variability signals are primarily expected to be from compact binaries and reflec… ▽ More We have conducted a survey of candidate hot subdwarf stars in the southern sky searching for fast transits, eclipses, and sinusoidal like variability in the Evryscope light curves. The survey aims to detect transit signals from Neptune size planets to gas-giants, and eclipses from M-dwarfs and brown dwarfs. The other variability signals are primarily expected to be from compact binaries and reflection effect binaries. Due to the small size of hot subdwarfs, transit and eclipse signals are expected to last only twenty minutes, but with large signal depths (up to completely eclipsing if the orientation is edge on). With its 2-minute cadence and continuous observing Evryscope is well placed to recover these fast transits and eclipses. The very large field of view (8150 sq. deg.) is critical to obtain enough hot subdwarf targets, despite their rarity. We identified 11,000 potential hot subdwarfs from the 9.3M Evryscope light curves for sources brighter than mg = 15. With our machine learning spectral classifier, we flagged high-confidence targets and estimate the total hot subdwarfs in the survey to be 1400. The light curve search detected three planet transit candidates, shown to have stellar companions from followup analysis. We discovered several new compact binaries (including two with unseen degenerate companions, and several others with potentially rare secondaries), two eclipsing binaries with M-dwarf companions, as well as new reflection effect binaries and others with sinusoidal like variability. The hot subdwarf discoveries identified here are spectroscopically confirmed and we verified the Evryscope discovery light curve with TESS light curves when available. Four of the discoveries are in the process of being published in separate followup papers, and we discuss the followup potential of several of the other discoveries. △ Less

Submitted 23 December, 2019; originally announced December 2019.

Comments: Accepted to ApJ, December 2019

TOI 564 b and TOI 905 b: Grazing and Fully Transiting Hot Jupiters Discovered by TESS

Authors: Allen B. Davis, Songhu Wang, Matias Jones, Jason D. Eastman, Maximilian N. Günther, Keivan G. Stassun, Brett C. Addison, Karen A. Collins, Samuel N. Quinn, David W. Latham, Trifon Trifonov, Sahar Shahaf, Tsevi Mazeh, Stephen R. Kane, Xian-Yu Wang, Thiam-Guan Tan, Andrei Tokovinin, Carl Ziegler, René Tronsgaard, Sarah Millholland, Bryndis Cruz, Perry Berlind, Michael L. Calkins, Gilbert A. Esquerdo, Kevin I. Collins , et al. (24 additional authors not shown)

Abstract: We report the discovery and confirmation of two new hot Jupiters discovered by the Transiting Exoplanet Survey Satellite (TESS): TOI 564 b and TOI 905 b. The transits of these two planets were initially observed by TESS with orbital periods of 1.651 d and 3.739 d, respectively. We conducted follow-up observations of each system from the ground, including photometry in multiple filters, speckle int… ▽ More We report the discovery and confirmation of two new hot Jupiters discovered by the Transiting Exoplanet Survey Satellite (TESS): TOI 564 b and TOI 905 b. The transits of these two planets were initially observed by TESS with orbital periods of 1.651 d and 3.739 d, respectively. We conducted follow-up observations of each system from the ground, including photometry in multiple filters, speckle interferometry, and radial velocity measurements. For TOI 564 b, our global fitting revealed a classical hot Jupiter with a mass of $1.463^{+0.10}_{-0.096}\ M_J$ and a radius of $1.02^{+0.71}_{-0.29}\ R_J$. TOI 905 b is a classical hot Jupiter as well, with a mass of $0.667^{+0.042}_{-0.041}\ M_J$ and radius of $1.171^{+0.053}_{-0.051}\ R_J$. Both planets orbit Sun-like, moderately bright, mid-G dwarf stars with V ~ 11. While TOI 905 b fully transits its star, we found that TOI 564 b has a very high transit impact parameter of $0.994^{+0.083}_{-0.049}$, making it one of only ~20 known systems to exhibit a grazing transit and one of the brightest host stars among them. TOI 564 b is therefore one of the most attractive systems to search for additional non-transiting, smaller planets by exploiting the sensitivity of grazing transits to small changes in inclination and transit duration over the time scale of several years. △ Less

Submitted 20 December, 2019; originally announced December 2019.

Comments: 21 pages and 10 figures. Submitted to AJ

Induction of Subgoal Automata for Reinforcement Learning

Authors: Daniel Furelos-Blanco, Mark Law, Alessandra Russo, Krysia Broda, Anders Jonsson

Abstract: In this work we present ISA, a novel approach for learning and exploiting subgoals in reinforcement learning (RL). Our method relies on inducing an automaton whose transitions are subgoals expressed as propositional formulas over a set of observable events. A state-of-the-art inductive logic programming system is used to learn the automaton from observation traces perceived by the RL agent. The re… ▽ More In this work we present ISA, a novel approach for learning and exploiting subgoals in reinforcement learning (RL). Our method relies on inducing an automaton whose transitions are subgoals expressed as propositional formulas over a set of observable events. A state-of-the-art inductive logic programming system is used to learn the automaton from observation traces perceived by the RL agent. The reinforcement learning and automaton learning processes are interleaved: a new refined automaton is learned whenever the RL agent generates a trace not recognized by the current automaton. We evaluate ISA in several gridworld problems and show that it performs similarly to a method for which automata are given in advance. We also show that the learned automata can be exploited to speed up convergence through reward sha** and transfer learning across multiple tasks. Finally, we analyze the running time and the number of traces that ISA needs to learn an automata, and the impact that the number of observable events has on the learner's performance. △ Less

Submitted 29 November, 2019; originally announced November 2019.

Comments: Preprint accepted for publication to the 34th AAAI Conference on Artificial Intelligence (AAAI-20)

TOI-132 b: A short-period planet in the Neptune desert transiting a $V=11.3$ G-type star

Authors: Matías R. Díaz, James S. Jenkins, Davide Gandolfi, Eric D. Lopez, Maritza G. Soto, Pía Cortés-Zuleta, Zaira M. Berdiñas, Keivan G. Stassun, Karen A. Collins, José I. Vines, Carl Ziegler, Malcolm Fridlund, Eric J. N. Jensen, Felipe Murgas, Alexandre Santerne, Paul A. Wilson, Massimiliano Esposito, Artie P. Hatzes, Marshall C. Johnson, Kristine W. F. Lam, John H. Livingston, Vincent Van Eylen, Norio Narita, César Briceño, Kevin I. Collins , et al. (23 additional authors not shown)

Abstract: The Neptune desert is a feature seen in the radius-mass-period plane, whereby a notable dearth of short period, Neptune-like planets is found. Here we report the {\it TESS} discovery of a new short-period planet in the Neptune desert, orbiting the G-type dwarf TYC\,8003-1117-1 (TOI-132). {\it TESS} photometry shows transit-like dips at the level of $\sim$1400 ppm occurring every $\sim$2.11 days. H… ▽ More The Neptune desert is a feature seen in the radius-mass-period plane, whereby a notable dearth of short period, Neptune-like planets is found. Here we report the {\it TESS} discovery of a new short-period planet in the Neptune desert, orbiting the G-type dwarf TYC\,8003-1117-1 (TOI-132). {\it TESS} photometry shows transit-like dips at the level of $\sim$1400 ppm occurring every $\sim$2.11 days. High-precision radial velocity follow-up with HARPS confirmed the planetary nature of the transit signal and provided a semi-amplitude radial velocity variation of $\sim$11.5 m s$^{-1}$, which, when combined with the stellar mass of $0.97\pm0.06$ $M_{\odot}$, provides a planetary mass of 22.83$^{+1.81}_{-1.80}$ $M_{\oplus}$. Modeling the {\it TESS} high-quality light curve returns a planet radius of 3.43$^{+0.13}_{-0.14}$ $R_{\oplus}$, and therefore the planet bulk density is found to be 3.11$^{+0.44}_{-0.450}$ g cm$^{-3}$. Planet structure models suggest that the bulk of the planet mass is in the form of a rocky core, with an atmospheric mass fraction of 4.3$^{+1.2}_{-2.3}$\%. TOI-132 b is a {\it TESS} Level 1 Science Requirement candidate, and therefore priority follow-up will allow the search for additional planets in the system, whilst hel** to constrain low-mass planet formation and evolution models, particularly valuable for better understanding the Neptune desert. △ Less

Submitted 18 November, 2019; v1 submitted 5 November, 2019; originally announced November 2019.

Comments: 12 pages, 10 figures, 4 tables. Submitted to MNRAS. Comments welcome. Missing labels, Typos fixed

Adversarial Pulmonary Pathology Translation for Pairwise Chest X-ray Data Augmentation

Authors: Yunyan Xing, Zongyuan Ge, Rui Zeng, Dwarikanath Mahapatra, Jarrel Seah, Meng Law, Tom Drummond

Abstract: Recent works show that Generative Adversarial Networks (GANs) can be successfully applied to chest X-ray data augmentation for lung disease recognition. However, the implausible and distorted pathology features generated from the less than perfect generator may lead to wrong clinical decisions. Why not keep the original pathology region? We proposed a novel approach that allows our generative mode… ▽ More Recent works show that Generative Adversarial Networks (GANs) can be successfully applied to chest X-ray data augmentation for lung disease recognition. However, the implausible and distorted pathology features generated from the less than perfect generator may lead to wrong clinical decisions. Why not keep the original pathology region? We proposed a novel approach that allows our generative model to generate high quality plausible images that contain undistorted pathology areas. The main idea is to design a training scheme based on an image-to-image translation network to introduce variations of new lung features around the pathology ground-truth area. Moreover, our model is able to leverage both annotated disease images and unannotated healthy lung images for the purpose of generation. We demonstrate the effectiveness of our model on two tasks: (i) we invite certified radiologists to assess the quality of the generated synthetic images against real and other state-of-the-art generative models, and (ii) data augmentation to improve the performance of disease localisation. △ Less

Submitted 21 January, 2020; v1 submitted 10 October, 2019; originally announced October 2019.

Comments: Code: https://github.com/yunyanxing/pairwise_xray_augmentation - Accepted to the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) 2019

Robo-AO Kepler Asteroseismic Survey. II. Do Stellar Companions Inhibit Stellar Oscillations?

Authors: Jessica Schonhut-Stasik, Daniel Huber, Christoph Baranec, Claire Lamman, Maissa Salama, Reecca Jensen-Clem, Dmitry A. Duev, Reed Riddle, S. R. Kulkarni, Nicholas M. Law

Abstract: The Kepler space telescope observed over 15,000 stars for asteroseismic studies. Of these, 75% of dwarfs (and 8% of giants) were found to show anomalous behavior: such as suppressed oscillations (low amplitude) or no oscillations at all. The lack of solar-like oscillations may be a consequence of multiplicity, due to physical interactions with spectroscopic companions or due to the dilution of osc… ▽ More The Kepler space telescope observed over 15,000 stars for asteroseismic studies. Of these, 75% of dwarfs (and 8% of giants) were found to show anomalous behavior: such as suppressed oscillations (low amplitude) or no oscillations at all. The lack of solar-like oscillations may be a consequence of multiplicity, due to physical interactions with spectroscopic companions or due to the dilution of oscillation amplitudes from "wide'" (AO detected; visual) or spectroscopic companions introducing contaminating flux. We present a search for stellar companions to 327 of the Kepler asteroseismic sample, which were expected to display solar-like oscillations. We used direct imaging with Robo-AO, which can resolve secondary sources at ~0."15, and followed up detected companions with Keck AO. Directly imaged companion systems with both separations of $\leq$0."5 and amplitude dilutions >10% all have anomalous primaries, suggesting these oscillation signals are diluted by a sufficient amount of excess flux. We also used the high-resolution spectrometer ESPaDOnS at CFHT to search for spectroscopic binaries. We find tentative evidence for a higher fraction of spectroscopic binaries with high radial velocity scatter in anomalous systems, which would be consistent with previous results suggesting that oscillations are suppressed by tidal interactions in close eclipsing binaries. △ Less

Submitted 9 October, 2019; originally announced October 2019.

Comments: 19 pages, 13 of text. 12 plots, 6 tables. Accepted to ApJ

A Theoretical Analysis of the Number of Shots in Few-Shot Learning

Authors: Tianshi Cao, Marc Law, Sanja Fidler

Abstract: Few-shot classification is the task of predicting the category of an example from a set of few labeled examples. The number of labeled examples per category is called the number of shots (or shot number). Recent works tackle this task through meta-learning, where a meta-learner extracts information from observed tasks during meta-training to quickly adapt to new tasks during meta-testing. In this… ▽ More Few-shot classification is the task of predicting the category of an example from a set of few labeled examples. The number of labeled examples per category is called the number of shots (or shot number). Recent works tackle this task through meta-learning, where a meta-learner extracts information from observed tasks during meta-training to quickly adapt to new tasks during meta-testing. In this formulation, the number of shots exploited during meta-training has an impact on the recognition performance at meta-test time. Generally, the shot number used in meta-training should match the one used in meta-testing to obtain the best performance. We introduce a theoretical analysis of the impact of the shot number on Prototypical Networks, a state-of-the-art few-shot classification method. From our analysis, we propose a simple method that is robust to the choice of shot number used during meta-training, which is a crucial hyperparameter. The performance of our model trained for an arbitrary meta-training shot number shows great performance for different values of meta-testing shot numbers. We experimentally demonstrate our approach on different few-shot classification benchmarks. △ Less

Submitted 14 February, 2020; v1 submitted 25 September, 2019; originally announced September 2019.

Comments: 15 pages incl. appendix, 6 figures

Journal ref: 8th International Conference on Learning Representations 2020, Addis Ababa, Ethiopia

Optimal curtailed designs for single arm phase II clinical trials

Authors: Martin Law, Michael J. Grayling, Adrian P. Mander

Abstract: In single-arm phase II oncology trials, the most popular choice of design is Simon's two-stage design, which allows early stop** at one interim analysis. However, the expected trial sample size can be reduced further by allowing curtailment. Curtailment is stop** when the final go or no-go decision is certain, so-called non-stochastic curtailment, or very likely, known as stochastic curtailmen… ▽ More In single-arm phase II oncology trials, the most popular choice of design is Simon's two-stage design, which allows early stop** at one interim analysis. However, the expected trial sample size can be reduced further by allowing curtailment. Curtailment is stop** when the final go or no-go decision is certain, so-called non-stochastic curtailment, or very likely, known as stochastic curtailment. In the context of single-arm phase II oncology trials, stochastic curtailment has previously been restricted to stop** in the second stage and/or stop** for a no-go decision only. We introduce two designs that incorporate stochastic curtailment and allow stop** after every observation, for either a go or no-go decision. We obtain optimal stop** boundaries by searching over a range of potential conditional powers, beyond which the trial will stop for a go or no-go decision. This search is novel: firstly, the search is undertaken over a range of values unique to each possible design realisation. Secondly, these values are evaluated taking into account the possibility of early stop**. Finally, each design realisation's operating characteristics are obtained exactly. The proposed designs are compared to existing designs in a real data example. They are also compared under three scenarios, both with respect to four single optimality criteria and using a loss function. The proposed designs are superior in almost all cases. Optimising for the expected sample size under either the null or alternative hypothesis, the saving compared to the popular Simon's design ranges from 22% to 55%. △ Less

Submitted 6 September, 2019; originally announced September 2019.

EVR-CB-001: An evolving, progenitor, white dwarf compact binary discovered with the Evryscope

Authors: Jeffrey K. Ratzloff, Brad N. Barlow, Thomas Kupfer, Kyle A. Corcoran, Stephan Geier, Evan Bauer, Henry T. Corbett, Ward S. Howard, Amy Glazier, Nicholas M. Law

Abstract: We present EVR-CB-001, the discovery of a compact binary with an extremely low mass ($.21 \pm 0.05 M_{\odot}$) helium core white dwarf progenitor (pre-He WD) and an unseen low mass ($.32 \pm 0.06 M_{\odot}$) helium white dwarf (He WD) companion. He WDs are thought to evolve from the remnant helium-rich core of a main-sequence star stripped during the giant phase by a close companion. Low mass He W… ▽ More We present EVR-CB-001, the discovery of a compact binary with an extremely low mass ($.21 \pm 0.05 M_{\odot}$) helium core white dwarf progenitor (pre-He WD) and an unseen low mass ($.32 \pm 0.06 M_{\odot}$) helium white dwarf (He WD) companion. He WDs are thought to evolve from the remnant helium-rich core of a main-sequence star stripped during the giant phase by a close companion. Low mass He WDs are exotic objects (only about .2$\%$ of WDs are thought to be less than .3 $M_{\odot}$), and are expected to be found in compact binaries. Pre-He WDs are even rarer, and occupy the intermediate phase after the core is stripped, but before the star becomes a fully degenerate WD and with a larger radius ($\approx .2 R_{\odot}$) than a typical WD. The primary component of EVR-CB-001 (the pre-He WD) was originally thought to be a hot subdwarf (sdB) star from its blue color and under-luminous magnitude, characteristic of sdBs. The mass, temperature ($T_{\rm eff}=18,500 \pm 500 K$), and surface gravity ($\log(g)=4.96 \pm 0.04$) solutions from this work are lower than values for typical hot subdwarfs. The primary is likely to be a post-RGB, pre-He WD contracting into a He WD, and at a stage that places it nearest to sdBs on color-magnitude and $T_{\rm eff}$-$\log(g)$ diagrams. EVR-CB-001 is expected to evolve into a fully double degenerate, compact system that should spin down and potentially evolve into a single hot subdwarf star. Single hot subdwarfs are observed, but progenitor systems have been elusive. △ Less

Submitted 3 October, 2019; v1 submitted 4 September, 2019; originally announced September 2019.

Comments: 14 pages, 11 figures. Published in The Astrophysical Journal

Dispersions of Many-Body Bethe strings

Authors: Anup Kumar Bera, Jianda Wu, Wang Yang, Zhe Wang, Robert Bewley, Martin Boehm, Maciej Bartkowiak, Oleksandr Prokhnenko, Bastian Klemke, A. T. M. Nazmul Islam, Joseph Mathew Law, Bella Lake

Abstract: Complex bound states of magnetic excitations, known as Bethe string, were predicted almost a century ago to exist in one-dimensional quantum magnets 1. The dispersions of the string states have so far remained the subject of intensive theoretical studies 2-7. By performing neutron scattering experiments on the one-dimensional Heisenberg-Ising antiferromagnet SrCo2V2O8 in high longitudinal magnetic… ▽ More Complex bound states of magnetic excitations, known as Bethe string, were predicted almost a century ago to exist in one-dimensional quantum magnets 1. The dispersions of the string states have so far remained the subject of intensive theoretical studies 2-7. By performing neutron scattering experiments on the one-dimensional Heisenberg-Ising antiferromagnet SrCo2V2O8 in high longitudinal magnetic fields, we reveal in detail the dispersion relations of the string states over the full Brillouin zone, as well as their magnetic field dependences. Furthermore the characteristic energy, the scattering intensity and linewidth of the observed string states exhibit excellent agreement with our precise Bethe Ansatz calculations. Our results establish the important role of string states in the quantum spin dynamics of one-dimensional systems, and will invoke studies of their dynamical properties in more general many-body systems. △ Less

Submitted 31 August, 2019; originally announced September 2019.

Comments: Main manuscript: 6 pages, 4 figures, Supplementary information: 8 pages, 6 figures, 1 table

Disorder and Critical Current Variability in Josephson Junctions

Authors: Miguel Antonio Sulangi, T. A. Weingartner, N. Pokhrel, E. Patrick, M. Law, P. J. Hirschfeld

Abstract: We investigate theoretically the origins of observed variations in the critical currents of Nb/Al-AlO$_x$/Nb Josephson junctions in terms of various types of disorder. We consider the following disorder sources: vacancies within the Al layer; thickness variations in the AlO$_x$ layer; and "pinholes" (i.e., point contacts) within the AlO$_x$ layer. The calculations are all performed by solving the… ▽ More We investigate theoretically the origins of observed variations in the critical currents of Nb/Al-AlO$_x$/Nb Josephson junctions in terms of various types of disorder. We consider the following disorder sources: vacancies within the Al layer; thickness variations in the AlO$_x$ layer; and "pinholes" (i.e., point contacts) within the AlO$_x$ layer. The calculations are all performed by solving the microscopic Bogoliubov-de Gennes Hamiltonian self-consistently. It is found that a small concentration of vacancies within the Al layer is sufficient to suppress the critical current, while the presence of a small number of thick regions of the oxide layer induces a similar effect as well. The pinhole scenario is found to result in anomalous behavior that resembles neither that of a pure tunnel junction nor that of an SNS junction, but a regime that interpolates between these two limits. We comment on the degree to which each of the three scenarios describes the actual situation present in these junctions. △ Less

Submitted 13 February, 2020; v1 submitted 22 August, 2019; originally announced August 2019.

Comments: 10 pages, 12 figures

Journal ref: Journal of Applied Physics 127, 033901 (2020)

Low-Frequency Electronic Noise in Superlattice and Random-Packed Thin Films of Colloidal Quantum Dots

Authors: Adane Geremew, Caroline Qian, Alex Abelson, Sergey Rumyantsev, Fariborz Kargar, Matt Law, Alexander A. Balandin

Abstract: We report measurements of low-frequency electronic noise in ordered superlattice, weakly-ordered and random-packed thin films of 6.5 nm PbSe quantum dots prepared using several different ligand chemistries. For all samples, the normalized noise spectral density of the dark current revealed a Lorentzian component, reminiscent of the generation-recombination noise, superimposed on the 1/f background… ▽ More We report measurements of low-frequency electronic noise in ordered superlattice, weakly-ordered and random-packed thin films of 6.5 nm PbSe quantum dots prepared using several different ligand chemistries. For all samples, the normalized noise spectral density of the dark current revealed a Lorentzian component, reminiscent of the generation-recombination noise, superimposed on the 1/f background (f is the frequency). An activation energy of 0.3 eV was extracted from the temperature dependence of the noise spectra. The noise level in the ordered films was lower than that in the weakly-ordered and random-packed films. A large variation in the magnitude of the noise spectral density was also observed in samples with different ligand treatments. The obtained results are important for application of colloidal quantum dot films in photodetectors. △ Less

Submitted 10 August, 2019; originally announced August 2019.

Comments: 24 pages, 6 figures and supplemental info

Journal ref: Nanoscale, 11, 20171 (2019)

Video Face Clustering with Unknown Number of Clusters

Authors: Makarand Tapaswi, Marc T. Law, Sanja Fidler

Abstract: Understanding videos such as TV series and movies requires analyzing who the characters are and what they are doing. We address the challenging problem of clustering face tracks based on their identity. Different from previous work in this area, we choose to operate in a realistic and difficult setting where: (i) the number of characters is not known a priori; and (ii) face tracks belonging to min… ▽ More Understanding videos such as TV series and movies requires analyzing who the characters are and what they are doing. We address the challenging problem of clustering face tracks based on their identity. Different from previous work in this area, we choose to operate in a realistic and difficult setting where: (i) the number of characters is not known a priori; and (ii) face tracks belonging to minor or background characters are not discarded. To this end, we propose Ball Cluster Learning (BCL), a supervised approach to carve the embedding space into balls of equal size, one for each cluster. The learned ball radius is easily translated to a stop** criterion for iterative merging algorithms. This gives BCL the ability to estimate the number of clusters as well as their assignment, achieving promising results on commonly used datasets. We also present a thorough discussion of how existing metric learning literature can be adapted for this task. △ Less

Submitted 20 August, 2019; v1 submitted 9 August, 2019; originally announced August 2019.

Comments: Accepted to ICCV 2019, code and data at https://github.com/makarandtapaswi/BallClustering_ICCV2019

Estimating the Random Effect in Big Data Mixed Models

Authors: Michael Law, Ya'acov Ritov

Abstract: We consider three problems in high-dimensional Gaussian linear mixed models. Without any assumptions on the design for the fixed effects, we construct an asymptotic $F$-statistic for testing whether a collection of random effects is zero, derive an asymptotic confidence interval for a single random effect at the parametric rate $\sqrt{n}$, and propose an empirical Bayes estimator for a part of the… ▽ More We consider three problems in high-dimensional Gaussian linear mixed models. Without any assumptions on the design for the fixed effects, we construct an asymptotic $F$-statistic for testing whether a collection of random effects is zero, derive an asymptotic confidence interval for a single random effect at the parametric rate $\sqrt{n}$, and propose an empirical Bayes estimator for a part of the mean vector in ANOVA type models that performs asymptotically as well as the oracle Bayes estimator. We support our results with numerical simulations and provide comparisons with oracle estimators. The procedures developed are applied to the Trends in International Mathematics and Sciences Study (TIMSS) data. △ Less

Submitted 27 July, 2019; originally announced July 2019.

EvryFlare II: Rotation Periods of the Cool Flare Stars in TESS Across Half the Southern Sky

Authors: Ward S. Howard, Hank Corbett, Nicholas M. Law, Jeffrey K. Ratzloff, Nathan Galliher, Amy Glazier, Octavi Fors, Daniel del Ser, Joshua Haislip

Abstract: We measure rotation periods and sinusoidal amplitudes in Evryscope light curves for 122 two-minute K5-M4 TESS targets selected for strong flaring. The Evryscope array of telescopes has observed all bright nearby stars in the South, producing two-minute cadence light curves since 2016. Long-term, high-cadence observations of rotating flare stars probe the complex relationship between stellar rotati… ▽ More We measure rotation periods and sinusoidal amplitudes in Evryscope light curves for 122 two-minute K5-M4 TESS targets selected for strong flaring. The Evryscope array of telescopes has observed all bright nearby stars in the South, producing two-minute cadence light curves since 2016. Long-term, high-cadence observations of rotating flare stars probe the complex relationship between stellar rotation, starspots, and superflares. We detect periods from 0.3487 to 104 d, and observe amplitudes from 0.008 to 0.216 g' mag. We find the Evryscope amplitudes are larger than those in TESS with the effect correlated to stellar mass (p-value=0.01). We compute the Rossby number (Ro), and find our sample selected for flaring has twice as many intermediate rotators (0.04<Ro<0.4) as fast (Ro<0.04) or slow (Ro>0.44) rotators; this may be astrophysical or a result of period-detection sensitivity. We discover 30 fast, 59 intermediate, and 33 slow rotators. We measure a median starspot coverage of 13% of the stellar hemisphere and constrain the minimum magnetic field strength consistent with our flare energies and spot coverage to be 500 G, with later-type stars exhibiting lower values than earlier-types. We observe a possible change in superflare rates at intermediate periods. However, we do not conclusively confirm the increased activity of intermediate rotators seen in previous studies. We split all rotators at Ro~0.2 into Prot<10 d and Prot>10 d bins to confirm short-period rotators exhibit higher superflare rates, larger flare energies, and higher starspot coverage than do long-period rotators, at p-values of 3.2 X 10^-5, 1.0 X 10^-5, and 0.01, respectively. △ Less

Submitted 26 May, 2020; v1 submitted 24 July, 2019; originally announced July 2019.

Comments: 16 pages, 8 figures, 3 tables. Ancillary machine-readable files included. Accepted for publication in ApJ (proofs submitted). Includes significant new material, including starspot color that depends on stellar mass, more rotation periods, potential changes in activity during spin-down, and examples of binary rotators

Inductive general game playing

Authors: Andrew Cropper, Richard Evans, Mark Law

Abstract: General game playing (GGP) is a framework for evaluating an agent's general intelligence across a wide range of tasks. In the GGP competition, an agent is given the rules of a game (described as a logic program) that it has never seen before. The task is for the agent to play the game, thus generating game traces. The winner of the GGP competition is the agent that gets the best total score over a… ▽ More General game playing (GGP) is a framework for evaluating an agent's general intelligence across a wide range of tasks. In the GGP competition, an agent is given the rules of a game (described as a logic program) that it has never seen before. The task is for the agent to play the game, thus generating game traces. The winner of the GGP competition is the agent that gets the best total score over all the games. In this paper, we invert this task: a learner is given game traces and the task is to learn the rules that could produce the traces. This problem is central to inductive general game playing (IGGP). We introduce a technique that automatically generates IGGP tasks from GGP games. We introduce an IGGP dataset which contains traces from 50 diverse games, such as Sudoku, Sokoban, and Checkers. We claim that IGGP is difficult for existing inductive logic programming (ILP) approaches. To support this claim, we evaluate existing ILP systems on our dataset. Our empirical results show that most of the games cannot be correctly learned by existing systems. The best performing system solves only 40% of the tasks perfectly. Our results suggest that IGGP poses many challenges to existing approaches. Furthermore, because we can automatically generate IGGP tasks from GGP games, our dataset will continue to grow with the GGP competition, as new games are added every year. We therefore think that the IGGP problem and dataset will be valuable for motivating and evaluating future research. △ Less

Submitted 23 June, 2019; originally announced June 2019.

Comments: Accepted for the Machine Learning journal

A Hot Saturn Near (but unassociated with) the Open Cluster NGC 1817

Authors: Rayna Rampalli, Andrew Vanderburg, Allyson Bieryla, David W. Latham, Samuel N. Quinn, Christoph Baranec, Perry Berlind, Michael L. Calkins, William D. Cochran, Dmitry A. Duev, Michael Endl, Gilbert A. Esquerdo, Rebecca Jensen-Clem, Nicholas M. Law, Andrew W. Mayo, Reed Riddle, Maïssa Salama

Abstract: We report on the discovery of a hot Saturn-sized planet (9.916 +/- 0.985 R_earth) around a late F star, EPIC 246865365, observed in Campaign 13 of the K2 mission. We began studying this planet candidate because prior to the release of Gaia DR2, the host star was thought to have been a member (> 90% membership probability) of the approximately 1 Gyr open cluster NGC 1817 based on its kinematics and… ▽ More We report on the discovery of a hot Saturn-sized planet (9.916 +/- 0.985 R_earth) around a late F star, EPIC 246865365, observed in Campaign 13 of the K2 mission. We began studying this planet candidate because prior to the release of Gaia DR2, the host star was thought to have been a member (> 90% membership probability) of the approximately 1 Gyr open cluster NGC 1817 based on its kinematics and photometric distance. We identify the host star (among three stars within the K2 photometric aperture) using seeing-limited photometry and rule out false positive scenarios using adaptive optics imaging and radial velocity observations. We statistically validate EPIC 246865365b by calculating a false positive probability rate of 0.01%. However, we also show using new kinematic measurements provided by Gaia DR2 and our measured radial velocity of the system that EPIC 246865365 is unassociated with the cluster NGC 1817. Therefore, the long-running search for a giant transiting planet in an open cluster remains fruitless. Finally, we note that our use of seeing-limited photometry is a good demonstration of similar techniques that are already being used to follow up TESS planet candidates, especially in crowded regions. △ Less

Submitted 5 June, 2019; originally announced June 2019.

Comments: 11 pages, 4 figures, 1 table. Accepted in AJ

Investigating the origin of the spectral line profiles of the Hot Wolf-Rayet Star WR2

Authors: A. -N. Chené, N. St-Louis, A. F. J. Moffat, O. Schnurr, P. A. Crowther, G. A. Wade, N. D. Richardson, C. Baranec, C. A. Ziegler, N. M. Law, R. Riddle, G. A. Rate, É. Artigau, E. Alecian, BinaMIcS collaboration

Abstract: The hot WN star WR2 (HD6327) has been claimed to have many singular characteristics. To explain its unusually rounded and relatively weak emission line profiles, it has been proposed that WR2 is rotating close to break-up with a magnetically confined wind. Alternatively, the line profiles could be explained by the dilution of WR2's spectrum by that of a companion. In this paper, we present a study… ▽ More The hot WN star WR2 (HD6327) has been claimed to have many singular characteristics. To explain its unusually rounded and relatively weak emission line profiles, it has been proposed that WR2 is rotating close to break-up with a magnetically confined wind. Alternatively, the line profiles could be explained by the dilution of WR2's spectrum by that of a companion. In this paper, we present a study of WR2 using near-infrared AO imaging and optical spectroscopy and polarimetry. Our spectra reveal the presence of weak photospheric absorption lines from a ~B2.5-4V companion, which however contributes only ~5-10% to the total light, suggesting that the companion is a background object. Therefore, its flux cannot be causing any significant dilution of the WR star's emission lines. The absence of intrinsic linear continuum polarization from WR2 does not support the proposed fast rotation. Our Stokes V spectrum was not of sufficient quality to test the presence of a moderately strong organized magnetic field but our new modelling indicates that to confine the wind the putative magnetic field must be significantly stronger than was previously suggested sufficiently strong as to make its presence implausible. △ Less

Submitted 14 May, 2019; originally announced May 2019.

Comments: 12 pages, 9 figures

Journal ref: MNRAS, 484, 5834 (2019)

Variables in the Southern Polar Region Evryscope 2016 Dataset

Authors: Jeffrey K. Ratzloff, Henry T. Corbett, Nicholas M. Law, Brad N. Barlow, Amy Glazier, Ward S. Howard, Octavi Fors, Daniel del Ser, Trifon Trifonov

Abstract: The regions around the celestial poles offer the ability to find and characterize long-term variables from ground-based observatories. We used multi-year Evryscope data to search for high-amplitude (~5% or greater) variable objects among 160,000 bright stars (Mv < 14.5) near the South Celestial Pole. We developed a machine learning based spectral classifier to identify eclipse and transit candidat… ▽ More The regions around the celestial poles offer the ability to find and characterize long-term variables from ground-based observatories. We used multi-year Evryscope data to search for high-amplitude (~5% or greater) variable objects among 160,000 bright stars (Mv < 14.5) near the South Celestial Pole. We developed a machine learning based spectral classifier to identify eclipse and transit candidates with M-dwarf or K-dwarf host stars - and potential low-mass secondary stars or gas giant planets. The large amplitude transit signals from low-mass companions of smaller dwarf host stars lessens the photometric precision and systematics removal requirements necessary for detection, and increases the discoveries from long-term observations with modest light curve precision. The Evryscope is a robotic telescope array that observes the Southern sky continuously at 2-minute cadence, searching for stellar variability, transients, transits around exotic stars and other observationally challenging astrophysical variables. In this study, covering all stars 9 < Mv < 14.5, in declinations -75 to -90 deg, we recover 346 known variables and discover 303 new variables, including 168 eclipsing binaries. We characterize the discoveries and provide the amplitudes, periods, and variability type. A 1.7 Jupiter radius planet candidate with a late K-dwarf primary was found and the transit signal was verified with the PROMPT telescope network. Further followup revealed this object to be a likely grazing eclipsing binary system with nearly identical primary and secondary K5 stars. Radial velocity measurements from the Goodman Spectrograph on the 4.1 meter SOAR telescope of the likely-lowest-mass targets reveal that six of the eclipsing binary discoveries are low-mass (.06 - .37 solar mass) secondaries with K-dwarf primaries, strong candidates for precision mass-radius measurements. △ Less

Submitted 7 May, 2019; originally announced May 2019.

Comments: 32 pages, 17 figures, accepted to PASP

Building the Evryscope: Hardware Design and Performance

Authors: Jeffrey K. Ratzloff, Nicholas M. Law, Octavi Fors, Henry T. Corbett, Ward S. Howard, Daniel del Ser, Joshua Haislip

Abstract: The Evryscope is a telescope array designed to open a new parameter space in optical astronomy, detecting short timescale events across extremely large sky areas simultaneously. The system consists of a 780 MPix 22-camera array with an 8150 sq. deg. field of view, 13" per pixel sampling, and the ability to detect objects down to Mg=16 in each 2 minute dark-sky exposure. The Evryscope, covering 18,… ▽ More The Evryscope is a telescope array designed to open a new parameter space in optical astronomy, detecting short timescale events across extremely large sky areas simultaneously. The system consists of a 780 MPix 22-camera array with an 8150 sq. deg. field of view, 13" per pixel sampling, and the ability to detect objects down to Mg=16 in each 2 minute dark-sky exposure. The Evryscope, covering 18,400 sq.deg. with hours of high-cadence exposure time each night, is designed to find the rare events that require all-sky monitoring, including transiting exoplanets around exotic stars like white dwarfs and hot subdwarfs, stellar activity of all types within our galaxy, nearby supernovae, and other transient events such as gamma ray bursts and gravitational-wave electromagnetic counterparts. The system averages 5000 images per night with ~300,000 sources per image, and to date has taken over 3.0M images, totaling 250TB of raw data. The resulting light curve database has light curves for 9.3M targets, averaging 32,600 epochs per target through 2018. This paper summarizes the hardware and performance of the Evryscope, including the lessons learned during telescope design, electronics design, a procedure for the precision polar alignment of mounts for Evryscope-like systems, robotic control and operations, and safety and performance-optimization systems. We measure the on-sky performance of the Evryscope, discuss its data-analysis pipelines, and present some example variable star and eclipsing binary discoveries from the telescope. We also discuss new discoveries of very rare objects including 2 hot subdwarf eclipsing binaries with late M-dwarf secondaries (HW Vir systems), 2 white dwarf / hot subdwarf short-period binaries, and 4 hot subdwarf reflection binaries. We conclude with the status of our transit surveys, M-dwarf flare survey, and transient detection. △ Less

Submitted 26 April, 2019; originally announced April 2019.

Comments: 24 pages, 24 figures, accepted PASP

EvryFlare I: Long-term Evryscope Monitoring of Flares from the Cool Stars Across Half the Southern Sky

Authors: Ward S. Howard, Hank Corbett, Nicholas M. Law, Jeffrey K. Ratzloff, Amy L. Glazier, Octavi Fors, Daniel del Ser, Joshua Haislip

Abstract: We search for superflares from 4,068 cool stars in 2+ years of Evryscope photometry, focusing on those with high-cadence data from both Evryscope and TESS. The Evryscope array of small telescopes observed 575 flares from 284 stars, with a median energy of 10^34.0 erg. Since 2016, Evryscope has enabled the detection of rare events from all stars observed by TESS through multi-year, high-cadence con… ▽ More We search for superflares from 4,068 cool stars in 2+ years of Evryscope photometry, focusing on those with high-cadence data from both Evryscope and TESS. The Evryscope array of small telescopes observed 575 flares from 284 stars, with a median energy of 10^34.0 erg. Since 2016, Evryscope has enabled the detection of rare events from all stars observed by TESS through multi-year, high-cadence continuous observing. We report ~2X the previous largest number of 10^34 erg high-cadence flares from nearby cool stars. We find 8 flares with amplitudes of 3+ g' magnitudes, with the largest reaching 5.6 magnitudes and releasing 10^36.2 erg. We observe a 10^34 erg superflare from TOI-455 (LTT 1445), a mid-M with a rocky planet candidate. We measure the superflare rate per flare-star and quantify the average flaring of active stars as a function of spectral type, including superflare rates, FFDs, and typical flare amplitudes in g'. We confirm superflare morphology is broadly consistent with magnetic re-connection. We estimate starspot coverage necessary to produce superflares, and hypothesize maximum-allowed superflare energies and waiting-times between flares corresponding to 100% coverage of the stellar hemisphere. We observe decreased flaring at high galactic latitudes. We explore the effects of superflares on ozone loss to planetary atmospheres: we observe 1 superflare with sufficient energy to photo-dissociate all ozone in an Earth-like atmosphere in one event. We find 17 stars that may deplete an Earth-like atmosphere via repeated flaring. Of the 1822 stars around which TESS may discover temperate rocky planets, we observe 14.6% +/- 2% emit large flares. △ Less

Submitted 23 July, 2019; v1 submitted 23 April, 2019; originally announced April 2019.

Comments: 18 pages, 14 figures, 5 tables. Ancillary machine-readable files included. Accepted for publication in ApJ. Includes new sections on starspot coverage and superflare emission from starspots

HD 213885b: A transiting 1-day-period super-Earth with an Earth-like composition around a bright ($V=7.9$) star unveiled by TESS

Authors: Néstor Espinoza, Rafael Brahm, Thomas Henning, Andrés Jordán, Caroline Dorn, Felipe Rojas, Paula Sarkis, Diana Kossakowski, Martin Schlecker, Matías Díaz, James S. Jenkins, Claudia Aguilera-Gomez, Jon M. Jenkins, Joseph D. Twicken, Karen A. Collins, Jack Lissauer, David J. Armstrong, Vardan Adibekyan, David Barrado, Susana C. C. Barros, Matthew Battley, Daniel Bayliss, François Bouchy, Edward Bryant, Benjamin F. Cooke , et al. (38 additional authors not shown)

Abstract: We report the discovery of the 1.008-day, ultra-short period (USP) super-Earth HD 213885b (TOI-141b) orbiting the bright ($V=7.9$) star HD 213885 (TOI-141, TIC 403224672), detected using photometry from the recently launched TESS mission. Using FEROS, HARPS and CORALIE radial-velocities, we measure a precise mass of $8.8\pm0.6$ $M_\oplus$ for this $1.74 \pm 0.05$ $R_\oplus$ exoplanet, which provid… ▽ More We report the discovery of the 1.008-day, ultra-short period (USP) super-Earth HD 213885b (TOI-141b) orbiting the bright ($V=7.9$) star HD 213885 (TOI-141, TIC 403224672), detected using photometry from the recently launched TESS mission. Using FEROS, HARPS and CORALIE radial-velocities, we measure a precise mass of $8.8\pm0.6$ $M_\oplus$ for this $1.74 \pm 0.05$ $R_\oplus$ exoplanet, which provides enough information to constrain its bulk composition, which is similar to Earth's but enriched in iron. The radius, mass and stellar irradiation of HD 213885b are, given our data, very similar to 55 Cancri e, making this exoplanet a good target to perform comparative exoplanetology of short period, highly irradiated super-Earths. Our precise radial-velocities reveal an additional $4.78$-day signal which we interpret as arising from a second, non-transiting planet in the system, HD 213885c (TOI-141c), whose minimum mass of $19.95\pm 1.4$ $M_\oplus$ makes it consistent with being a Neptune-mass exoplanet. The HD 213885 system is very interesting from the perspective of future atmospheric characterization, being the second brightest star to host an ultra-short period transiting super-Earth (with the brightest star being, in fact, 55 Cancri). Prospects for characterization with present and future observatories are discussed. △ Less

Submitted 4 November, 2019; v1 submitted 18 March, 2019; originally announced March 2019.

Comments: MNRAS accepted version

MSC Class: 85-XX

The Elusive Majority of Young Moving Groups. I. Young Binaries and Lithium-Rich Stars in the Solar Neighborhood

Authors: Brendan P. Bowler, Sasha Hinkley, Carl Ziegler, Christoph Baranec, John E. Gizis, Nicholas M. Law, Michael C. Liu, Viyang S. Shah, Evgenya L. Shkolnik, Basmah Riaz, Reed Riddle

Abstract: Young stars in the solar neighborhood serve as nearby probes of stellar evolution and represent promising targets to directly image self-luminous giant planets. We have carried out an all-sky search for late-type ($\approx$K7$-$M5) stars within 100 pc selected primarily on the basis of activity indicators from $GALEX$ and $ROSAT$. Approximately two thousand active and potentially young stars are i… ▽ More Young stars in the solar neighborhood serve as nearby probes of stellar evolution and represent promising targets to directly image self-luminous giant planets. We have carried out an all-sky search for late-type ($\approx$K7$-$M5) stars within 100 pc selected primarily on the basis of activity indicators from $GALEX$ and $ROSAT$. Approximately two thousand active and potentially young stars are identified, over 600 of which we have followed up with low-resolution optical spectroscopy and over 1000 with diffraction-limited imaging using Robo-AO at the Palomar 1.5-m telescope. Strong lithium is present in 58 stars, implying ages spanning $\approx$10$-$200 Myr. Most of these lithium-rich stars are new or previously known members of young moving groups including TWA, $β$ Pic, Tuc-Hor, Carina, Columba, Argus, AB Dor, Upper Centaurus Lupus, and Lower Centaurus Crux; the rest appear to be young low-mass stars without connections to established kinematic groups. Over 200 close binaries are identified down to 0.2$''$ $-$ the vast majority of which are new $-$ and will be valuable for dynamical mass measurements of young stars with continued orbit monitoring in the future. △ Less

Submitted 14 March, 2019; originally announced March 2019.

Comments: Accepted to ApJ

Inference Without Compatibility

Authors: Michael Law, Ya'acov Ritov

Abstract: We consider hypotheses testing problems for three parameters in high-dimensional linear models with minimal sparsity assumptions of their type but without any compatibility conditions. Under this framework, we construct the first $\sqrt{n}$-consistent estimators for low-dimensional coefficients, the signal strength, and the noise level. We support our results using numerical simulations and provid… ▽ More We consider hypotheses testing problems for three parameters in high-dimensional linear models with minimal sparsity assumptions of their type but without any compatibility conditions. Under this framework, we construct the first $\sqrt{n}$-consistent estimators for low-dimensional coefficients, the signal strength, and the noise level. We support our results using numerical simulations and provide comparisons with other estimators. △ Less

Submitted 21 January, 2020; v1 submitted 14 March, 2019; originally announced March 2019.

Centroid-based deep metric learning for speaker recognition

Authors: Jixuan Wang, Kuan-Chieh Wang, Marc Law, Frank Rudzicz, Michael Brudno

Abstract: Speaker embedding models that utilize neural networks to map utterances to a space where distances reflect similarity between speakers have driven recent progress in the speaker recognition task. However, there is still a significant performance gap between recognizing speakers in the training set and unseen speakers. The latter case corresponds to the few-shot learning task, where a trained model… ▽ More Speaker embedding models that utilize neural networks to map utterances to a space where distances reflect similarity between speakers have driven recent progress in the speaker recognition task. However, there is still a significant performance gap between recognizing speakers in the training set and unseen speakers. The latter case corresponds to the few-shot learning task, where a trained model is evaluated on unseen classes. Here, we optimize a speaker embedding model with prototypical network loss (PNL), a state-of-the-art approach for the few-shot image classification task. The resulting embedding model outperforms the state-of-the-art triplet loss based models in both speaker verification and identification tasks, for both seen and unseen speakers. △ Less

Submitted 6 February, 2019; originally announced February 2019.

Comments: ICASSP 2019 (44th International Conference on Acoustics, Speech, and Signal Processing)

An Eccentric Massive Jupiter Orbiting a Sub-Giant on a 9.5 Day Period Discovered in the Transiting Exoplanet Survey Satellite Full Frame Images

Authors: Joseph E. Rodriguez, Samuel N. Quinn, Chelsea X. Huang, Andrew Vanderburg, Kaloyan Penev, Rafael Brahm, Andrés Jordán, Mma Ikwut-Ukwa, Shelly Tsirulik, David W. Latham, Keivan G. Stassun, Avi Shporer, Carl Ziegler, Elisabeth Matthews, Jason D. Eastman, B. Scott Gaudi, Karen A. Collins, Natalia Guerrero, Howard M. Relles, Thomas Barclay, Natalie M. Batalha, Perry Berlind, Allyson Bieryla, L. G. Bouma, Patricia T Boyd , et al. (49 additional authors not shown)

Abstract: We report the discovery of TOI-172 b from the Transiting Exoplanet Survey Satellite (TESS) mission, a massive hot Jupiter transiting a slightly evolved G-star with a 9.48-day orbital period. This is the first planet to be confirmed from analysis of only the TESS full frame images, because the host star was not chosen as a two minute cadence target. From a global analysis of the TESS photometry and… ▽ More We report the discovery of TOI-172 b from the Transiting Exoplanet Survey Satellite (TESS) mission, a massive hot Jupiter transiting a slightly evolved G-star with a 9.48-day orbital period. This is the first planet to be confirmed from analysis of only the TESS full frame images, because the host star was not chosen as a two minute cadence target. From a global analysis of the TESS photometry and follow-up observations carried out by the TESS Follow-up Observing Program Working Group, TOI-172 (TIC 29857954) is a slightly evolved star with an effective temperature of $T_{\rm eff}$ =$5645\pm50$ K, a mass of $M_{\star}$ = $1.128^{+0.065}_{-0.061}$ $M_{\odot}$, radius of $R_{\star}$ = $1.777^{+0.047}_{-0.044}$ $R_{\odot}$, a surface gravity of $\log$ $g_{\star}$ = $3.993^{+0.027}_{-0.028}$, and an age of $7.4^{+1.6}_{-1.5}$ Gyr. Its planetary companion (TOI-172 b) has a radius of $R_{\rm P}$ = $0.965^{+0.032}_{-0.029}$ $R_{\rm J}$, a mass of $M_{\rm P}$ = $5.42^{+0.22}_{-0.20}$ $M_{\rm J}$, and is on an eccentric orbit ($e = 0.3806^{+0.0093}_{-0.0090}$). TOI-172 b is one of the few known massive giant planets on a highly eccentric short-period orbit. Future study of the atmosphere of this planet and its system architecture offer opportunities to understand the formation and evolution of similar systems. △ Less

Submitted 2 July, 2019; v1 submitted 28 January, 2019; originally announced January 2019.

Comments: 14 pages, 8 figures, 4 tables, Published in AJ

Journal ref: The Astronomical Journal, 2019, Volume 157, Issue 5, 191, 13

Bright Opportunities for Atmospheric Characterization of Small Planets: Masses and Radii of K2-3 b, c, d and GJ3470 b from Radial Velocity Measurements and Spitzer Transits

Authors: Molly R. Kosiarek, Ian J. M. Crossfield, Kevin K. Hardegree-Ullman, John H. Livingston, Bjorn Benneke, Sarah Blunt, Gregory W. Henry, Ward S. Howard, David Berardo, Benjamin J. Fulton, Lea A. Hirsch, Andrew W. Howard, Howard Isaacson, Erik A. Petigura, Evan Sinukoff, Lauren Weiss, X. Bonfils, Courtney D. Dressing, Heather A. Knutson, Joshua E. Schlieder, Michael Werner, Varoujan Gorjian, Jessica Krick, Farisa Y. Morales, Nicola Astudillo-Defru , et al. (14 additional authors not shown)

Abstract: We report improved masses, radii, and densities for four planets in two bright M-dwarf systems, K2-3 and GJ3470, derived from a combination of new radial velocity and transit observations. Supplementing K2 photometry with follow-up Spitzer transit observations refined the transit ephemerides of K2-3 b, c, and d by over a factor of 10. We analyze ground-based photometry from the Evryscope and Fairb… ▽ More We report improved masses, radii, and densities for four planets in two bright M-dwarf systems, K2-3 and GJ3470, derived from a combination of new radial velocity and transit observations. Supplementing K2 photometry with follow-up Spitzer transit observations refined the transit ephemerides of K2-3 b, c, and d by over a factor of 10. We analyze ground-based photometry from the Evryscope and Fairborn Observatory to determine the characteristic stellar activity timescales for our Gaussian Process fit, including the stellar rotation period and activity region decay timescale. The stellar rotation signals for both stars are evident in the radial velocity data and are included in our fit using a Gaussian process trained on the photometry. We find the masses of K2-3 b, K2-3 c and GJ3470 b to be 6.48$^{+0.99}_{-0.93}$, 2.14$^{+1.08}_{-1.04}$, and 12.58$^{+1.31}_{-1.28}$ M$_\oplus$ respectively. K2-3 d was not significantly detected and has a 3-$σ$ upper limit of 2.80 M$_\oplus$. These two systems are training cases for future TESS systems; due to the low planet densities ($ρ$ $<$ 3.7 g cm$^{-3}$) and bright host stars (K $<$ 9 mag), they are among the best candidates for transmission spectroscopy in order to characterize the atmospheric compositions of small planets. △ Less

Submitted 14 February, 2019; v1 submitted 19 December, 2018; originally announced December 2018.

Comments: 21 pages, 11 figures, accepted to AJ

Journal ref: AJ 157, 97 (2019)

HD2685 b: A Hot-Jupiter orbiting an early F-type star detected by TESS

Authors: M. I. Jones, R. Brahm, N. Espinoza, S. Wang, A. Shporer, T. Henning, A. Jordan, P. Sarkis, L. A. Paredes, J. Hodari-Sadiki, T. Henry, B. Cruz, L. D. Nielsen, F. Bouchy, F. Pepe, D. Segransan, O. Turner, S. Udry, G. Bakos, D. Osip, V. Suc, C. Ziegler, A. Tokovinin, N. M. Law, A. W. Mann , et al. (18 additional authors not shown)

Abstract: We report on the confirmation of a transiting giant planet around the relatively hot (Teff = 6801 $\pm$ 56 K) star HD2685, whose transit signal was detected in Sector 1 data of the TESS mission. We confirmed the planetary nature of the transit signal by using Doppler velocimetric measurements with CHIRON, CORALIE and FEROS, as well as photometric data with CHAT and LCOGT. From the photometry and r… ▽ More We report on the confirmation of a transiting giant planet around the relatively hot (Teff = 6801 $\pm$ 56 K) star HD2685, whose transit signal was detected in Sector 1 data of the TESS mission. We confirmed the planetary nature of the transit signal by using Doppler velocimetric measurements with CHIRON, CORALIE and FEROS, as well as photometric data with CHAT and LCOGT. From the photometry and radial velocities joint analysis, we derived the following parameters for HD2685 $b$: $P$=4.12692$\pm$0.00004 days, M$_P$=1.18 $\pm$ 0.09 $M_J$ and $R_P$=1.44 $\pm$ 0.01 $R_J$. This system is a typical example of an inflated transiting Hot-Jupiter in a circular orbit. Given the host star apparent visual magnitude ($V$ = 9.6 mag), this is one of the brightest known stars hosting a transiting Hot-Jupiter, and a good example of the upcoming systems that will be detected by TESS during the two-year primary mission. This is also an excellent target for future ground and space based atmospheric characterization as well as a good candidate for measuring the projected spin-orbit misalignment angle via the Rossiter-McLaughlin effect. △ Less

Submitted 13 November, 2018; originally announced November 2018.

Comments: Submitted for publication to A&A