Showing 1–50 of 168 results for author: Aggarwal, K

Polaris: A Safety-focused LLM Constellation Architecture for Healthcare

Authors: Subhabrata Mukherjee, Paul Gamble, Markel Sanz Ausin, Neel Kant, Kriti Aggarwal, Neha Manjunath, Debajyoti Datta, Zhengliang Liu, Jiayuan Ding, Sophia Busacca, Cezanne Bianco, Swapnil Sharma, Rae Lasko, Michelle Voisard, Sanchay Harneja, Darya Filippova, Gerry Meixiong, Kevin Cha, Amir Youssefi, Meyhaa Buvanesh, Howard Weingram, Sebastian Bierman-Lytle, Harpreet Singh Mangat, Kim Parikh, Saad Godil , et al. (1 additional authors not shown)

Abstract: We develop Polaris, the first safety-focused LLM constellation for real-time patient-AI healthcare conversations. Unlike prior LLM works in healthcare focusing on tasks like question answering, our work specifically focuses on long multi-turn voice conversations. Our one-trillion parameter constellation system is composed of several multibillion parameter LLMs as co-operative agents: a stateful pr… ▽ More We develop Polaris, the first safety-focused LLM constellation for real-time patient-AI healthcare conversations. Unlike prior LLM works in healthcare focusing on tasks like question answering, our work specifically focuses on long multi-turn voice conversations. Our one-trillion parameter constellation system is composed of several multibillion parameter LLMs as co-operative agents: a stateful primary agent that focuses on driving an engaging conversation and several specialist support agents focused on healthcare tasks performed by nurses to increase safety and reduce hallucinations. We develop a sophisticated training protocol for iterative co-training of the agents that optimize for diverse objectives. We train our models on proprietary data, clinical care plans, healthcare regulatory documents, medical manuals, and other medical reasoning documents. We align our models to speak like medical professionals, using organic healthcare conversations and simulated ones between patient actors and experienced nurses. This allows our system to express unique capabilities such as rapport building, trust building, empathy and bedside manner. Finally, we present the first comprehensive clinician evaluation of an LLM system for healthcare. We recruited over 1100 U.S. licensed nurses and over 130 U.S. licensed physicians to perform end-to-end conversational evaluations of our system by posing as patients and rating the system on several measures. We demonstrate Polaris performs on par with human nurses on aggregate across dimensions such as medical safety, clinical readiness, conversational quality, and bedside manner. Additionally, we conduct a challenging task-based evaluation of the individual specialist support agents, where we demonstrate our LLM agents significantly outperform a much larger general-purpose LLM (GPT-4) as well as from its own medium-size class (LLaMA-2 70B). △ Less

Submitted 20 March, 2024; originally announced March 2024.

Computational Study on the Impact of Gasoline-Ethanol Blending on Autoignition and Soot/NOx Emissions under Gasoline Compression Ignition Conditions

Authors: Krishna C. Kalvakala, Harsimran Singh, Pinaki Pal, Jorge P. Gonzalez, Christopher P. Kolodziej, Suresh K. Aggarwal

Abstract: Computational fluid dynamics (CFD) simulations of a single-cylinder gasoline compression ignition engine are performed to investigate the impact of gasoline-ethanol blending on autoignition, nitrogen oxide (NOx), and soot emissions under low-load conditions. A four-component toluene primary reference fuel (TPRF) + ethanol (ETPRF) surrogate (with 10% ethanol by volume; E10) is employed to represent… ▽ More Computational fluid dynamics (CFD) simulations of a single-cylinder gasoline compression ignition engine are performed to investigate the impact of gasoline-ethanol blending on autoignition, nitrogen oxide (NOx), and soot emissions under low-load conditions. A four-component toluene primary reference fuel (TPRF) + ethanol (ETPRF) surrogate (with 10% ethanol by volume; E10) is employed to represent the test gasoline (RD5-87). A 3D engine CFD model employing finite-rate chemistry with a skeletal kinetic mechanism, adaptive mesh refinement (AMR), and hybrid method of moments (HMOM) is adopted to capture in-cylinder combustion and soot/NOx emissions. The engine CFD model is validated against experimental data for three gasoline-ethanol blends: E10, E30 and E100, with varying ethanol content by volume. Model validation is carried out for multiple start-of-injection (SOI) timings (-21, -27, -36, and -45 crank angle degrees after top-dead-center (aTDC)) with respect to in-cylinder pressure, heat release rate, combustion phasing, NOx and soot emissions. For late injection timings (-21 and -27oaTDC), E30 yields higher soot than E10; while the trend reverses for early injection cases (-36 and -45oaTDC). E100 yields the lowest amount of soot among all fuels irrespective of SOI timing. Further, E10 shows a non-monotonic trend in soot emissions with SOI timing: SOI-36>SOI-45>SOI-21>SOI-27, while soot emissions from E30 exhibit monotonic decrease with advancing SOI timing. NOx emissions from various fuels follow a trend of E10>E30>E100. NOx emissions increase as SOI timing is advanced for all fuels, with an anomaly for E10 and E100 where NOx decreases when SOI is advanced beyond -36oaTDC. Detailed analysis of the numerical results is performed to investigate the emission trends and elucidate the impact of chemical composition and physical properties on autoignition and emissions characteristics. △ Less

Submitted 15 March, 2024; originally announced March 2024.

Coupling a single spin to high-frequency motion

Authors: Federico Fedele, Federico Cerisola, Lea Bresque, Florian Vigneau, Juliette Monsel, Jorge Tabanera, Kushagra Aggarwal, Jonathan Dexter, Sofia Sevitz, Joe Dunlop, Alexia Auffèves, Juan Parrondo, András Pályi, Janet Anders, Natalia Ares

Abstract: Coupling a single spin to high-frequency mechanical motion is a fundamental bottleneck of applications such as quantum sensing, intermediate and long-distance spin-spin coupling, and classical and quantum information processing. Previous experiments have only shown single spin coupling to low-frequency mechanical resonators, such as diamond cantilevers. High-frequency mechanical resonators, having… ▽ More Coupling a single spin to high-frequency mechanical motion is a fundamental bottleneck of applications such as quantum sensing, intermediate and long-distance spin-spin coupling, and classical and quantum information processing. Previous experiments have only shown single spin coupling to low-frequency mechanical resonators, such as diamond cantilevers. High-frequency mechanical resonators, having the ability to access the quantum regime, open a range of possibilities when coupled to single spins, including readout and storage of quantum states. Here we report the first experimental demonstration of spin-mechanical coupling to a high-frequency resonator. We achieve this all-electrically on a fully suspended carbon nanotube device. A new mechanism gives rise to this coupling, which stems from spin-orbit coupling, and it is not mediated by strain. We observe both resonant and off-resonant coupling as a shift and broadening of the electric dipole spin resonance (EDSR), respectively. We develop a complete theoretical model taking into account the tensor form of the coupling and non-linearity in the motion. Our results propel spin-mechanical platforms to an uncharted regime. The interaction we reveal provides the full toolbox for promising applications ranging from the demonstration of macroscopic superpositions, to the operation of fully quantum engines, to quantum simulators. △ Less

Submitted 29 February, 2024; originally announced February 2024.

Comments: Comments welcome

Nightside clouds and disequilibrium chemistry on the hot Jupiter WASP-43b

Authors: Taylor J. Bell, Nicolas Crouzet, Patricio E. Cubillos, Laura Kreidberg, Anjali A. A. Piette, Michael T. Roman, Joanna K. Barstow, Jasmina Blecic, Ludmila Carone, Louis-Philippe Coulombe, Elsa Ducrot, Mark Hammond, João M. Mendonça, Julianne I. Moses, Vivien Parmentier, Kevin B. Stevenson, Lucas Teinturier, Michael Zhang, Natalie M. Batalha, Jacob L. Bean, Björn Benneke, Benjamin Charnay, Katy L. Chubb, Brice-Olivier Demory, Peter Gao , et al. (58 additional authors not shown)

Abstract: Hot Jupiters are among the best-studied exoplanets, but it is still poorly understood how their chemical composition and cloud properties vary with longitude. Theoretical models predict that clouds may condense on the nightside and that molecular abundances can be driven out of equilibrium by zonal winds. Here we report a phase-resolved emission spectrum of the hot Jupiter WASP-43b measured from 5… ▽ More Hot Jupiters are among the best-studied exoplanets, but it is still poorly understood how their chemical composition and cloud properties vary with longitude. Theoretical models predict that clouds may condense on the nightside and that molecular abundances can be driven out of equilibrium by zonal winds. Here we report a phase-resolved emission spectrum of the hot Jupiter WASP-43b measured from 5-12 $μ$m with JWST's Mid-Infrared Instrument (MIRI). The spectra reveal a large day-night temperature contrast (with average brightness temperatures of 1524$\pm$35 and 863$\pm$23 Kelvin, respectively) and evidence for water absorption at all orbital phases. Comparisons with three-dimensional atmospheric models show that both the phase curve shape and emission spectra strongly suggest the presence of nightside clouds which become optically thick to thermal emission at pressures greater than ~100 mbar. The dayside is consistent with a cloudless atmosphere above the mid-infrared photosphere. Contrary to expectations from equilibrium chemistry but consistent with disequilibrium kinetics models, methane is not detected on the nightside (2$σ$ upper limit of 1-6 parts per million, depending on model assumptions). △ Less

Submitted 23 January, 2024; originally announced January 2024.

Comments: 61 pages, 13 figures, 4 tables. This preprint has been submitted to and accepted in principle for publication in Nature Astronomy without significant changes

An unidentified Fermi source emitting radio bursts in the Galactic bulge

Authors: Reshma Anna-Thomas, Sarah Burke-Spolaor, Casey J. Law, F. K. Schinzel, Kshitij Aggarwal, Geoffrey C. Bower, Liam Connor, Paul B. Demorest

Abstract: We report on the detection of radio bursts from the Galactic bulge using the real-time transient detection and localization system, realfast. The pulses were detected commensally on the Karl G. Jansky Very Large Array during a survey of unidentified Fermi $γ$-ray sources. The bursts were localized to subarcsecond precision using realfast fast-sampled imaging. Follow-up observations with the Green… ▽ More We report on the detection of radio bursts from the Galactic bulge using the real-time transient detection and localization system, realfast. The pulses were detected commensally on the Karl G. Jansky Very Large Array during a survey of unidentified Fermi $γ$-ray sources. The bursts were localized to subarcsecond precision using realfast fast-sampled imaging. Follow-up observations with the Green Bank Telescope detected additional bursts from the same source. The bursts do not exhibit periodicity in a search up to periods of 480s, assuming a duty cycle of < 20%. The pulses are nearly 100% linearly polarized, show circular polarization up to 12%, have a steep radio spectral index of -2.7, and exhibit variable scattering on timescales of months. The arcsecond-level realfast localization links the source confidently with the Fermi $γ$-ray source and places it nearby (though not coincident with) an XMM-Newton X-ray source. Based on the source's overall properties, we discuss various options for the nature of this object and propose that it could be a young pulsar, magnetar, or a binary pulsar system. △ Less

Submitted 4 January, 2024; originally announced January 2024.

Comments: Submitted to The Astrophysical Journal

ODIN: A Single Model for 2D and 3D Segmentation

Authors: Ayush Jain, Pushkal Katara, Nikolaos Gkanatsios, Adam W. Harley, Gabriel Sarch, Kriti Aggarwal, Vishrav Chaudhary, Katerina Fragkiadaki

Abstract: State-of-the-art models on contemporary 3D segmentation benchmarks like ScanNet consume and label dataset-provided 3D point clouds, obtained through post processing of sensed multiview RGB-D images. They are typically trained in-domain, forego large-scale 2D pre-training and outperform alternatives that featurize the posed RGB-D multiview images instead. The gap in performance between methods that… ▽ More State-of-the-art models on contemporary 3D segmentation benchmarks like ScanNet consume and label dataset-provided 3D point clouds, obtained through post processing of sensed multiview RGB-D images. They are typically trained in-domain, forego large-scale 2D pre-training and outperform alternatives that featurize the posed RGB-D multiview images instead. The gap in performance between methods that consume posed images versus post-processed 3D point clouds has fueled the belief that 2D and 3D perception require distinct model architectures. In this paper, we challenge this view and propose ODIN (Omni-Dimensional INstance segmentation), a model that can segment and label both 2D RGB images and 3D point clouds, using a transformer architecture that alternates between 2D within-view and 3D cross-view information fusion. Our model differentiates 2D and 3D feature operations through the positional encodings of the tokens involved, which capture pixel coordinates for 2D patch tokens and 3D coordinates for 3D feature tokens. ODIN achieves state-of-the-art performance on ScanNet200, Matterport3D and AI2THOR 3D instance segmentation benchmarks, and competitive performance on ScanNet, S3DIS and COCO. It outperforms all previous works by a wide margin when the sensed 3D point cloud is used in place of the point cloud sampled from 3D mesh. When used as the 3D perception engine in an instructable embodied agent architecture, it sets a new state-of-the-art on the TEACh action-from-dialogue benchmark. Our code and checkpoints can be found at the project website (https://odin-seg.github.io). △ Less

Submitted 25 June, 2024; v1 submitted 4 January, 2024; originally announced January 2024.

Comments: Camera Ready (CVPR 2024, Highlight)

Orca 2: Teaching Small Language Models How to Reason

Authors: Arindam Mitra, Luciano Del Corro, Shweti Mahajan, Andres Codas, Clarisse Simoes, Sahaj Agarwal, Xuxi Chen, Anastasia Razdaibiedina, Erik Jones, Kriti Aggarwal, Hamid Palangi, Guoqing Zheng, Corby Rosset, Hamed Khanpour, Ahmed Awadallah

Abstract: Orca 1 learns from rich signals, such as explanation traces, allowing it to outperform conventional instruction-tuned models on benchmarks like BigBench Hard and AGIEval. In Orca 2, we continue exploring how improved training signals can enhance smaller LMs' reasoning abilities. Research on training small LMs has often relied on imitation learning to replicate the output of more capable models. We… ▽ More Orca 1 learns from rich signals, such as explanation traces, allowing it to outperform conventional instruction-tuned models on benchmarks like BigBench Hard and AGIEval. In Orca 2, we continue exploring how improved training signals can enhance smaller LMs' reasoning abilities. Research on training small LMs has often relied on imitation learning to replicate the output of more capable models. We contend that excessive emphasis on imitation may restrict the potential of smaller models. We seek to teach small LMs to employ different solution strategies for different tasks, potentially different from the one used by the larger model. For example, while larger models might provide a direct answer to a complex task, smaller models may not have the same capacity. In Orca 2, we teach the model various reasoning techniques (step-by-step, recall then generate, recall-reason-generate, direct answer, etc.). More crucially, we aim to help the model learn to determine the most effective solution strategy for each task. We evaluate Orca 2 using a comprehensive set of 15 diverse benchmarks (corresponding to approximately 100 tasks and over 36,000 unique prompts). Orca 2 significantly surpasses models of similar size and attains performance levels similar or better to those of models 5-10x larger, as assessed on complex tasks that test advanced reasoning abilities in zero-shot settings. make Orca 2 weights publicly available at aka.ms/orca-lm to support research on the development, evaluation, and alignment of smaller LMs △ Less

Submitted 21 November, 2023; v1 submitted 18 November, 2023; originally announced November 2023.

Comments: Added url to model weights fixed typo in Author name

Efficient Continual Pre-training for Building Domain Specific Large Language Models

Authors: Yong Xie, Karan Aggarwal, Aitzaz Ahmad

Abstract: Large language models (LLMs) have demonstrated remarkable open-domain capabilities. Traditionally, LLMs tailored for a domain are trained from scratch to excel at handling domain-specific tasks. In this work, we explore an alternative strategy of continual pre-training as a means to develop domain-specific LLMs. We introduce FinPythia-6.9B, developed through domain-adaptive continual pre-training… ▽ More Large language models (LLMs) have demonstrated remarkable open-domain capabilities. Traditionally, LLMs tailored for a domain are trained from scratch to excel at handling domain-specific tasks. In this work, we explore an alternative strategy of continual pre-training as a means to develop domain-specific LLMs. We introduce FinPythia-6.9B, developed through domain-adaptive continual pre-training on the financial domain. Continual pre-trained FinPythia showcases consistent improvements on financial tasks over the original foundational model. We further explore simple but effective data selection strategies for continual pre-training. Our data selection strategies outperforms vanilla continual pre-training's performance with just 10% of corpus size and cost, without any degradation on open-domain standard tasks. Our work proposes an alternative solution to building domain-specific LLMs from scratch in a cost-effective manner. △ Less

Submitted 14 November, 2023; originally announced November 2023.

The Petabyte Project

Authors: Evan F. Lewis, Sarah Burke-Spolaor, Maura McLaughlin, Duncan Lorimer, Kshitij Aggarwal, Devansh Agarwal, Joseph Kania, Nate Garver-Daniels, Joseph P. Glaser

Abstract: Transient radio sources, such as fast radio bursts, intermittent pulsars, and rotating radio transients, can offer a wealth of information regarding extreme emission physics as well as the intervening interstellar and/or intergalactic medium. Vital steps towards understanding these objects include characterizing their source populations and estimating their event rates across observing frequencies… ▽ More Transient radio sources, such as fast radio bursts, intermittent pulsars, and rotating radio transients, can offer a wealth of information regarding extreme emission physics as well as the intervening interstellar and/or intergalactic medium. Vital steps towards understanding these objects include characterizing their source populations and estimating their event rates across observing frequencies. However, previous efforts have been undertaken mostly by individual survey teams at disparate observing frequencies and telescopes, and with non-uniform algorithms for searching and characterization. The Petabyte Project (TPP) aims to address these issues by uniformly reprocessing data from several petabytes of radio transient surveys covering two decades of observing frequency (300 MHz-20 GHz). The TPP will provide robust event rate analyses, in-depth assessment of survey and pipeline completeness, as well as revealing discoveries from archival and ongoing radio surveys. We present an overview of TPP's processing pipeline, scope, and our potential to make new discoveries. △ Less

Submitted 23 August, 2023; originally announced August 2023.

Comments: 5 pages, 1 figure. Accepted for publication in Proceedings for IAU Symposium 369: The Dawn of Cosmology & Multi-Messenger Studies with Fast Radio Bursts

Temporal and Spectral Properties of the Persistent Radio Source Associated with FRB 20190520B with the VLA

Authors: Xian Zhang, Wenfei Yu, Casey Law, Di Li, Shami Chatterjee, Paul Demorest, Zhen Yan, Chenhui Niu, Kshitij Aggarwal, Reshma Anna-Thomas, Sarah Burke-Spolaor, Liam Connor, Chao-wei Tsai, Weiwei Zhu, Gan Luo

Abstract: Among more than 800 known fast radio bursts (FRBs), only two, namely FRB 20121102A and FRB 20190520B, are confirmed to be associated with a persistent radio sources (PRS). Here we report evidence of apparent temporal variability in the PRS associated with the bursting FRB 20190520B based on the Karl G. Jansky Very Large Array (VLA) observations taken in 2020 and 2021. Based on the analysis of epoc… ▽ More Among more than 800 known fast radio bursts (FRBs), only two, namely FRB 20121102A and FRB 20190520B, are confirmed to be associated with a persistent radio sources (PRS). Here we report evidence of apparent temporal variability in the PRS associated with the bursting FRB 20190520B based on the Karl G. Jansky Very Large Array (VLA) observations taken in 2020 and 2021. Based on the analysis of epoch-to-epoch variability of the PRS at L, S, C, and X band in 1-12 GHz, we detected not only overall marginal variability but also a likely radio flux decrease ($\sim$ 3.2 $σ$) between the observations taken in 2020 and 2021 at 3 GHz. Assuming no spectral variation in the PRS during these observations, we found the evidence for an overall broadband radio flux decrease by about 20 percent between the 2020 and the 2021 observations, suggesting that the PRS probably evolves on the yearly time scale. If we attribute the marginal variability at 3 GHz as intrinsic or due to scintillation, the size of potential variable component of the PRS is constrained to be sub-parsec. On the other hand, the size of the PRS can be also constrained to be larger than about 0.22 parsec from the averaged radio spectrum and the integrated radio luminosity in the 1-12 GHz band based on equipartition and self-absorption arguments. We discuss potential origins of the PRS and suggest that an accreting compact object origin might be able to explain the PRS's temporal and spectral properties. Confirmation of variability or flux decline of the PRS would be critical to our understanding of the PRS and its relation to the bursting source. △ Less

Submitted 23 October, 2023; v1 submitted 30 July, 2023; originally announced July 2023.

Comments: 12 pages, 3 figures, accepted for publication in ApJ

Parity-conserving Cooper-pair transport and ideal superconducting diode in planar Germanium

Authors: Marco Valentini, Oliver Sagi, Levon Baghumyan, Thijs de Gijsel, Jason Jung, Stefano Calcaterra, Andrea Ballabio, Juan Aguilera Servin, Kushagra Aggarwal, Marian Janik, Thomas Adletzberger, Rubén Seoane Souto, Martin Leijnse, Jeroen Danon, Constantin Schrade, Erik Bakkers, Daniel Chrastina, Giovanni Isella, Georgios Katsaros

Abstract: Superconductor/semiconductor hybrid devices have attracted increasing interest in the past years. Superconducting electronics aims to complement semiconductor technology, while hybrid architectures are at the forefront of new ideas such as topological superconductivity and protected qubits. In this work, we engineer the induced superconductivity in two-dimensional germanium hole gas by varying the… ▽ More Superconductor/semiconductor hybrid devices have attracted increasing interest in the past years. Superconducting electronics aims to complement semiconductor technology, while hybrid architectures are at the forefront of new ideas such as topological superconductivity and protected qubits. In this work, we engineer the induced superconductivity in two-dimensional germanium hole gas by varying the distance between the quantum well and the aluminum. We demonstrate a hard superconducting gap and realize an electrically and flux tunable superconducting diode using a superconducting quantum interference device (SQUID). This allows to tune the current phase relation (CPR), to a regime where single Cooper pair tunneling is suppressed, creating a $\sin \left( 2 \varphi \right)$ CPR. Shapiro experiments complement this interpretation and the microwave drive allows to create a diode with 100% efficiency. The reported results open up the path towards integration of spin qubit devices, microwave resonators and (protected) superconducting qubits on a silicon technology compatible platform. △ Less

Submitted 16 November, 2023; v1 submitted 12 June, 2023; originally announced June 2023.

DUBLIN -- Document Understanding By Language-Image Network

Authors: Kriti Aggarwal, Aditi Khandelwal, Kumar Tanmay, Owais Mohammed Khan, Qiang Liu, Monojit Choudhury, Hardik Hansrajbhai Chauhan, Subhojit Som, Vishrav Chaudhary, Saurabh Tiwary

Abstract: Visual document understanding is a complex task that involves analyzing both the text and the visual elements in document images. Existing models often rely on manual feature engineering or domain-specific pipelines, which limit their generalization ability across different document types and languages. In this paper, we propose DUBLIN, which is pretrained on web pages using three novel objectives… ▽ More Visual document understanding is a complex task that involves analyzing both the text and the visual elements in document images. Existing models often rely on manual feature engineering or domain-specific pipelines, which limit their generalization ability across different document types and languages. In this paper, we propose DUBLIN, which is pretrained on web pages using three novel objectives: Masked Document Text Generation Task, Bounding Box Task, and Rendered Question Answering Task, that leverage both the spatial and semantic information in the document images. Our model achieves competitive or state-of-the-art results on several benchmarks, such as Web-Based Structural Reading Comprehension, Document Visual Question Answering, Key Information Extraction, Diagram Understanding, and Table Question Answering. In particular, we show that DUBLIN is the first pixel-based model to achieve an EM of 77.75 and F1 of 84.25 on the WebSRC dataset. We also show that our model outperforms the current pixel-based SOTA models on DocVQA, InfographicsVQA, OCR-VQA and AI2D datasets by 4.6%, 6.5%, 2.6% and 21%, respectively. We also achieve competitive performance on RVL-CDIP document classification. Moreover, we create new baselines for text-based datasets by rendering them as document images to promote research in this direction. △ Less

Submitted 27 October, 2023; v1 submitted 23 May, 2023; originally announced May 2023.

ACM Class: F.2.2; I.2.7

Comment on "Atomic structure and electron impact excitation of Al-like ions (Ga--Br)" by HB Wang and G Jiang in At. Data Nucl. Data Tables 148 (2022) 101532

Authors: K. M. Aggarwal, K. W. Smith

Abstract: In a recent paper, Wang and Jiang (At. Data Nucl. Data Tables 148 (2022) 101532) have reported data for energy levels, radiative rates (A-values), and effective collision strengths ($Υ$) for some transitions of five Al-like ions, namely Ga~XIX, Ge~XX, As~XXI, Se~XXII, and Br~XXIII. On a closer examination we find that their reported data for energy levels and A-values are generally correct, but no… ▽ More In a recent paper, Wang and Jiang (At. Data Nucl. Data Tables 148 (2022) 101532) have reported data for energy levels, radiative rates (A-values), and effective collision strengths ($Υ$) for some transitions of five Al-like ions, namely Ga~XIX, Ge~XX, As~XXI, Se~XXII, and Br~XXIII. On a closer examination we find that their reported data for energy levels and A-values are generally correct, but not for $Υ$. Their $Υ$ values, for all transitions (allowed or forbidden) and for all ions, invariably decrease at higher temperatures. This is mainly because they have adopted a limited range of electron energies for the calculations of collision strengths. We demonstrate this with our calculations with the Flexible Atomic Code (FAC), and conclude that their $Υ$ values are inaccurate, unreliable, and should not be adopted in any applications or modelling analysis. △ Less

Submitted 19 May, 2023; originally announced May 2023.

Comments: The paper will be published in ADNDT which has 23 pages including 2 Tables and 5 Figures (a--e)

FIREBALL: A Dataset of Dungeons and Dragons Actual-Play with Structured Game State Information

Authors: Andrew Zhu, Karmanya Aggarwal, Alexander Feng, Lara J. Martin, Chris Callison-Burch

Abstract: Dungeons & Dragons (D&D) is a tabletop roleplaying game with complex natural language interactions between players and hidden state information. Recent work has shown that large language models (LLMs) that have access to state information can generate higher quality game turns than LLMs that use dialog history alone. However, previous work used game state information that was heuristically created… ▽ More Dungeons & Dragons (D&D) is a tabletop roleplaying game with complex natural language interactions between players and hidden state information. Recent work has shown that large language models (LLMs) that have access to state information can generate higher quality game turns than LLMs that use dialog history alone. However, previous work used game state information that was heuristically created and was not a true gold standard game state. We present FIREBALL, a large dataset containing nearly 25,000 unique sessions from real D&D gameplay on Discord with true game state info. We recorded game play sessions of players who used the Avrae bot, which was developed to aid people in playing D&D online, capturing language, game commands and underlying game state information. We demonstrate that FIREBALL can improve natural language generation (NLG) by using Avrae state information, improving both automated metrics and human judgments of quality. Additionally, we show that LLMs can generate executable Avrae commands, particularly after finetuning. △ Less

Submitted 25 May, 2023; v1 submitted 2 May, 2023; originally announced May 2023.

Comments: 21 pages, 2 figures. Accepted at ACL 2023

Journal ref: Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2023, pp. 4171-4193

Detection of carbon monoxide's 4.6 micron fundamental band structure in WASP-39b's atmosphere with JWST NIRSpec G395H

Authors: David Grant, Joshua D. Lothringer, Hannah R. Wakeford, Munazza K. Alam, Lili Alderson, Jacob L. Bean, Björn Benneke, Jean-Michel Désert, Tansu Daylan, Laura Flagg, Renyu Hu, Julie Inglis, James Kirk, Laura Kreidberg, Mercedes López-Morales, Luigi Mancini, Thomas Mikal-Evans, Karan Molaverdikhani, Enric Palle, Benjamin V. Rackham, Seth Redfield, Kevin B. Stevenson, Jeff Valenti, Nicole L. Wallack, Keshav Aggarwal , et al. (6 additional authors not shown)

Abstract: Carbon monoxide (CO) is predicted to be the dominant carbon-bearing molecule in giant planet atmospheres, and, along with water, is important for discerning the oxygen and therefore carbon-to-oxygen ratio of these planets. The fundamental absorption mode of CO has a broad double-branched structure composed of many individual absorption lines from 4.3 to 5.1 $\mathrmμ$m, which can now be spectrosco… ▽ More Carbon monoxide (CO) is predicted to be the dominant carbon-bearing molecule in giant planet atmospheres, and, along with water, is important for discerning the oxygen and therefore carbon-to-oxygen ratio of these planets. The fundamental absorption mode of CO has a broad double-branched structure composed of many individual absorption lines from 4.3 to 5.1 $\mathrmμ$m, which can now be spectroscopically measured with JWST. Here we present a technique for detecting the rotational sub-band structure of CO at medium resolution with the NIRSpec G395H instrument. We use a single transit observation of the hot Jupiter WASP-39b from the JWST Transiting Exoplanet Community Early Release Science (JTEC ERS) program at the native resolution of the instrument ($R \,{\sim} 2700$) to resolve the CO absorption structure. We robustly detect absorption by CO, with an increase in transit depth of 264 $\pm$ 68 ppm, in agreement with the predicted CO contribution from the best-fit model at low resolution. This detection confirms our theoretical expectations that CO is the dominant carbon-bearing molecule in WASP-39b's atmosphere, and further supports the conclusions of low C/O and super-solar metallicities presented in the JTEC ERS papers for WASP-39b. △ Less

Submitted 24 April, 2023; originally announced April 2023.

Comments: 11 pages, 5 figures, accepted for publication in ApJL

Repeatability of image quality in very low field MRI

Authors: Pavan Poojar Kunal Aggarwal, Marina Manso Jimeno, Sairam Geethanath

Abstract: We investigated the repeatability of image quality metrics such as SNR, image uniformity, and geometrical distortion at 0.05T over ten days and three sessions per day. The measurements included temperature, humidity, transmit frequency, off-resonance maps, and 3D turbo spin echo (TSE) images of an in vitro phantom. This resulted in a protocol with nine pulse sequences. We also acquired a 3T data s… ▽ More We investigated the repeatability of image quality metrics such as SNR, image uniformity, and geometrical distortion at 0.05T over ten days and three sessions per day. The measurements included temperature, humidity, transmit frequency, off-resonance maps, and 3D turbo spin echo (TSE) images of an in vitro phantom. This resulted in a protocol with nine pulse sequences. We also acquired a 3T data set for reference. The image quality metrics included computing SNR, image non-uniformity, and eccentricity (to assess geometrical distortion) to investigate the repeatability of 0.05T image quality. The image reconstruction included drift correction, k-space filtering, and off-resonance correction. We computed the coefficient of variation (CV) of the experimental parameters and the resulting image quality metrics to assess repeatability. The range of temperature measured during the study was within 1.50C. The off-resonance maps acquired before and after the 3D TSE showed similar hotspots and changed mainly by a global constant. The SNR measurements were highly repeatable across sessions and over the ten days, quantified by a CV of 4.9%. The magnetic field inhomogeneity effects quantified by eccentricity showed a CV of 13.7% but less than 5.1% in two of the three sessions over ten days. The use of conjugate phase reconstruction mitigated geometrical distortion artifacts. The repeatability of image uniformity was moderate at 10.6%, with two of three sessions resulting in a CV of less than 7.8%. Temperature and humidity did not significantly affect SNR and mean frequency drift within the ranges of these environmental factors investigated. We found that humidity and temperature in the range investigated did not impact SNR and frequency. Our findings indicate high repeatability for SNR and magnetic field homogeneity; and moderate repeatability for image uniformity. △ Less

Submitted 14 April, 2023; originally announced April 2023.

Comments: 9 Figures and 2 tables

Filling out the missing gaps: Time Series Imputation with Semi-Supervised Learning

Authors: Karan Aggarwal, Jaideep Srivastava

Abstract: Missing data in time series is a challenging issue affecting time series analysis. Missing data occurs due to problems like data drops or sensor malfunctioning. Imputation methods are used to fill in these values, with quality of imputation having a significant impact on downstream tasks like classification. In this work, we propose a semi-supervised imputation method, ST-Impute, that uses both un… ▽ More Missing data in time series is a challenging issue affecting time series analysis. Missing data occurs due to problems like data drops or sensor malfunctioning. Imputation methods are used to fill in these values, with quality of imputation having a significant impact on downstream tasks like classification. In this work, we propose a semi-supervised imputation method, ST-Impute, that uses both unlabeled data along with downstream task's labeled data. ST-Impute is based on sparse self-attention and trains on tasks that mimic the imputation process. Our results indicate that the proposed method outperforms the existing supervised and unsupervised time series imputation methods measured on the imputation quality as well as on the downstream tasks ingesting imputed time series. △ Less

Submitted 9 April, 2023; originally announced April 2023.

Embarrassingly Simple MixUp for Time-series

Authors: Karan Aggarwal, Jaideep Srivastava

Abstract: Labeling time series data is an expensive task because of domain expertise and dynamic nature of the data. Hence, we often have to deal with limited labeled data settings. Data augmentation techniques have been successfully deployed in domains like computer vision to exploit the use of existing labeled data. We adapt one of the most commonly used technique called MixUp, in the time series domain.… ▽ More Labeling time series data is an expensive task because of domain expertise and dynamic nature of the data. Hence, we often have to deal with limited labeled data settings. Data augmentation techniques have been successfully deployed in domains like computer vision to exploit the use of existing labeled data. We adapt one of the most commonly used technique called MixUp, in the time series domain. Our proposed, MixUp++ and LatentMixUp++, use simple modifications to perform interpolation in raw time series and classification model's latent space, respectively. We also extend these methods with semi-supervised learning to exploit unlabeled data. We observe significant improvements of 1\% - 15\% on time series classification on two public datasets, for both low labeled data as well as high labeled data regimes, with LatentMixUp++. △ Less

Submitted 9 April, 2023; originally announced April 2023.

Language Is Not All You Need: Aligning Perception with Language Models

Authors: Shaohan Huang, Li Dong, Wenhui Wang, Yaru Hao, Saksham Singhal, Shuming Ma, Tengchao Lv, Lei Cui, Owais Khan Mohammed, Barun Patra, Qiang Liu, Kriti Aggarwal, Zewen Chi, Johan Bjorck, Vishrav Chaudhary, Subhojit Som, Xia Song, Furu Wei

Abstract: A big convergence of language, multimodal perception, action, and world modeling is a key step toward artificial general intelligence. In this work, we introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zero-shot). Specifically, we train Kosmos-1 from scratch on web-scale multimodal co… ▽ More A big convergence of language, multimodal perception, action, and world modeling is a key step toward artificial general intelligence. In this work, we introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zero-shot). Specifically, we train Kosmos-1 from scratch on web-scale multimodal corpora, including arbitrarily interleaved text and images, image-caption pairs, and text data. We evaluate various settings, including zero-shot, few-shot, and multimodal chain-of-thought prompting, on a wide range of tasks without any gradient updates or finetuning. Experimental results show that Kosmos-1 achieves impressive performance on (i) language understanding, generation, and even OCR-free NLP (directly fed with document images), (ii) perception-language tasks, including multimodal dialogue, image captioning, visual question answering, and (iii) vision tasks, such as image recognition with descriptions (specifying classification via text instructions). We also show that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge from language to multimodal, and from multimodal to language. In addition, we introduce a dataset of Raven IQ test, which diagnoses the nonverbal reasoning capability of MLLMs. △ Less

Submitted 1 March, 2023; v1 submitted 27 February, 2023; originally announced February 2023.

Controlling Personality Style in Dialogue with Zero-Shot Prompt-Based Learning

Authors: Angela Ramirez, Mamon Alsalihy, Kartik Aggarwal, Cecilia Li, Liren Wu, Marilyn Walker

Abstract: Prompt-based or in-context learning has achieved high zero-shot performance on many natural language generation (NLG) tasks. Here we explore the performance of prompt-based learning for simultaneously controlling the personality and the semantic accuracy of an NLG for task-oriented dialogue. We experiment with prompt-based learning on the PERSONAGE restaurant recommendation corpus to generate sema… ▽ More Prompt-based or in-context learning has achieved high zero-shot performance on many natural language generation (NLG) tasks. Here we explore the performance of prompt-based learning for simultaneously controlling the personality and the semantic accuracy of an NLG for task-oriented dialogue. We experiment with prompt-based learning on the PERSONAGE restaurant recommendation corpus to generate semantically and stylistically-controlled text for 5 different Big-5 personality types: agreeable, disagreeable, conscientious, unconscientious, and extravert. We test two different classes of discrete prompts to generate utterances for a particular personality style: (1) prompts that demonstrate generating directly from a meaning representation that includes a personality specification; and (2) prompts that rely on first converting the meaning representation to a textual pseudo-reference, and then using the pseudo-reference in a textual style transfer (TST) prompt. In each case, we show that we can vastly improve performance by over-generating outputs and ranking them, testing several ranking functions based on automatic metrics for semantic accuracy, personality-match, and fluency. We also test whether NLG personality demonstrations from the restaurant domain can be used with meaning representations for the video game domain to generate personality stylized utterances about video games. Our findings show that the TST prompts produces the highest semantic accuracy (78.46% for restaurants and 87.6% for video games) and personality accuracy (100% for restaurants and 97% for video games). Our results on transferring personality style to video game utterances are surprisingly good. To our knowledge, there is no previous work testing the application of prompt-based learning to simultaneously controlling both style and semantic accuracy in NLG. △ Less

Submitted 7 February, 2023; originally announced February 2023.

Comments: To appear at International Workshop on Spoken Dialogue Systems Technology, 2023

A broadband thermal emission spectrum of the ultra-hot Jupiter WASP-18b

Authors: Louis-Philippe Coulombe, Björn Benneke, Ryan Challener, Anjali A. A. Piette, Lindsey S. Wiser, Megan Mansfield, Ryan J. MacDonald, Hayley Beltz, Adina D. Feinstein, Michael Radica, Arjun B. Savel, Leonardo A. Dos Santos, Jacob L. Bean, Vivien Parmentier, Ian Wong, Emily Rauscher, Thaddeus D. Komacek, Eliza M. -R. Kempton, Xianyu Tan, Mark Hammond, Neil T. Lewis, Michael R. Line, Elspeth K. H. Lee, Hinna Shivkumar, Ian J. M. Crossfield , et al. (51 additional authors not shown)

Abstract: Close-in giant exoplanets with temperatures greater than 2,000 K (''ultra-hot Jupiters'') have been the subject of extensive efforts to determine their atmospheric properties using thermal emission measurements from the Hubble and Spitzer Space Telescopes. However, previous studies have yielded inconsistent results because the small sizes of the spectral features and the limited information conten… ▽ More Close-in giant exoplanets with temperatures greater than 2,000 K (''ultra-hot Jupiters'') have been the subject of extensive efforts to determine their atmospheric properties using thermal emission measurements from the Hubble and Spitzer Space Telescopes. However, previous studies have yielded inconsistent results because the small sizes of the spectral features and the limited information content of the data resulted in high sensitivity to the varying assumptions made in the treatment of instrument systematics and the atmospheric retrieval analysis. Here we present a dayside thermal emission spectrum of the ultra-hot Jupiter WASP-18b obtained with the NIRISS instrument on JWST. The data span 0.85 to 2.85 $μ$m in wavelength at an average resolving power of 400 and exhibit minimal systematics. The spectrum shows three water emission features (at $>$6$σ$ confidence) and evidence for optical opacity, possibly due to H$^-$, TiO, and VO (combined significance of 3.8$σ$). Models that fit the data require a thermal inversion, molecular dissociation as predicted by chemical equilibrium, a solar heavy element abundance (''metallicity'', M/H = 1.03$_{-0.51}^{+1.11}$ $\times$ solar), and a carbon-to-oxygen (C/O) ratio less than unity. The data also yield a dayside brightness temperature map, which shows a peak in temperature near the sub-stellar point that decreases steeply and symmetrically with longitude toward the terminators. △ Less

Submitted 20 January, 2023; v1 submitted 19 January, 2023; originally announced January 2023.

Comments: JWST ERS bright star observations. Uploaded to inform JWST Cycle 2 proposals. Manuscript under review. 50 pages, 14 figures, 2 tables

Develo** and deploying deep learning models in brain MRI: a review

Authors: Kunal Aggarwal, Marina Manso Jimeno, Keerthi Sravan Ravi, Gilberto Gonzalez, Sairam Geethanath

Abstract: Magnetic Resonance Imaging (MRI) of the brain has benefited from deep learning (DL) to alleviate the burden on radiologists and MR technologists, and improve throughput. The easy accessibility of DL tools have resulted in the rapid increase of DL models and subsequent peer-reviewed publications. However, the rate of deployment in clinical settings is low. Therefore, this review attempts to bring t… ▽ More Magnetic Resonance Imaging (MRI) of the brain has benefited from deep learning (DL) to alleviate the burden on radiologists and MR technologists, and improve throughput. The easy accessibility of DL tools have resulted in the rapid increase of DL models and subsequent peer-reviewed publications. However, the rate of deployment in clinical settings is low. Therefore, this review attempts to bring together the ideas from data collection to deployment into the clinic building on the guidelines and principles that accreditation agencies have espoused. We introduce the need for and the role of DL to deliver accessible MRI. This is followed by a brief review of DL examples in the context of neuropathologies. Based on these studies and others, we collate the prerequisites to develop and deploy DL models for brain MRI. We then delve into the guiding principles to practice good machine learning practices in the context of neuroimaging with a focus on explainability. A checklist based on the FDA's good machine learning practices is provided as a summary of these guidelines. Finally, we review the current challenges and future opportunities in DL for brain MRI. △ Less

Submitted 3 January, 2023; originally announced January 2023.

Comments: 4 figures and 1 table

Photochemically-produced SO$_2$ in the atmosphere of WASP-39b

Authors: Shang-Min Tsai, Elspeth K. H. Lee, Diana Powell, Peter Gao, Xi Zhang, Julianne Moses, Eric Hébrard, Olivia Venot, Vivien Parmentier, Sean Jordan, Renyu Hu, Munazza K. Alam, Lili Alderson, Natalie M. Batalha, Jacob L. Bean, Björn Benneke, Carver J. Bierson, Ryan P. Brady, Ludmila Carone, Aarynn L. Carter, Katy L. Chubb, Julie Inglis, Jérémy Leconte, Mercedes Lopez-Morales, Yamila Miguel , et al. (60 additional authors not shown)

Abstract: Photochemistry is a fundamental process of planetary atmospheres that regulates the atmospheric composition and stability. However, no unambiguous photochemical products have been detected in exoplanet atmospheres to date. Recent observations from the JWST Transiting Exoplanet Early Release Science Program found a spectral absorption feature at 4.05 $μ$m arising from SO$_2$ in the atmosphere of WA… ▽ More Photochemistry is a fundamental process of planetary atmospheres that regulates the atmospheric composition and stability. However, no unambiguous photochemical products have been detected in exoplanet atmospheres to date. Recent observations from the JWST Transiting Exoplanet Early Release Science Program found a spectral absorption feature at 4.05 $μ$m arising from SO$_2$ in the atmosphere of WASP-39b. WASP-39b is a 1.27-Jupiter-radii, Saturn-mass (0.28 M$_J$) gas giant exoplanet orbiting a Sun-like star with an equilibrium temperature of $\sim$1100 K. The most plausible way of generating SO$_2$ in such an atmosphere is through photochemical processes. Here we show that the SO$_2$ distribution computed by a suite of photochemical models robustly explains the 4.05 $μ$m spectral feature identified by JWST transmission observations with NIRSpec PRISM (2.7$σ$) and G395H (4.5$σ$). SO$_2$ is produced by successive oxidation of sulphur radicals freed when hydrogen sulphide (H$_2$S) is destroyed. The sensitivity of the SO$_2$ feature to the enrichment of the atmosphere by heavy elements (metallicity) suggests that it can be used as a tracer of atmospheric properties, with WASP-39b exhibiting an inferred metallicity of $\sim$10$\times$ solar. We further point out that SO$_2$ also shows observable features at ultraviolet and thermal infrared wavelengths not available from the existing observations. △ Less

Submitted 24 March, 2023; v1 submitted 18 November, 2022; originally announced November 2022.

Comments: 39 pages, 14 figures, accepted to be published in Nature

Early Release Science of the Exoplanet WASP-39b with JWST NIRSpec G395H

Authors: Lili Alderson, Hannah R. Wakeford, Munazza K. Alam, Natasha E. Batalha, Joshua D. Lothringer, Jea Adams Redai, Saugata Barat, Jonathan Brande, Mario Damiano, Tansu Daylan, Néstor Espinoza, Laura Flagg, Jayesh M. Goyal, David Grant, Renyu Hu, Julie Inglis, Elspeth K. H. Lee, Thomas Mikal-Evans, Lakeisha Ramos-Rosado, Pierre-Alexis Roy, Nicole L. Wallack, Natalie M. Batalha, Jacob L. Bean, Björn Benneke, Zachory K. Berta-Thompson , et al. (67 additional authors not shown)

Abstract: Measuring the abundances of carbon and oxygen in exoplanet atmospheres is considered a crucial avenue for unlocking the formation and evolution of exoplanetary systems. Access to an exoplanet's chemical inventory requires high-precision observations, often inferred from individual molecular detections with low-resolution space-based and high-resolution ground-based facilities. Here we report the m… ▽ More Measuring the abundances of carbon and oxygen in exoplanet atmospheres is considered a crucial avenue for unlocking the formation and evolution of exoplanetary systems. Access to an exoplanet's chemical inventory requires high-precision observations, often inferred from individual molecular detections with low-resolution space-based and high-resolution ground-based facilities. Here we report the medium-resolution (R$\sim$600) transmission spectrum of an exoplanet atmosphere between 3-5 $μ$m covering multiple absorption features for the Saturn-mass exoplanet WASP-39b, obtained with JWST NIRSpec G395H. Our observations achieve 1.46x photon precision, providing an average transit depth uncertainty of 221 ppm per spectroscopic bin, and present minimal impacts from systematic effects. We detect significant absorption from CO$_2$ (28.5$σ$) and H$_2$O (21.5$σ$), and identify SO$_2$ as the source of absorption at 4.1 $μ$m (4.8$σ$). Best-fit atmospheric models range between 3 and 10x solar metallicity, with sub-solar to solar C/O ratios. These results, including the detection of SO$_2$, underscore the importance of characterising the chemistry in exoplanet atmospheres, and showcase NIRSpec G395H as an excellent mode for time series observations over this critical wavelength range. △ Less

Submitted 18 November, 2022; originally announced November 2022.

Comments: 44 pages, 11 figures, 3 tables. Resubmitted after revision to Nature

Early Release Science of the exoplanet WASP-39b with JWST NIRSpec PRISM

Authors: Z. Rustamkulov, D. K. Sing, S. Mukherjee, E. M. May, J. Kirk, E. Schlawin, M. R. Line, C. Piaulet, A. L. Carter, N. E. Batalha, J. M. Goyal, M. López-Morales, J. D. Lothringer, R. J. MacDonald, S. E. Moran, K. B. Stevenson, H. R. Wakeford, N. Espinoza, J. L. Bean, N. M. Batalha, B. Benneke, Z. K. Berta-Thompson, I. J. M. Crossfield, P. Gao, L. Kreidberg , et al. (69 additional authors not shown)

Abstract: Transmission spectroscopy of exoplanets has revealed signatures of water vapor, aerosols, and alkali metals in a few dozen exoplanet atmospheres. However, these previous inferences with the Hubble and Spitzer Space Telescopes were hindered by the observations' relatively narrow wavelength range and spectral resolving power, which precluded the unambiguous identification of other chemical species… ▽ More Transmission spectroscopy of exoplanets has revealed signatures of water vapor, aerosols, and alkali metals in a few dozen exoplanet atmospheres. However, these previous inferences with the Hubble and Spitzer Space Telescopes were hindered by the observations' relatively narrow wavelength range and spectral resolving power, which precluded the unambiguous identification of other chemical species$-$in particular the primary carbon-bearing molecules. Here we report a broad-wavelength 0.5-5.5 $μ$m atmospheric transmission spectrum of WASP-39 b, a 1200 K, roughly Saturn-mass, Jupiter-radius exoplanet, measured with JWST NIRSpec's PRISM mode as part of the JWST Transiting Exoplanet Community Early Release Science Team program. We robustly detect multiple chemical species at high significance, including Na (19$σ$), H$_2$O (33$σ$), CO$_2$ (28$σ$), and CO (7$σ$). The non-detection of CH$_4$, combined with a strong CO$_2$ feature, favours atmospheric models with a super-solar atmospheric metallicity. An unanticipated absorption feature at 4$μ$m is best explained by SO$_2$ (2.7$σ$), which could be a tracer of atmospheric photochemistry. These observations demonstrate JWST's sensitivity to a rich diversity of exoplanet compositions and chemical processes. △ Less

Submitted 18 November, 2022; originally announced November 2022.

Comments: 41 pages, 4 main figures, 10 extended data figures, 4 tables. Under review in Nature

Stability of long-sustained oscillations induced by electron tunneling

Authors: Jorge Tabanera-Bravo, Florian Vigneau, Juliette Monsel, Kushagra Aggarwal, Léa Bresque, Federico Fedele, Federico Cerisola, G. A. D. Briggs, Janet Anders, Alexia Aufèves, Juan M. R. Parrondo, Natalia Ares

Abstract: Self-oscillations are the result of an efficient mechanism generating periodic motion from a constant power source. In quantum devices, these oscillations may arise due to the interaction between single electron dynamics and mechanical motion. We show that, due to the complexity of this mechanism, these self-oscillations may irrupt, vanish, or exhibit a bistable behaviour causing hysteresis cycles… ▽ More Self-oscillations are the result of an efficient mechanism generating periodic motion from a constant power source. In quantum devices, these oscillations may arise due to the interaction between single electron dynamics and mechanical motion. We show that, due to the complexity of this mechanism, these self-oscillations may irrupt, vanish, or exhibit a bistable behaviour causing hysteresis cycles. We observe these hysteresis cycles and characterize the stability of different regimes in both single and double quantum dot configurations. In particular cases, we find these oscillations stable for over 20 seconds, many orders of magnitude above electronic and mechanical characteristic timescales, revealing the robustness of the mechanism at play. △ Less

Submitted 26 March, 2024; v1 submitted 8 November, 2022; originally announced November 2022.

Comments: Revised version: 13 pages, 12 figures, includes the complete paper and the Supplemental Material

Journal ref: Phys. Rev. Research 6, 013291 (2024)

Identification of carbon dioxide in an exoplanet atmosphere

Authors: The JWST Transiting Exoplanet Community Early Release Science Team, Eva-Maria Ahrer, Lili Alderson, Natalie M. Batalha, Natasha E. Batalha, Jacob L. Bean, Thomas G. Beatty, Taylor J. Bell, Björn Benneke, Zachory K. Berta-Thompson, Aarynn L. Carter, Ian J. M. Crossfield, Néstor Espinoza, Adina D. Feinstein, Jonathan J. Fortney, Neale P. Gibson, Jayesh M. Goyal, Eliza M. -R. Kempton, James Kirk, Laura Kreidberg, Mercedes López-Morales, Michael R. Line, Joshua D. Lothringer, Sarah E. Moran, Sagnick Mukherjee , et al. (107 additional authors not shown)

Abstract: Carbon dioxide (CO2) is a key chemical species that is found in a wide range of planetary atmospheres. In the context of exoplanets, CO2 is an indicator of the metal enrichment (i.e., elements heavier than helium, also called "metallicity"), and thus formation processes of the primary atmospheres of hot gas giants. It is also one of the most promising species to detect in the secondary atmospheres… ▽ More Carbon dioxide (CO2) is a key chemical species that is found in a wide range of planetary atmospheres. In the context of exoplanets, CO2 is an indicator of the metal enrichment (i.e., elements heavier than helium, also called "metallicity"), and thus formation processes of the primary atmospheres of hot gas giants. It is also one of the most promising species to detect in the secondary atmospheres of terrestrial exoplanets. Previous photometric measurements of transiting planets with the Spitzer Space Telescope have given hints of the presence of CO2 but have not yielded definitive detections due to the lack of unambiguous spectroscopic identification. Here we present the detection of CO2 in the atmosphere of the gas giant exoplanet WASP-39b from transmission spectroscopy observations obtained with JWST as part of the Early Release Science Program (ERS). The data used in this study span 3.0 to 5.5 μm in wavelength and show a prominent CO2 absorption feature at 4.3 μm (26σ significance). The overall spectrum is well matched by one-dimensional, 10x solar metallicity models that assume radiative-convective-thermochemical equilibrium and have moderate cloud opacity. These models predict that the atmosphere should have water, carbon monoxide, and hydrogen sulfide in addition to CO2, but little methane. Furthermore, we also tentatively detect a small absorption feature near 4.0 μm that is not reproduced by these models. △ Less

Submitted 24 August, 2022; originally announced August 2022.

Comments: 27 pages, 6 figures, Accepted for publication in Nature, data and models available at https://doi.10.5281/zenodo.6959427

Image as a Foreign Language: BEiT Pretraining for All Vision and Vision-Language Tasks

Authors: Wenhui Wang, Hangbo Bao, Li Dong, Johan Bjorck, Zhiliang Peng, Qiang Liu, Kriti Aggarwal, Owais Khan Mohammed, Saksham Singhal, Subhojit Som, Furu Wei

Abstract: A big convergence of language, vision, and multimodal pretraining is emerging. In this work, we introduce a general-purpose multimodal foundation model BEiT-3, which achieves state-of-the-art transfer performance on both vision and vision-language tasks. Specifically, we advance the big convergence from three aspects: backbone architecture, pretraining task, and model scaling up. We introduce Mult… ▽ More A big convergence of language, vision, and multimodal pretraining is emerging. In this work, we introduce a general-purpose multimodal foundation model BEiT-3, which achieves state-of-the-art transfer performance on both vision and vision-language tasks. Specifically, we advance the big convergence from three aspects: backbone architecture, pretraining task, and model scaling up. We introduce Multiway Transformers for general-purpose modeling, where the modular architecture enables both deep fusion and modality-specific encoding. Based on the shared backbone, we perform masked "language" modeling on images (Imglish), texts (English), and image-text pairs ("parallel sentences") in a unified manner. Experimental results show that BEiT-3 obtains state-of-the-art performance on object detection (COCO), semantic segmentation (ADE20K), image classification (ImageNet), visual reasoning (NLVR2), visual question answering (VQAv2), image captioning (COCO), and cross-modal retrieval (Flickr30K, COCO). △ Less

Submitted 30 August, 2022; v1 submitted 22 August, 2022; originally announced August 2022.

Comments: 18 pages

Short second moment bound and Subconvexity for GL(3) $L$-functions

Authors: Keshav Aggarwal, Wing Hong Leung, Ritabrata Munshi

Abstract: Let $π$ be a Hecke cusp form for $\mathrm{SL}_3(\mathbb{Z})$. We bound the second moment average of $L(s,π)$ over a short interval to obtain the subconvexity estimate $$ L(1/2+it, π) \ll_{π, \varepsilon} (1+|t|)^{3/4-1/8+\varepsilon}. $$ Let $π$ be a Hecke cusp form for $\mathrm{SL}_3(\mathbb{Z})$. We bound the second moment average of $L(s,π)$ over a short interval to obtain the subconvexity estimate $$ L(1/2+it, π) \ll_{π, \varepsilon} (1+|t|)^{3/4-1/8+\varepsilon}. $$ △ Less

Submitted 15 June, 2022; v1 submitted 13 June, 2022; originally announced June 2022.

MSC Class: 11F66; 11M99

Masked Image Modeling Advances 3D Medical Image Analysis

Authors: Zekai Chen, Devansh Agarwal, Kshitij Aggarwal, Wiem Safta, Samit Hirawat, Venkat Sethuraman, Mariann Micsinai Balan, Kevin Brown

Abstract: Recently, masked image modeling (MIM) has gained considerable attention due to its capacity to learn from vast amounts of unlabeled data and has been demonstrated to be effective on a wide variety of vision tasks involving natural images. Meanwhile, the potential of self-supervised learning in modeling 3D medical images is anticipated to be immense due to the high quantities of unlabeled images, a… ▽ More Recently, masked image modeling (MIM) has gained considerable attention due to its capacity to learn from vast amounts of unlabeled data and has been demonstrated to be effective on a wide variety of vision tasks involving natural images. Meanwhile, the potential of self-supervised learning in modeling 3D medical images is anticipated to be immense due to the high quantities of unlabeled images, and the expense and difficulty of quality labels. However, MIM's applicability to medical images remains uncertain. In this paper, we demonstrate that masked image modeling approaches can also advance 3D medical images analysis in addition to natural images. We study how masked image modeling strategies leverage performance from the viewpoints of 3D medical image segmentation as a representative downstream task: i) when compared to naive contrastive learning, masked image modeling approaches accelerate the convergence of supervised training even faster (1.40$\times$) and ultimately produce a higher dice score; ii) predicting raw voxel values with a high masking ratio and a relatively smaller patch size is non-trivial self-supervised pretext-task for medical images modeling; iii) a lightweight decoder or projection head design for reconstruction is powerful for masked image modeling on 3D medical images which speeds up training and reduce cost; iv) finally, we also investigate the effectiveness of MIM methods under different practical scenarios where different image resolutions and labeled data ratios are applied. △ Less

Submitted 23 August, 2022; v1 submitted 25 April, 2022; originally announced April 2022.

Comments: 8 pages, 6 figures, 9 tables; Accepted by WACV2023

On the radio spectra of Galactic millisecond pulsars

Authors: Kshitij Aggarwal, Duncan. R. Lorimer

Abstract: With recent advances in the sensitivity of radio surveys of the Galactic disk, the number of millisecond pulsars (MSPs) has increased substantially in recent years such that it is now possible to study their demographic properties in more detail than in the past. We investigate what can be learned about the radio spectra of the MSP population. Using a sample of 179 MSPs detected in eleven surveys… ▽ More With recent advances in the sensitivity of radio surveys of the Galactic disk, the number of millisecond pulsars (MSPs) has increased substantially in recent years such that it is now possible to study their demographic properties in more detail than in the past. We investigate what can be learned about the radio spectra of the MSP population. Using a sample of 179 MSPs detected in eleven surveys carried out at radio frequencies in the range 0.135-6.6 GHz, we carry out detailed modeling of MSP radio spectral behaviour in this range. Employing Markov Chain Monte Carlo simulations to explore a multi-dimensional parameter space, and accurately accounting for observational selection effects, we find strong evidence in favour of the MSP population having a two-component power-law spectral model scaling with frequency, $ν$. Specifically, we find that MSP flux density spectra are approximately independent of frequency below 320 MHz, and proportional to $ν^{-1.5}$ at higher frequencies. This parameterization performs significantly better than single power-law models which over predict the number of MSPs seen in low-frequency (100-200 MHz) surveys. We compared our results with earlier work, and current understanding of the normal pulsar population, and use our model to make predictions for MSP yields in upcoming surveys. We demonstrate that the observed sample of MSPs could triple in the coming decade. △ Less

Submitted 8 March, 2022; originally announced March 2022.

Comments: 9 pages, 9 figures, 4 tables, submitted to MNRAS

Magnetic field reversal in the turbulent environment around a repeating fast radio burst

Authors: Reshma Anna-Thomas, Liam Connor, Shi Dai, Yi Feng, Sarah Burke-Spolaor, Paz Beniamini, Yuan-Pei Yang, Yongkun Zhang, Kshitij Aggarwal, Casey J. Law, Di Li, Chenhui Niu, Shami Chatterjee, Marilyn Cruces, Ran Duan, Miroslav D. Filipovi, George Hobbs, Ryan S. Lynch, Chenchen Miao, Jiarui Niu, Stella K. Ocker, Chao-Wei Tsai, Pei Wang, Mengyao Xue, Jumei Yao , et al. (5 additional authors not shown)

Abstract: Fast radio bursts (FRBs) are brief, intense flashes of radio waves from unidentified extragalactic sources. Polarized FRBs originate in highly magnetized environments. We report observations of the repeating FRB 20190520B spanning seventeen months , which show its amount of Faraday rotation is highly variable and twice changes its sign. The FRB also depolarizes below radio frequencies around 1 to… ▽ More Fast radio bursts (FRBs) are brief, intense flashes of radio waves from unidentified extragalactic sources. Polarized FRBs originate in highly magnetized environments. We report observations of the repeating FRB 20190520B spanning seventeen months , which show its amount of Faraday rotation is highly variable and twice changes its sign. The FRB also depolarizes below radio frequencies around 1 to 3 GHz. We interpret these properties as due to change in the parallel component of the integrated magnetic field along the line-of-sight, including reversals. This could result from propagation through a turbulent, magnetized screen of plasma located between $10^{-5}$ to 100 parsecs of the FRB source. This is consistent with the bursts passing through the stellar wind of a binary companion of the FRB source. △ Less

Submitted 12 May, 2023; v1 submitted 22 February, 2022; originally announced February 2022.

Comments: arXiv:2203.08151 has been merged into this

Journal ref: Science 380,599-603(2023)

Deep Image Prior using Stein's Unbiased Risk Estimator: SURE-DIP

Authors: Maneesh John, Hemant Kumar Aggarwal, Qing Zou, Mathews Jacob

Abstract: Deep learning algorithms that rely on extensive training data are revolutionizing image recovery from ill-posed measurements. Training data is scarce in many imaging applications, including ultra-high-resolution imaging. The deep image prior (DIP) algorithm was introduced for single-shot image recovery, completely eliminating the need for training data. A challenge with this scheme is the need for… ▽ More Deep learning algorithms that rely on extensive training data are revolutionizing image recovery from ill-posed measurements. Training data is scarce in many imaging applications, including ultra-high-resolution imaging. The deep image prior (DIP) algorithm was introduced for single-shot image recovery, completely eliminating the need for training data. A challenge with this scheme is the need for early stop** to minimize the overfitting of the CNN parameters to the noise in the measurements. We introduce a generalized Stein's unbiased risk estimate (GSURE) loss metric to minimize the overfitting. Our experiments show that the SURE-DIP approach minimizes the overfitting issues, thus offering significantly improved performance over classical DIP schemes. We also use the SURE-DIP approach with model-based unrolling architectures, which offers improved performance over direct inversion schemes. △ Less

Submitted 21 November, 2021; originally announced November 2021.

VLMo: Unified Vision-Language Pre-Training with Mixture-of-Modality-Experts

Authors: Hangbo Bao, Wenhui Wang, Li Dong, Qiang Liu, Owais Khan Mohammed, Kriti Aggarwal, Subhojit Som, Furu Wei

Abstract: We present a unified Vision-Language pretrained Model (VLMo) that jointly learns a dual encoder and a fusion encoder with a modular Transformer network. Specifically, we introduce Mixture-of-Modality-Experts (MoME) Transformer, where each block contains a pool of modality-specific experts and a shared self-attention layer. Because of the modeling flexibility of MoME, pretrained VLMo can be fine-tu… ▽ More We present a unified Vision-Language pretrained Model (VLMo) that jointly learns a dual encoder and a fusion encoder with a modular Transformer network. Specifically, we introduce Mixture-of-Modality-Experts (MoME) Transformer, where each block contains a pool of modality-specific experts and a shared self-attention layer. Because of the modeling flexibility of MoME, pretrained VLMo can be fine-tuned as a fusion encoder for vision-language classification tasks, or used as a dual encoder for efficient image-text retrieval. Moreover, we propose a stagewise pre-training strategy, which effectively leverages large-scale image-only and text-only data besides image-text pairs. Experimental results show that VLMo achieves state-of-the-art results on various vision-language tasks, including VQA, NLVR2 and image-text retrieval. The code and pretrained models are available at https://aka.ms/vlmo. △ Less

Submitted 27 May, 2022; v1 submitted 3 November, 2021; originally announced November 2021.

Comments: Work in progress

On the Fast Radio Burst and Persistent Radio Source Populations

Authors: C. J. Law, L. Connor, K. Aggarwal

Abstract: The first Fast Radio Burst (FRB) to be precisely localized was associated with a luminous persistent radio source (PRS). Recently, a second FRB/PRS association was discovered for another repeating source of FRBs. However, it is not clear what makes FRBs or PRS or how they are related. We compile FRB and PRS properties to consider the population of FRB/PRS sources. We suggest a practical definition… ▽ More The first Fast Radio Burst (FRB) to be precisely localized was associated with a luminous persistent radio source (PRS). Recently, a second FRB/PRS association was discovered for another repeating source of FRBs. However, it is not clear what makes FRBs or PRS or how they are related. We compile FRB and PRS properties to consider the population of FRB/PRS sources. We suggest a practical definition for PRS as FRB associations with luminosity greater than $10^{29}$ erg s$^{-1}$ Hz$^{-1}$ that is not attributed to star-formation activity in the host galaxy. We model the probability distribution of the fraction of FRBs with PRS for repeaters and non-repeaters, showing there is not yet evidence for repeaters to be preferentially associated with PRS. We discuss how FRB/PRS sources may be distinguished by the combination of active repetition and an excess dispersion measure local to the FRB environment. We use CHIME/FRB event statistics to bound the mean per-source repetition rate of FRBs to be between 25 and 440 yr$^{-1}$. We use this to provide a bound on the density of FRB-emitting sources in the local universe of between $2.2\times10^2$ and $5.2\times10^4$ Gpc$^{-3}$ assuming a pulsar-like beam width for FRB emission. This density implies that PRS may comprise as much as 1\% of compact, luminous radio sources detected in the local universe. The cosmic density and phenomenology of PRS are similar to that of the newly-discovered, off-nuclear "wandering" AGN. We argue that it is likely that some PRS have already been detected and misidentified as AGN. △ Less

Submitted 28 October, 2021; originally announced October 2021.

Comments: Submitted. 12 pages, 3 figures

Search for fast radio transients using Arecibo drift-scan observations at 1.4 GHz

Authors: B. B. P. Perera, A. J. Smith, S. Vaddi, R. Carballo-Rubio, A. McGilvray, A. Venkataraman, D. Anish Roshi, P. K. Manoharan, P. Perillat, E. Lieb, D. R. Lorimer, M. A. McLaughlin, D. Agarwal, K. Aggarwal, S. M. Ransom

Abstract: We conducted a drift-scan observation campaign using the 305-m Arecibo telescope in January and March 2020 when the observatory was temporarily closed during the intense earthquakes and the initial outbreak of the COVID-19 pandemic, respectively. The primary objective of the survey was to search for fast radio transients, including Fast Radio Bursts (FRBs) and Rotating Radio Transients (RRATs). We… ▽ More We conducted a drift-scan observation campaign using the 305-m Arecibo telescope in January and March 2020 when the observatory was temporarily closed during the intense earthquakes and the initial outbreak of the COVID-19 pandemic, respectively. The primary objective of the survey was to search for fast radio transients, including Fast Radio Bursts (FRBs) and Rotating Radio Transients (RRATs). We used the 7-beam ALFA receiver to observe different sections of the sky within the declination region $\sim$(10$-$20) deg on 23 nights and collected 160 hours of data in total. We searched our data for single-pulse transients, covering up to a maximum dispersion measure of 11 000 pc cm$^{-3}$ at which the dispersion delay across the entire bandwidth is equal to the 13 s transit length of our observations. The analysis produced more than 18 million candidates. Machine learning techniques sorted the radio frequency interference and possibly astrophysical candidates, allowing us to visually inspect and confirm the candidate transients. We found no evidence for new astrophysical transients in our data. We also searched for emission from repeated transient signals, but found no evidence for such sources. We detected single pulses from two known pulsars in our observations and their measured flux densities are consistent with the expected values. Based on our observations and sensitivity, we estimated the upper limit for the FRB rate to be $<$2.8$\times10^5$ sky$^{-1}$ day$^{-1}$ above a fluence of 0.16 Jy ms at 1.4 GHz, which is consistent with the rates from other telescopes and surveys. △ Less

Submitted 27 October, 2021; originally announced October 2021.

Comments: 11 pages, 8 figures. Accepted 2021 October 27 for publication in MNRAS

A repeating fast radio burst associated with a persistent radio source

Authors: C. -H. Niu, K. Aggarwal, D. Li, X. Zhang, S. Chatterjee, C. -W. Tsai, W. Yu, C. J. Law, S. Burke-Spolaor, J. M. Cordes, Y. -K. Zhang, S. Ocker, J. -M. Yao, P. Wang, Y. Feng, Y. Niino, C. Bochenek, M. Cruces, L. Connor, J. -A. Jiang, S. Dai, R. Luo, G. -D. Li, C. -C. Miao, J. -R. Niu , et al. (10 additional authors not shown)

Abstract: The dispersive sweep of fast radio bursts (FRBs) has been used to probe the ionized baryon content of the intergalactic medium, which is assumed to dominate the total extragalactic dispersion. While the host galaxy contributions to dispersion measure (DM) appear to be small for most FRBs, in at least one case there is evidence for an extreme magneto-ionic local environment and a compact persistent… ▽ More The dispersive sweep of fast radio bursts (FRBs) has been used to probe the ionized baryon content of the intergalactic medium, which is assumed to dominate the total extragalactic dispersion. While the host galaxy contributions to dispersion measure (DM) appear to be small for most FRBs, in at least one case there is evidence for an extreme magneto-ionic local environment and a compact persistent radio source. Here we report the detection and localization of the repeating FRB 20190520B, which is co-located with a compact, persistent radio source and associated with a dwarf host galaxy of high specific star formation rate at a redshift $z=0.241\pm0.001$. The estimated host galaxy DM $\approx 903^{+72}_{-111}$ pc cm$^{-3}$, nearly an order of magnitude higher than the average of FRB host galaxies, far exceeds the DM contribution of the intergalactic medium. Caution is thus warranted in inferring redshifts for FRBs without accurate host galaxy identifications. △ Less

Submitted 20 April, 2022; v1 submitted 14 October, 2021; originally announced October 2021.

Comments: Accepted, Version 3

Observational effects of banded repeating FRBs

Authors: Kshitij Aggarwal

Abstract: Recent observations have shown that repeating Fast Radio Bursts (FRBs) exhibit band-limited emission, whose frequency-dependent amplitude can be modeled using a Gaussian function. In this analysis, we show that banded emission of FRBs can lead to incompleteness across the observing band. This biases the detected sample of bursts and can explain the various shapes of cumulative energy distributions… ▽ More Recent observations have shown that repeating Fast Radio Bursts (FRBs) exhibit band-limited emission, whose frequency-dependent amplitude can be modeled using a Gaussian function. In this analysis, we show that banded emission of FRBs can lead to incompleteness across the observing band. This biases the detected sample of bursts and can explain the various shapes of cumulative energy distributions seen for repeating FRBs. We assume a Gaussian shape of the burst spectra and used simulations to demonstrate the above bias using an FRB 121102-like example. We recovered energy distributions that showed a break in power-law and flattening of power-law at low energies, based on the fluence threshold of the observations. We provide recommendations for single-pulse searches and analysis of repeating FRBs to account for this incompleteness. Primarily, we recommend that burst spectra should be modeled to estimate the intrinsic fluence and bandwidth of the burst robustly. Also, bursts that lie mainly within the observing band should be used for analyses of energy distributions. We show that the bimodality reported in the distribution of energies of FRB 121102 by Li et al. (2021) disappears when burst bandwidth, instead of the center frequency of the observation, is used to estimate energy. Sub-banded searches will also aid in detecting band-limited bursts. All the analysis scripts used in this work are available in a Github repository. △ Less

Submitted 26 September, 2021; v1 submitted 10 August, 2021; originally announced August 2021.

Comments: 10 pages, 5 figures, Accepted for publication in ApJL

Characterizing the FRB host galaxy population and its connection to transients in the local and extragalactic Universe

Authors: Shivani Bhandari, Kasper E. Heintz, Kshitij Aggarwal, Lachlan Marnoch, Cherie K. Day, Jessica Sydnor, Sarah Burke-Spolaor, Casey J. Law, J. Xavier Prochaska, Nicolas Tejos, Keith W. Bannister, Bryan J. Butler, Adam T. Deller, R. D. Ekers, Chris Flynn, Wen-fai Fong, Clancy W. James, T. Joseph W. Lazio, Rui Luo, Elizabeth K. Mahony, Stuart D. Ryder, Elaine M. Sadler, Ryan M. Shannon, **Lin Han, Kejia Lee , et al. (1 additional authors not shown)

Abstract: We present the localization and host galaxies of one repeating and two apparently non-repeating Fast Radio Bursts. FRB20180301A was detected and localized with the Karl G. Jansky Very Large Array to a star-forming galaxy at $z=0.3304$. FRB20191228A, and FRB20200906A were detected and localized by the Australian Square Kilometre Array Pathfinder to host galaxies at $z=0.2430$ and $z=0.3688$, respec… ▽ More We present the localization and host galaxies of one repeating and two apparently non-repeating Fast Radio Bursts. FRB20180301A was detected and localized with the Karl G. Jansky Very Large Array to a star-forming galaxy at $z=0.3304$. FRB20191228A, and FRB20200906A were detected and localized by the Australian Square Kilometre Array Pathfinder to host galaxies at $z=0.2430$ and $z=0.3688$, respectively. We combine these with 13 other well-localized FRBs in the literature, and analyze the host galaxy properties. We find no significant differences in the host properties of repeating and apparently non-repeating FRBs. FRB hosts are moderately star-forming, with masses slightly offset from the star-forming main-sequence. Star formation and low-ionization nuclear emission-line region (LINER) emission are major sources of ionization in FRB host galaxies, with the former dominant in repeating FRB hosts. FRB hosts do not track stellar mass and star formation as seen in field galaxies (more than 95% confidence). FRBs are rare in massive red galaxies, suggesting that progenitor formation channels are not solely dominated by delayed channels which lag star formation by Gigayears. The global properties of FRB hosts are indistinguishable from core-collapse supernovae (CCSNe) and short gamma-ray bursts (SGRBs) hosts, and the spatial offset (from galaxy centers) of FRBs is mostly inconsistent with that of the Galactic neutron star population (95% confidence). The spatial offsets of FRBs (normalized to the galaxy effective radius) also differ from those of globular clusters (GCs) in late- and early-type galaxies with 95% confidence. △ Less

Submitted 16 November, 2021; v1 submitted 3 August, 2021; originally announced August 2021.

Comments: 27 pages, 12 figures, 8 tables. Accepted for publication in the Astrophysical Journal

Machine Learning Advances aiding Recognition and Classification of Indian Monuments and Landmarks

Authors: Aditya Jyoti Paul, Smaranjit Ghose, Kanishka Aggarwal, Niketha Nethaji, Shivam Pal, Arnab Dutta Purkayastha

Abstract: Tourism in India plays a quintessential role in the country's economy with an estimated 9.2% GDP share for the year 2018. With a yearly growth rate of 6.2%, the industry holds a huge potential for being the primary driver of the economy as observed in the nations of the Middle East like the United Arab Emirates. The historical and cultural diversity exhibited throughout the geography of the nation… ▽ More Tourism in India plays a quintessential role in the country's economy with an estimated 9.2% GDP share for the year 2018. With a yearly growth rate of 6.2%, the industry holds a huge potential for being the primary driver of the economy as observed in the nations of the Middle East like the United Arab Emirates. The historical and cultural diversity exhibited throughout the geography of the nation is a unique spectacle for people around the world and therefore serves to attract tourists in tens of millions in number every year. Traditionally, tour guides or academic professionals who study these heritage monuments were responsible for providing information to the visitors regarding their architectural and historical significance. However, unfortunately this system has several caveats when considered on a large scale such as unavailability of sufficient trained people, lack of accurate information, failure to convey the richness of details in an attractive format etc. Recently, machine learning approaches revolving around the usage of monument pictures have been shown to be useful for rudimentary analysis of heritage sights. This paper serves as a survey of the research endeavors undertaken in this direction which would eventually provide insights for building an automated decision system that could be utilized to make the experience of tourism in India more modernized for visitors. △ Less

Submitted 29 July, 2021; originally announced July 2021.

Comments: Currently under review

DECIFE: Detecting Collusive Users Involved in Blackmarket Following Services on Twitter

Authors: Hridoy Sankar Dutta, Kartik Aggarwal, Tanmoy Chakraborty

Abstract: The popularity of Twitter has fostered the emergence of various fraudulent user activities - one such activity is to artificially bolster the social reputation of Twitter profiles by gaining a large number of followers within a short time span. Many users want to gain followers to increase the visibility and reach of their profiles to wide audiences. This has provoked several blackmarket services… ▽ More The popularity of Twitter has fostered the emergence of various fraudulent user activities - one such activity is to artificially bolster the social reputation of Twitter profiles by gaining a large number of followers within a short time span. Many users want to gain followers to increase the visibility and reach of their profiles to wide audiences. This has provoked several blackmarket services to garner huge attention by providing artificial followers via the network of agreeable and compromised accounts in a collusive manner. Their activity is difficult to detect as the blackmarket services shape their behavior in such a way that users who are part of these services disguise themselves as genuine users. In this paper, we propose DECIFE, a framework to detect collusive users involved in producing 'following' activities through blackmarket services with the intention to gain collusive followers in return. We first construct a heterogeneous user-tweet-topic network to leverage the follower/followee relationships and linguistic properties of a user. The heterogeneous network is then decomposed to form four different subgraphs that capture the semantic relations between the users. An attention-based subgraph aggregation network is proposed to learn and combine the node representations from each subgraph. The combined representation is finally passed on to a hypersphere learning objective to detect collusive users. Comprehensive experiments on our curated dataset are conducted to validate the effectiveness of DECIFE by comparing it with other state-of-the-art approaches. To our knowledge, this is the first attempt to detect collusive users involved in blackmarket 'following services' on Twitter. △ Less

Submitted 24 July, 2021; originally announced July 2021.

Comprehensive analysis of a dense sample of FRB 121102 bursts

Authors: Kshitij Aggarwal, Devansh Agarwal, Evan F. Lewis, Reshma Anna-Thomas, Jacob Cardinal Tremblay, Sarah Burke-Spolaor, Maura A. McLaughlin, Duncan R. Lorimer

Abstract: We present an analysis of a densely repeating sample of bursts from the first repeating fast radio burst, FRB 121102. We reanalysed the data used by Gourdji et al. (2019) and detected 93 additional bursts using our single-pulse search pipeline. In total, we detected 133 bursts in three hours of data at a center frequency of 1.4 GHz using the Arecibo telescope, and develop robust modeling strategie… ▽ More We present an analysis of a densely repeating sample of bursts from the first repeating fast radio burst, FRB 121102. We reanalysed the data used by Gourdji et al. (2019) and detected 93 additional bursts using our single-pulse search pipeline. In total, we detected 133 bursts in three hours of data at a center frequency of 1.4 GHz using the Arecibo telescope, and develop robust modeling strategies to constrain the spectro-temporal properties of all the bursts in the sample. Most of the burst profiles show a scattering tail, and burst spectra are well modeled by a Gaussian with a median width of 230 MHz. We find a lack of emission below 1300 MHz, consistent with previous studies of FRB 121102. We also find that the peak of the log-normal distribution of wait times decreases from 207 s to 75 s using our larger sample of bursts, as compared to that of Gourdji et al. (2019). Our observations do not favor either Poissonian or Weibull distributions for the burst rate distribution. We searched for periodicity in the bursts using multiple techniques but did not detect any significant period. The cumulative burst energy distribution exhibits a broken power-law shape, with the lower and higher-energy slopes of $-0.4\pm0.1$ and $-1.8\pm0.2$, with the break at $(2.3\pm0.2)\times 10^{37}$ ergs. We provide our burst fitting routines as a python package BURSTFIT that can be used to model the spectrogram of any complex FRB or pulsar pulse using robust fitting techniques. All the other analysis scripts and results are publicly available. △ Less

Submitted 23 September, 2021; v1 submitted 12 July, 2021; originally announced July 2021.

Comments: 27 pages, 13 figures, 5 Tables; Accepted for publication in ApJ

The host galaxy and persistent radio counterpart of FRB 20201124A

Authors: Vikram Ravi, Casey J. Law, Dongzi Li, Kshitij Aggarwal, Sarah Burke-Spolaor, Liam Connor, T. Joseph W. Lazio, Dana Simard, Jean Somalwar, Shriharsh P. Tendulkar

Abstract: The physical properties of fast radio burst (FRB) host galaxies provide important clues towards the nature of FRB sources. The 16 FRB hosts identified thus far span three orders of magnitude in mass and specific star-formation rate, implicating a ubiquitously occurring progenitor object. FRBs localised with ~arcsecond accuracy also enable effective searches for associated multi-wavelength and mult… ▽ More The physical properties of fast radio burst (FRB) host galaxies provide important clues towards the nature of FRB sources. The 16 FRB hosts identified thus far span three orders of magnitude in mass and specific star-formation rate, implicating a ubiquitously occurring progenitor object. FRBs localised with ~arcsecond accuracy also enable effective searches for associated multi-wavelength and multi-timescale counterparts, such as the persistent radio source associated with FRB 20121102A. Here we present a localisation of the repeating source FRB 20201124A, and its association with a host galaxy (SDSS J050803.48+260338.0, z=0.098) and persistent radio source. The galaxy is massive ($\sim3\times10^{10} M_{\odot}$), star-forming (few solar masses per year), and dusty. Very Large Array and Very Long Baseline Array observations of the persistent radio source measure a luminosity of $1.2\times10^{29}$ erg s$^{-1}$ Hz$^{-1}$, and show that is extended on scales $\gtrsim50$ mas. We associate this radio emission with the ongoing star-formation activity in SDSS J050803.48+260338.0. Deeper, more detailed observations are required to better utilise the milliarcsecond-scale localisation of FRB 20201124A reported from the European VLBI Network, and determine the origin of the large dispersion measure ($150-220$ pc cm$^{-3}$) contributed by the host. SDSS J050803.48+260338.0 is an order of magnitude more massive than any galaxy or stellar system previously associated with a repeating FRB source, but is comparable to the hosts of so far non-repeating FRBs, further building the link between the two apparent populations. △ Less

Submitted 17 June, 2021; originally announced June 2021.

Comments: 9 pages, 6 figures, 1 table, submitted to MNRAS

A repeating fast radio burst source in a globular cluster

Authors: F. Kirsten, B. Marcote, K. Nimmo, J. W. T. Hessels, M. Bhardwaj, S. P. Tendulkar, A. Keimpema, J. Yang, M. P. Snelders, P. Scholz, A. B. Pearlman, C. J. Law, W. M. Peters, M. Giroletti, Z. Paragi, C. Bassa, D. M. Hewitt, U. Bach, V. Bezrukovs, M. Burgay, S. T. Buttaccio, J. E. Conway, A. Corongiu, R. Feiler, O. Forssén , et al. (41 additional authors not shown)

Abstract: Fast radio bursts (FRBs) are exceptionally luminous flashes of unknown physical origin, reaching us from other galaxies (Petroff et al. 2019). Most FRBs have only ever been seen once, while others flash repeatedly, though sporadically (Spitler et al. 2016, CHIME/FRB Collaboration et al. 2021). Many models invoke magnetically powered neutron stars (magnetars) as the engines producing FRB emission (… ▽ More Fast radio bursts (FRBs) are exceptionally luminous flashes of unknown physical origin, reaching us from other galaxies (Petroff et al. 2019). Most FRBs have only ever been seen once, while others flash repeatedly, though sporadically (Spitler et al. 2016, CHIME/FRB Collaboration et al. 2021). Many models invoke magnetically powered neutron stars (magnetars) as the engines producing FRB emission (Margalit & Metzger 2018, CHIME/FRB Collaboration et al. 2020). Recently, CHIME/FRB announced the discovery (Bhardwaj et al. 2021) of the repeating FRB 20200120E, coming from the direction of the nearby grand design spiral galaxy M81. Four potential counterparts at other observing wavelengths were identified (Bhardwaj et al. 2021) but no definitive association with these sources, or M81, could be made. Here we report an extremely precise localisation of FRB 20200120E, which allows us to associate it with a globular cluster (GC) in the M81 galactic system and to place it ~2pc offset from the optical center of light of the GC. This confirms (Bhardwaj et al. 2021) that FRB 20200120E is 40 times closer than any other known extragalactic FRB. Because such GCs host old stellar populations, this association strongly challenges FRB models that invoke young magnetars formed in a core-collapse supernova as powering FRB emission. We propose, instead, that FRB 20200120E is a highly magnetised neutron star formed via either accretion-induced collapse of a white dwarf or via merger of compact stars in a binary system (Margalit et al. 2019). Alternative scenarios involving compact binary systems, efficiently formed inside globular clusters, could also be responsible for the observed bursts. △ Less

Submitted 29 September, 2021; v1 submitted 24 May, 2021; originally announced May 2021.

Comments: Submitted. Comments welcome

Does Putting a Linguist in the Loop Improve NLU Data Collection?

Authors: Alicia Parrish, William Huang, Omar Agha, Soo-Hwan Lee, Nikita Nangia, Alex Warstadt, Karmanya Aggarwal, Emily Allaway, Tal Linzen, Samuel R. Bowman

Abstract: Many crowdsourced NLP datasets contain systematic gaps and biases that are identified only after data collection is complete. Identifying these issues from early data samples during crowdsourcing should make mitigation more efficient, especially when done iteratively. We take natural language inference as a test case and ask whether it is beneficial to put a linguist `in the loop' during data coll… ▽ More Many crowdsourced NLP datasets contain systematic gaps and biases that are identified only after data collection is complete. Identifying these issues from early data samples during crowdsourcing should make mitigation more efficient, especially when done iteratively. We take natural language inference as a test case and ask whether it is beneficial to put a linguist `in the loop' during data collection to dynamically identify and address gaps in the data by introducing novel constraints on the task. We directly compare three data collection protocols: (i) a baseline protocol, (ii) a linguist-in-the-loop intervention with iteratively-updated constraints on the task, and (iii) an extension of linguist-in-the-loop that provides direct interaction between linguists and crowdworkers via a chatroom. The datasets collected with linguist involvement are more reliably challenging than baseline, without loss of quality. But we see no evidence that using this data in training leads to better out-of-domain model performance, and the addition of a chat platform has no measurable effect on the resulting dataset. We suggest integrating expert analysis \textit{during} data collection so that the expert can dynamically address gaps and biases in the dataset. △ Less

Submitted 14 April, 2021; originally announced April 2021.

Comments: 14 pages, 10 figures

Robust Assessment of Clustering Methods for Fast Radio Transient Candidates

Authors: Kshitij Aggarwal, Sarah Burke-Spolaor, Casey J. Law, Geoffrey C. Bower, Bryan J. Butler, Paul B. Demorest, T. Joseph W. Lazio, Justin Linford, Jessica Sydnor, Reshma Anna-Thomas

Abstract: Fast radio transient search algorithms identify signals of interest by iterating and applying a threshold on a set of matched filters. These filters are defined by properties of the transient such as time and dispersion. A real transient can trigger hundreds of search trials, each of which has to be post-processed for visualization and classification tasks. In this paper, we have explored a range… ▽ More Fast radio transient search algorithms identify signals of interest by iterating and applying a threshold on a set of matched filters. These filters are defined by properties of the transient such as time and dispersion. A real transient can trigger hundreds of search trials, each of which has to be post-processed for visualization and classification tasks. In this paper, we have explored a range of unsupervised clustering algorithms to cluster these redundant candidate detections. We demonstrate this for Realfast, the commensal fast transient search system at the Very Large Array. We use four features for clustering: sky position (l, m), time and dispersion measure (DM). We develop a custom performance metric that makes sure that the candidates are clustered into a small number of pure clusters, i.e, clusters with either astrophysical or noise candidates. We then use this performance metric to compare eight different clustering algorithms. We show that using sky location along with DM/time improves clustering performance by $\sim$10% as compared to the traditional DM/time-based clustering. Therefore, positional information should be used during clustering if it can be made available. We conduct several tests to compare the performance and generalisability of clustering algorithms to other transient datasets and propose a strategy that can be used to choose an algorithm. Our performance metric and clustering strategy can be easily extended to different single-pulse search pipelines and other astronomy and non-astronomy-based applications. △ Less

Submitted 14 April, 2021; originally announced April 2021.

Comments: 17 pages, 7 figures, Submitted to AAS Journals (ApJ)

Multi-wavelength follow-up of FRB 180309

Authors: Kshitij Aggarwal, Sarah Burke-Spolaor, Nicolas Tejos, Giuliano Pignata, J. Xavier Prochaska, Vikram Ravi, Jane F. Kaczmarek, Stefan Oslowski

Abstract: We report on the results of multi-wavelength follow-up observations with Gemini, VLA, and ATCA, to search for a host galaxy and any persistent radio emission associated with FRB 180309. This FRB is among the most luminous FRB detections to date, with a luminosity of $> 8.7\times 10^{32}$ erg Hz$^{-1}$ at the dispersion-based redshift upper limit of 0.32. We used the high-significance detection of… ▽ More We report on the results of multi-wavelength follow-up observations with Gemini, VLA, and ATCA, to search for a host galaxy and any persistent radio emission associated with FRB 180309. This FRB is among the most luminous FRB detections to date, with a luminosity of $> 8.7\times 10^{32}$ erg Hz$^{-1}$ at the dispersion-based redshift upper limit of 0.32. We used the high-significance detection of FRB 180309 with the Parkes Telescope and a beam model of the Parkes Multibeam Receiver to improve the localization of the FRB to a region spanning approximately $\sim2'\times2'$. We aimed to seek bright galaxies within this region to determine the strongest candidates as the originator of this highly luminous FRB. We identified optical sources within the localization region above our r-band magnitude limit of 24.27, fourteen of which have photometric redshifts whose fitted mean is consistent with the redshift upper limit ($z < 0.32$) of our FRB. Two of these galaxies are coincident with marginally detected "persistent" radio sources of flux density 24.3$μ$Jy beam$^{-1}$ and 22.1$μ$Jy beam$^{-1}$ respectively. Our redshift-dependent limit on the luminosity of any associated persistent radio source is comparable to the luminosity limits for other localized FRBs. We analyze several properties of the candidate hosts we identified, including chance association probability, redshift, and presence of radio emission, however it remains possible that any of these galaxies could be the host of this FRB. Follow-up spectroscopy on these objects to explore their H$α$ emission and ionization contents, as well as to obtain more precisely measured redshifts, may be able to isolate a single host for this luminous FRB. △ Less

Submitted 21 September, 2021; v1 submitted 8 April, 2021; originally announced April 2021.

Comments: 14 pages, 7 figures; Accepted for publication in AAS Journals (ApJ)

Ultrastrong coupling between electron tunneling and mechanical motion

Authors: Florian Vigneau, Juliette Monsel, Jorge Tabanera, Kushagra Aggarwal, Léa Bresque, Federico Fedele, G. A. D Briggs, Janet Anders, Juan M. R. Parrondo, Alexia Auffèves, Natalia Ares

Abstract: The ultrastrong coupling of single-electron tunneling and nanomechanical motion opens exciting opportunities to explore fundamental questions and develop new platforms for quantum technologies. We have measured and modeled this electromechanical coupling in a fully-suspended carbon nanotube device and report a ratio of $g_\mathrm{m}/ω_\mathrm{m} = 2.72 \pm 0.14$, where… ▽ More The ultrastrong coupling of single-electron tunneling and nanomechanical motion opens exciting opportunities to explore fundamental questions and develop new platforms for quantum technologies. We have measured and modeled this electromechanical coupling in a fully-suspended carbon nanotube device and report a ratio of $g_\mathrm{m}/ω_\mathrm{m} = 2.72 \pm 0.14$, where $g_\mathrm{m}/2π= 0.80\pm 0.04$ GHz is the coupling strength and $ω_\mathrm{m}/2π=294.5$ MHz is the mechanical resonance frequency. This is well within the ultrastrong coupling regime and the highest among all other electromechanical platforms. We show that, although this regime was present in similar fully-suspended carbon nanotube devices, it went unnoticed. Even higher ratios could be achieved with improvement on device design. △ Less

Submitted 6 October, 2022; v1 submitted 28 March, 2021; originally announced March 2021.

Comments: 13 pages, 11 figures This new version contains the same model and analysis but applied to new experimental data compared to the previous version. This is why the coupling ratio is now 2.72 instead of 1.3 and most of the experimental parameters have changed

Journal ref: Phys. Rev. Research 4, 043168 (2022)

Electron Impact Excitation of O III: An Assessment

Authors: K M Aggarwal

Abstract: Tayal and Zatsarinny [Astrophys. J. 850 (2017) 147] have reported results for energy levels, radiative rates (A-values), lifetimes, and effective collision strengths ($Υ$) for transitions among 202 levels of C-like O~III. For the calculations they have adopted the multi-configuration Hartree-Fock (MCHF) code for the energy levels and A-values, and B-spline $R$-matrix (BSR) code for $Υ$. Their repo… ▽ More Tayal and Zatsarinny [Astrophys. J. 850 (2017) 147] have reported results for energy levels, radiative rates (A-values), lifetimes, and effective collision strengths ($Υ$) for transitions among 202 levels of C-like O~III. For the calculations they have adopted the multi-configuration Hartree-Fock (MCHF) code for the energy levels and A-values, and B-spline $R$-matrix (BSR) code for $Υ$. Their reported results cover a (much) larger range of levels/transitions than generally available in the literature, and appear to be accurate for energy levels and A-values. However, the magnitude and behaviour of $Υ$ do not appear to be correct for several transitions. We demonstrate this through our independent calculations by adopting the flexible atomic code (FAC) and recommend a fresh calculation for this important ion. △ Less

Submitted 3 March, 2021; originally announced March 2021.

Comments: 12 pages of text including 11 figures

Probabilistic Association of Transients to their Hosts (PATH)

Authors: Kshitij Aggarwal, Tamás Budavári, Adam T. Deller, Tarraneh Eftekhari, Clancy W. James, J. Xavier Prochaska, Shriharsh P. Tendulkar

Abstract: We introduce a new method to estimate the probability that an extragalactic transient source is associated with a candidate host galaxy. This approach relies solely on simple observables: sky coordinates and their uncertainties, galaxy fluxes and angular sizes. The formalism invokes Bayes' rule to calculate the posterior probability P(O_i|x) from the galaxy prior P(O), observables x, and an assume… ▽ More We introduce a new method to estimate the probability that an extragalactic transient source is associated with a candidate host galaxy. This approach relies solely on simple observables: sky coordinates and their uncertainties, galaxy fluxes and angular sizes. The formalism invokes Bayes' rule to calculate the posterior probability P(O_i|x) from the galaxy prior P(O), observables x, and an assumed model for the true distribution of transients in/around their host galaxies. Using simulated transients placed in the well-studied COSMOS field, we consider several agnostic and physically motivated priors and offset distributions to explore the method sensitivity. We then apply the methodology to the set of 13~fast radio bursts (FRBs) localized with an uncertainty of several arcseconds. Our methodology finds nine of these are securely associated to a single host galaxy, P(O_i|x)>0.95. We examine the observed and intrinsic properties of these secure FRB hosts, recovering similar distributions as previous works. Furthermore, we find a strong correlation between the apparent magnitude of the securely identified host galaxies and the estimated cosmic dispersion measures of the corresponding FRBs, which results from the Macquart relation. Future work with FRBs will leverage this relation and other measures from the secure hosts as priors for future associations. The methodology is generic to transient type, localization error, and image quality. We encourage its application to other transients where host galaxy associations are critical to the science, e.g. gravitational wave events, gamma-ray bursts, and supernovae. We have encoded the technique in Python on GitHub: https://github.com/FRBs/astropath. △ Less

Submitted 21 February, 2021; originally announced February 2021.

Comments: In press, ApJ; comments still welcome; Visit https://github.com/FRBs/astropath to use and build PATH