-
Direct Preference Optimization for Suppressing Hallucinated Prior Exams in Radiology Report Generation
Authors:
Oishi Banerjee,
Hong-Yu Zhou,
Subathra Adithan,
Stephen Kwak,
Kay Wu,
Pranav Rajpurkar
Abstract:
Recent advances in generative vision-language models (VLMs) have exciting potential implications for AI in radiology, yet VLMs are also known to produce hallucinations, nonsensical text, and other unwanted behaviors that can waste clinicians' time and cause patient harm. Drawing on recent work on direct preference optimization (DPO), we propose a simple method for modifying the behavior of pretrai…
▽ More
Recent advances in generative vision-language models (VLMs) have exciting potential implications for AI in radiology, yet VLMs are also known to produce hallucinations, nonsensical text, and other unwanted behaviors that can waste clinicians' time and cause patient harm. Drawing on recent work on direct preference optimization (DPO), we propose a simple method for modifying the behavior of pretrained VLMs performing radiology report generation by suppressing unwanted types of generations. We apply our method to the prevention of hallucinations of prior exams, addressing a long-established problem behavior in models performing chest X-ray report generation. Across our experiments, we find that DPO fine-tuning achieves a 3.2-4.8x reduction in lines hallucinating prior exams while maintaining model performance on clinical accuracy metrics. Our work is, to the best of our knowledge, the first work to apply DPO to medical VLMs, providing a data- and compute- efficient way to suppress problem behaviors while maintaining overall clinical accuracy.
△ Less
Submitted 14 June, 2024; v1 submitted 10 June, 2024;
originally announced June 2024.
-
FineRadScore: A Radiology Report Line-by-Line Evaluation Technique Generating Corrections with Severity Scores
Authors:
Alyssa Huang,
Oishi Banerjee,
Kay Wu,
Eduardo Pontes Reis,
Pranav Rajpurkar
Abstract:
The current gold standard for evaluating generated chest x-ray (CXR) reports is through radiologist annotations. However, this process can be extremely time-consuming and costly, especially when evaluating large numbers of reports. In this work, we present FineRadScore, a Large Language Model (LLM)-based automated evaluation metric for generated CXR reports. Given a candidate report and a ground-t…
▽ More
The current gold standard for evaluating generated chest x-ray (CXR) reports is through radiologist annotations. However, this process can be extremely time-consuming and costly, especially when evaluating large numbers of reports. In this work, we present FineRadScore, a Large Language Model (LLM)-based automated evaluation metric for generated CXR reports. Given a candidate report and a ground-truth report, FineRadScore gives the minimum number of line-by-line corrections required to go from the candidate to the ground-truth report. Additionally, FineRadScore provides an error severity rating with each correction and generates comments explaining why the correction was needed. We demonstrate that FineRadScore's corrections and error severity scores align with radiologist opinions. We also show that, when used to judge the quality of the report as a whole, FineRadScore aligns with radiologists as well as current state-of-the-art automated CXR evaluation metrics. Finally, we analyze FineRadScore's shortcomings to provide suggestions for future improvements.
△ Less
Submitted 31 May, 2024;
originally announced May 2024.
-
BenchMD: A Benchmark for Unified Learning on Medical Images and Sensors
Authors:
Kathryn Wantlin,
Chenwei Wu,
Shih-Cheng Huang,
Oishi Banerjee,
Farah Dadabhoy,
Veeral Vipin Mehta,
Ryan Wonhee Han,
Fang Cao,
Raja R. Narayan,
Errol Colak,
Adewole Adamson,
Laura Heacock,
Geoffrey H. Tison,
Alex Tamkin,
Pranav Rajpurkar
Abstract:
Medical data poses a daunting challenge for AI algorithms: it exists in many different modalities, experiences frequent distribution shifts, and suffers from a scarcity of examples and labels. Recent advances, including transformers and self-supervised learning, promise a more universal approach that can be applied flexibly across these diverse conditions. To measure and drive progress in this dir…
▽ More
Medical data poses a daunting challenge for AI algorithms: it exists in many different modalities, experiences frequent distribution shifts, and suffers from a scarcity of examples and labels. Recent advances, including transformers and self-supervised learning, promise a more universal approach that can be applied flexibly across these diverse conditions. To measure and drive progress in this direction, we present BenchMD: a benchmark that tests how well unified, modality-agnostic methods, including architectures and training techniques (e.g. self-supervised learning, ImageNet pretraining),perform on a diverse array of clinically-relevant medical tasks. BenchMD combines 19 publicly available datasets for 7 medical modalities, including 1D sensor data, 2D images, and 3D volumetric scans. Our benchmark reflects real-world data constraints by evaluating methods across a range of dataset sizes, including challenging few-shot settings that incentivize the use of pretraining. Finally, we evaluate performance on out-of-distribution data collected at different hospitals than the training data, representing naturally-occurring distribution shifts that frequently degrade the performance of medical AI models. Our baseline results demonstrate that no unified learning technique achieves strong performance across all modalities, leaving ample room for improvement on the benchmark. Code is released at https://github.com/rajpurkarlab/BenchMD.
△ Less
Submitted 26 June, 2023; v1 submitted 17 April, 2023;
originally announced April 2023.
-
Moduli of curves on toric varieties and their stable cohomology
Authors:
Oishee Banerjee
Abstract:
We prove that the cohomology of the moduli space of morphisms of a fixed finite degree from a smooth projective curve $C$ of genus $g$ to a complete simplicial toric variety $\mathbb{P}(Σ)$, denoted by the rational polyhedral fan $Σ$, stabilizes. As an arithmetic consequence we obtain a resolution of the Batyrev-Manin conjecture for toric varieties over global function fields in all but finitely m…
▽ More
We prove that the cohomology of the moduli space of morphisms of a fixed finite degree from a smooth projective curve $C$ of genus $g$ to a complete simplicial toric variety $\mathbb{P}(Σ)$, denoted by the rational polyhedral fan $Σ$, stabilizes. As an arithmetic consequence we obtain a resolution of the Batyrev-Manin conjecture for toric varieties over global function fields in all but finitely many characteristics.
△ Less
Submitted 11 October, 2022;
originally announced October 2022.
-
Étale cohomological stability of the moduli space of stable elliptic surfaces
Authors:
Oishee Banerjee,
Jun-Yong Park,
Johannes Schmitt
Abstract:
We compute the (stable) étale cohomology of $\mathrm{Hom}_{n}(C, \mathcal{P}(\vecλ))$, the moduli stack of degree $n$ morphisms from a smooth projective curve $C$ to the weighted projective stack $\mathcal{P}(\vecλ)$, the latter being a stacky quotient defined by $\mathcal{P}(\vecλ) := \left[\mathbb{A}^N-\{0\}/\mathbb{G}_m\right]$, where $\mathbb{G}_m$ acts by weights…
▽ More
We compute the (stable) étale cohomology of $\mathrm{Hom}_{n}(C, \mathcal{P}(\vecλ))$, the moduli stack of degree $n$ morphisms from a smooth projective curve $C$ to the weighted projective stack $\mathcal{P}(\vecλ)$, the latter being a stacky quotient defined by $\mathcal{P}(\vecλ) := \left[\mathbb{A}^N-\{0\}/\mathbb{G}_m\right]$, where $\mathbb{G}_m$ acts by weights $\vecλ = (λ_0, \cdots, λ_N) \in \mathbb{Z}^N_{+}$. Our key ingredient is formulating and proving the étale cohomological descent over the category $ΔS$, the symmetric (semi)simplicial category. An immediate arithmetic consequence is the resolution of the geometric Batyrev--Manin type conjecture for weighted projective stacks over global function fields. Along the way, we also analyze the intersection theory on weighted projectivizations of vector bundles on smooth Deligne-Mumford stacks.
△ Less
Submitted 6 July, 2022;
originally announced July 2022.
-
Analysis of a Tau Neutrino Origin for the Near-Horizon Air Shower Events Observed by the Fourth Flight of the Antarctic Impulsive Transient Antenna (ANITA)
Authors:
R. Prechelt,
S. A. Wissel,
A. Romero-Wolf,
C. Burch,
P. W. Gorham,
P. Allison,
J. Alvarez-Muñiz,
O. Banerjee,
L. Batten,
J. J. Beatty,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
W. Carvalho Jr.,
C. H. Chen,
P. Chen,
Y. Chen,
J. M. Clem,
A. Connolly,
L. Cremonesi,
B. Dailey,
C. Deaconu,
P. F. Dowkontt
, et al. (43 additional authors not shown)
Abstract:
We study in detail the sensitivity of the Antarctic Impulsive Transient Antenna (ANITA) to possible $ν_τ$ point source fluxes detected via $τ$-lepton-induced air showers. This investigation is framed around the observation of four upward-going extensive air shower events very close to the horizon seen in ANITA-IV. We find that these four upgoing events are not observationally inconsistent with…
▽ More
We study in detail the sensitivity of the Antarctic Impulsive Transient Antenna (ANITA) to possible $ν_τ$ point source fluxes detected via $τ$-lepton-induced air showers. This investigation is framed around the observation of four upward-going extensive air shower events very close to the horizon seen in ANITA-IV. We find that these four upgoing events are not observationally inconsistent with $τ$-induced EASs from Earth-skimming $ν_τ$, both in their spectral properties as well as in their observed locations on the sky. These four events, as well as the overall diffuse and point source exposure to Earth-skimming $ν_τ$, are also compared against published ultrahigh-energy neutrino limits from the Pierre Auger Observatory. While none of these four events occurred at sky locations simultaneously visible by Auger, the implied fluence necessary for ANITA to observe these events is in strong tension with limits set by Auger across a wide range of energies and is additionally in tension with ANITA's Askaryan in-ice neutrino channel above $10^{19}$ eV. We conclude by discussing some of the technical challenges with simulating and analyzing these near horizon events and the potential for future observatories to observe similar events.
△ Less
Submitted 13 December, 2021;
originally announced December 2021.
-
CheXtransfer: Performance and Parameter Efficiency of ImageNet Models for Chest X-Ray Interpretation
Authors:
Alexander Ke,
William Ellsworth,
Oishi Banerjee,
Andrew Y. Ng,
Pranav Rajpurkar
Abstract:
Deep learning methods for chest X-ray interpretation typically rely on pretrained models developed for ImageNet. This paradigm assumes that better ImageNet architectures perform better on chest X-ray tasks and that ImageNet-pretrained weights provide a performance boost over random initialization. In this work, we compare the transfer performance and parameter efficiency of 16 popular convolutiona…
▽ More
Deep learning methods for chest X-ray interpretation typically rely on pretrained models developed for ImageNet. This paradigm assumes that better ImageNet architectures perform better on chest X-ray tasks and that ImageNet-pretrained weights provide a performance boost over random initialization. In this work, we compare the transfer performance and parameter efficiency of 16 popular convolutional architectures on a large chest X-ray dataset (CheXpert) to investigate these assumptions. First, we find no relationship between ImageNet performance and CheXpert performance for both models without pretraining and models with pretraining. Second, we find that, for models without pretraining, the choice of model family influences performance more than size within a family for medical imaging tasks. Third, we observe that ImageNet pretraining yields a statistically significant boost in performance across architectures, with a higher boost for smaller architectures. Fourth, we examine whether ImageNet architectures are unnecessarily large for CheXpert by truncating final blocks from pretrained models, and find that we can make models 3.25x more parameter-efficient on average without a statistically significant drop in performance. Our work contributes new experimental evidence about the relation of ImageNet to chest x-ray interpretation performance.
△ Less
Submitted 20 February, 2021; v1 submitted 17 January, 2021;
originally announced January 2021.
-
A search for ultrahigh-energy neutrinos associated with astrophysical sources using the third flight of ANITA
Authors:
C. Deaconu,
L. Batten,
P. Allison,
O. Banerjee,
J. J. Beatty,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. H. Chen,
P. Chen,
Y. Chen,
J. M. Clem,
A. Connolly,
L. Cremonesi,
B. Dailey,
P. F. Dowkontt,
B. D. Fox,
J. W. H. Gordon,
P. W. Gorham,
C. Hast,
B. Hill,
S. Y. Hsu,
J. J. Huang
, et al. (38 additional authors not shown)
Abstract:
The ANtarctic Impulsive Transient Antenna (ANITA) long-duration balloon experiment is sensitive to interactions of ultra high-energy (E > 10^{18} eV) neutrinos in the Antarctic ice sheet. The third flight of ANITA, lasting 22 days, began in December 2014. We develop a methodology to search for energetic neutrinos spatially and temporally coincident with potential source classes in ANITA data. This…
▽ More
The ANtarctic Impulsive Transient Antenna (ANITA) long-duration balloon experiment is sensitive to interactions of ultra high-energy (E > 10^{18} eV) neutrinos in the Antarctic ice sheet. The third flight of ANITA, lasting 22 days, began in December 2014. We develop a methodology to search for energetic neutrinos spatially and temporally coincident with potential source classes in ANITA data. This methodology is applied to several source classes: the TXS 0506+056 blazar and NGC 1068, the first potential TeV neutrino sources identified by IceCube, flaring high-energy blazars reported by the Fermi All-Sky Variability Analysis, gamma-ray bursts, and supernovae. Among searches within the five source classes, one candidate was identified as associated with SN 2015D, although not at a statistically significant level. We proceed to place upper limits on the source classes. We further comment on potential applications of this methodology to more sensitive future instruments.
△ Less
Submitted 15 March, 2021; v1 submitted 6 October, 2020;
originally announced October 2020.
-
Unusual Near-horizon Cosmic-ray-like Events Observed by ANITA-IV
Authors:
ANITA Collaboration,
P. W. Gorham,
A. Ludwig,
C. Deaconu,
P. Cao,
P. Allison,
O. Banerjee,
L. Batten,
D. Bhattacharya,
J. J. Beatty,
K. Belov,
W. R. Binns,
V. Bugaev,
C. H. Chen,
P. Chen,
Y. Chen,
J. M. Clem,
L. Cremonesi,
B. Dailey,
P. F. Dowkontt,
B. D. Fox,
J. W. H. Gordon,
C. Hast,
B. Hill,
S. Y. Hsu
, et al. (35 additional authors not shown)
Abstract:
ANITA's fourth long-duration balloon flight in late 2016 detected 29 cosmic-ray (CR)-like events on a background of $0.37^{+0.27}_{-0.17}$ anthropogenic events. CRs are mainly seen in reflection off the Antarctic ice sheets, creating a characteristic phase-inverted waveform polarity. However, four of the below-horizon CR-like events show anomalous non-inverted polarity, a $p = 5.3 \times 10^{-4}$…
▽ More
ANITA's fourth long-duration balloon flight in late 2016 detected 29 cosmic-ray (CR)-like events on a background of $0.37^{+0.27}_{-0.17}$ anthropogenic events. CRs are mainly seen in reflection off the Antarctic ice sheets, creating a characteristic phase-inverted waveform polarity. However, four of the below-horizon CR-like events show anomalous non-inverted polarity, a $p = 5.3 \times 10^{-4}$ chance if due to background. All anomalous events are from locations near the horizon; ANITA-IV observed no steeply-upcoming anomalous events similar to the two such events seen in prior flights.
△ Less
Submitted 19 November, 2020; v1 submitted 13 August, 2020;
originally announced August 2020.
-
On the cohomology of certain subspaces of $\mathit{Sym}^n(¶^1)$ and Occam's razor for Hodge structures
Authors:
Oishee Banerjee
Abstract:
In \cite{Vakil13} Vakil and Wood made several conjectures on the topology of symmetric powers of geometrically irreducible varieties based on their computations on motivic zeta functions. Two of those conjectures are about subspaces of $\Sym^n(¶^1)$. In this note, we disprove one of them thereby obtaining a counterexample to the principle of Occcam's razor for Hodge structures; and we prove that t…
▽ More
In \cite{Vakil13} Vakil and Wood made several conjectures on the topology of symmetric powers of geometrically irreducible varieties based on their computations on motivic zeta functions. Two of those conjectures are about subspaces of $\Sym^n(¶^1)$. In this note, we disprove one of them thereby obtaining a counterexample to the principle of Occcam's razor for Hodge structures; and we prove that the other conjecture, with a minor correction, holds true.
△ Less
Submitted 30 March, 2020;
originally announced March 2020.
-
Filtration of cohomology via symmetric semisimplicial spaces
Authors:
Oishee Banerjee
Abstract:
In the simplicial theory of hypercoverings, we replace the indexing category $Δ$ by the \emph{symmetric simplicial category} $ΔS$ and study (a class of) $ΔS$-hypercoverings, which we call \emph{spaces admitting symmetric (semi)simplicial filtration}. For $ΔS$-hypercoverings we construct a spectral sequence, somewhat like the Čech-to-derived category spectral sequence. The advantage of working on…
▽ More
In the simplicial theory of hypercoverings, we replace the indexing category $Δ$ by the \emph{symmetric simplicial category} $ΔS$ and study (a class of) $ΔS$-hypercoverings, which we call \emph{spaces admitting symmetric (semi)simplicial filtration}. For $ΔS$-hypercoverings we construct a spectral sequence, somewhat like the Čech-to-derived category spectral sequence. The advantage of working on $ΔS$ is that all of the combinatorial complexities that come with working on $Δ$ are bypassed, giving simpler, unified proof of known results like the computation of (in some cases, stable) singular cohomology (with $\mathbb{Q}$ coefficients) and et al e cohomology (with $\mathbb{Q}_{\ell}$ coefficients) of the moduli space of degree $n$ maps $C\to \mathbb{P}^r$, $C$ a smooth projective curve of genus $g$, of unordered configuration spaces etc. as well as new: that of the moduli space of smooth sections of a fixed $\mathfrak{g}^r_d$ that is $m$-very ample for some $m$.
△ Less
Submitted 10 August, 2023; v1 submitted 1 September, 2019;
originally announced September 2019.
-
The Simulation of the Sensitivity of the Antarctic Impulsive Transient Antenna (ANITA) to Askaryan Radiation from Cosmogenic Neutrinos Interacting in the Antarctic Ice
Authors:
L. Cremonesi,
A. Connolly,
P. Allison,
O. Banerjee,
L. Batten,
J. J. Beatty,
K. Bechtol,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. C. Chen,
C. H. Chen,
P. Chen,
J. M. Clem,
B. Dailey,
C. Deaconu,
P. F. Dowkontt,
B. D. Fox,
J. W. H. Gordon,
P. W. Gorham,
B. Hill,
J. J. Huang,
K. Hughes
, et al. (35 additional authors not shown)
Abstract:
A Monte Carlo simulation program for the radio detection of Ultra High Energy (UHE) neutrino interactions in the Antarctic ice as viewed by the Antarctic Impulsive Transient Antenna (ANITA) is described in this article. The program, icemc, provides an input spectrum of UHE neutrinos, the parametrization of the Askaryan radiation generated by their interaction in the ice, and the propagation of the…
▽ More
A Monte Carlo simulation program for the radio detection of Ultra High Energy (UHE) neutrino interactions in the Antarctic ice as viewed by the Antarctic Impulsive Transient Antenna (ANITA) is described in this article. The program, icemc, provides an input spectrum of UHE neutrinos, the parametrization of the Askaryan radiation generated by their interaction in the ice, and the propagation of the radiation through ice and air to a simulated model of the third and fourth ANITA flights. This paper provides an overview of the icemc simulation, descriptions of the physics models used and of the ANITA electronics processing chain, data/simulation comparisons to validate the predicted performance, and a summary of the impact of published results.
△ Less
Submitted 12 August, 2019; v1 submitted 26 March, 2019;
originally announced March 2019.
-
Constraints on the ultra-high energy cosmic neutrino flux from the fourth flight of ANITA
Authors:
P. W. Gorham,
P. Allison,
O. Banerjee,
L. Batten,
J. J. Beatty,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. C. Chen,
C. H. Chen,
P. Chen,
J. M. Clem,
A. Connolly,
L. Cremonesi,
B. Dailey,
C. Deaconu,
P. F. Dowkontt,
B. D. Fox,
J. W. H. Gordon,
C. Hast,
B. Hill,
S. Y. Hsu,
J. J. Huang
, et al. (35 additional authors not shown)
Abstract:
The ANtarctic Impulsive Transient Antenna (ANITA) NASA long-duration balloon payload completed its fourth flight in December 2016, after 28 days of flight time. ANITA is sensitive to impulsive broadband radio emission from interactions of ultra-high-energy neutrinos in polar ice (Askaryan emission). We present the results of two separate blind analyses searching for signals from Askaryan emission…
▽ More
The ANtarctic Impulsive Transient Antenna (ANITA) NASA long-duration balloon payload completed its fourth flight in December 2016, after 28 days of flight time. ANITA is sensitive to impulsive broadband radio emission from interactions of ultra-high-energy neutrinos in polar ice (Askaryan emission). We present the results of two separate blind analyses searching for signals from Askaryan emission in the data from the fourth flight of ANITA. The more sensitive analysis, with a better expected limit, has a background estimate of $0.64^{+0.69}_{-0.45}$ and an analysis efficiency of $82\pm2\%$. The second analysis has a background estimate of $0.34^{+0.66}_{-0.16}$ and an analysis efficiency of $71\pm6\%$. Each analysis found one event in the signal region, consistent with the background estimate for each analysis. The resulting limit further tightens the constraints on the diffuse flux of ultra-high-energy neutrinos at energies above $10^{19.5}$ eV.
△ Less
Submitted 11 February, 2019;
originally announced February 2019.
-
A comprehensive analysis of anomalous ANITA events disfavors a diffuse tau-neutrino flux origin
Authors:
A. Romero-Wolf,
S. A. Wissel,
H. Schoorlemmer,
W. R. Carvalho Jr,
J. Alvarez-Muñiz,
E. Zas,
P. Allison,
O. Banerjee,
L. Batten,
J. J. Beatty,
K. Bechtol,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. C. Chen,
C. H. Chen,
P. Chen,
J. M. Clem,
A. Connolly,
L. Cremonesi,
B. Dailey,
C. Deaconu,
P. F. Dowkontt
, et al. (38 additional authors not shown)
Abstract:
Recently, the ANITA collaboration reported on two upward-going extensive air shower events consistent with a primary particle that emerges from the surface of the ice. These events may be of $ν_τ$ origin, in which the neutrino interacts within the Earth to produce a $τ$ lepton that emerges from the Earth, decays in the atmosphere, and initiates an extensive air shower. In this paper we estimate an…
▽ More
Recently, the ANITA collaboration reported on two upward-going extensive air shower events consistent with a primary particle that emerges from the surface of the ice. These events may be of $ν_τ$ origin, in which the neutrino interacts within the Earth to produce a $τ$ lepton that emerges from the Earth, decays in the atmosphere, and initiates an extensive air shower. In this paper we estimate an upper bound on the ANITA acceptance to a diffuse $ν_τ$ flux detected via $τ$-lepton-induced air showers within the bounds of Standard Model (SM) uncertainties. By comparing this estimate with the acceptance of Pierre Auger Observatory and IceCube and assuming SM interactions, we conclude that a $ν_τ$ origin of these events would imply a neutrino flux at least two orders of magnitude above current bounds.
△ Less
Submitted 5 February, 2019; v1 submitted 17 November, 2018;
originally announced November 2018.
-
Cohomology of the space of polynomial maps on $\mathbb{A}^1$ with prescribed ramification
Authors:
Oishee Banerjee
Abstract:
In this paper we study the moduli spaces $Simp^m_n$ of degree $n+1$ morphisms $ \mathbb{A}^1_{K} \to \mathbb{A}^1_{K}$ with "ramification length $<m$" over an algebraically closed field $K$. For each $m$, the moduli space $Simp^m_n$ is a Zariski open subset of the space of degree $n+1$ polynomials over $K$ up to $Aut (\mathbb{A}^1_{K})$. It is, in a way, orthogonal to the many papers about polynom…
▽ More
In this paper we study the moduli spaces $Simp^m_n$ of degree $n+1$ morphisms $ \mathbb{A}^1_{K} \to \mathbb{A}^1_{K}$ with "ramification length $<m$" over an algebraically closed field $K$. For each $m$, the moduli space $Simp^m_n$ is a Zariski open subset of the space of degree $n+1$ polynomials over $K$ up to $Aut (\mathbb{A}^1_{K})$. It is, in a way, orthogonal to the many papers about polynomials with prescribed zeroes -- here we are prescribing, instead, the ramification data. Exploiting the topological properties of the poset that encodes the ramification behaviour, we use a sheaf-theoretic argument to compute $H^*(Simp^m_n(\mathbb{C}); \mathbb{Q})$ as well as the étale cohomology $H^*_{et}({Simp^m_n}_{/K}; \mathbb{Q}_{\ell})$ for $char K=0$ or $char K> n+1$. As a by-product we obtain that $H^*(Simp^m_n(\mathbb{C}); \mathbb{Q})$ is independent of $n$, thus implying rational cohomological stability. When $char K>0$ our methods compute $H^*_{et}(Simp^m_n; \mathbb{Q}_{\ell})$ provided $char K>n+1$ and show that the étale cohomology groups in positive characteristics do not stabilize.
△ Less
Submitted 7 November, 2018; v1 submitted 3 October, 2018;
originally announced October 2018.
-
Upward-Pointing Cosmic-Ray-like Events Observed with ANITA
Authors:
Andres Romero-Wolf,
P. W. Gorham,
J. Nam,
S. Hoover,
P. Allison,
O. Banerjee,
L. Batten,
J. J. Beatty,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. Chen,
P. Chen,
J. M. Clem,
A. Connolly,
B. Dailey,
C. Deaconu,
L. Cremonesi,
P. F. Dowkontt,
M. A. DuVernois,
R. C. Field,
B. D. Fox,
D. Goldstein
, et al. (51 additional authors not shown)
Abstract:
These proceedings address a recent publication by the ANITA collaboration of four upward- pointing cosmic-ray-like events observed in the first flight of ANITA. Three of these events were consistent with stratospheric cosmic-ray air showers where the axis of propagation does not inter- sect the surface of the Earth. The fourth event was consistent with a primary particle that emerges from the surf…
▽ More
These proceedings address a recent publication by the ANITA collaboration of four upward- pointing cosmic-ray-like events observed in the first flight of ANITA. Three of these events were consistent with stratospheric cosmic-ray air showers where the axis of propagation does not inter- sect the surface of the Earth. The fourth event was consistent with a primary particle that emerges from the surface of the ice suggesting a possible τ-lepton decay as the origin of this event. These proceedings follow-up on the modeling and testing of the hypothesis that this event was of τ neutrino origin.
△ Less
Submitted 30 September, 2018;
originally announced October 2018.
-
On ANITA's sensitivity to long-lived, charged massive particles
Authors:
Amy Connolly,
Patrick Allison,
Oindree Banerjee
Abstract:
We propose that the Antarctic Impulsive Transient Antenna (ANITA) can serve as a detector for long-lived, charged particles, through its measurement of extensive air showers from secondary leptons. To test this on an example model, we simulate the production of staus inside the earth from interactions between ultra-high energy neutrinos and nuclei. We propose that results of ANITA searches for upg…
▽ More
We propose that the Antarctic Impulsive Transient Antenna (ANITA) can serve as a detector for long-lived, charged particles, through its measurement of extensive air showers from secondary leptons. To test this on an example model, we simulate the production of staus inside the earth from interactions between ultra-high energy neutrinos and nuclei. We propose that results of ANITA searches for upgoing air showers can be interpreted in terms of constraints on long-lived, charged massive particles (CHAMPs) and consider a supersymmetric partner of the tau lepton, the stau, as an example of such a particle. Exploring the parameter space in stau mass and lifetimes, we find that the stau properties that lead to an observable signal in ANITA are highly energy dependent. At $10^{18.5}$ eV, we find that the best constraints on the product of the neutrino flux and the stau production cross section would be placed near $m_{\tildeτ}=$1 TeV and $τ_{\tildeτ}=$10 ns. Thus ANITA could be sensitive to new physics in a region of parameter space that is unconstrained by experiments at the Large Hadron Collider.
△ Less
Submitted 23 July, 2018;
originally announced July 2018.
-
Observation of an Unusual Upward-going Cosmic-ray-like Event in the Third Flight of ANITA
Authors:
P. W. Gorham,
B. Rotter,
P. Allison,
O. Banerjee,
L. Batten,
J. J. Beatty,
K. Bechtol,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. C. Chen,
C. H. Chen,
P. Chen,
J. M. Clem,
A. Connolly,
L. Cremonesi,
B. Dailey,
C. Deaconu,
P. F. Dowkontt,
B. D. Fox,
J. W. H. Gordon,
C. Hast,
B. Hill
, et al. (38 additional authors not shown)
Abstract:
We report on an upward traveling, radio-detected cosmic-ray-like impulsive event with characteristics closely matching an extensive air shower. This event, observed in the third flight of the Antarctic Impulsive Transient Antenna (ANITA), a NASA-sponsored long-duration balloon payload, is consistent with a similar event reported in a previous flight. These events may be produced by the atmospheric…
▽ More
We report on an upward traveling, radio-detected cosmic-ray-like impulsive event with characteristics closely matching an extensive air shower. This event, observed in the third flight of the Antarctic Impulsive Transient Antenna (ANITA), a NASA-sponsored long-duration balloon payload, is consistent with a similar event reported in a previous flight. These events may be produced by the atmospheric decay of an upward-propagating $τ$-lepton produced by a $ν_τ$ interaction, although their relatively steep arrival angles create tension with the standard model (SM) neutrino cross section. Each of the two events have $a~posteriori$ background estimates of $\lesssim 10^{-2}$ events. If these are generated by $τ$-lepton decay, then either the charged-current $ν_τ$ cross section is suppressed at EeV energies, or the events arise at moments when the peak flux of a transient neutrino source was much larger than the typical expected cosmogenic background neutrinos.
△ Less
Submitted 13 March, 2018;
originally announced March 2018.
-
Constraints on the diffuse high-energy neutrino flux from the third flight of ANITA
Authors:
P. W. Gorham,
P. Allison,
O. Banerjee,
L. Batten,
J. J. Beatty,
K. Bechtol,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. C. Chen,
C. H. Chen,
P. Chen,
J. M. Clem,
A. Connolly,
L. Cremonesi,
B. Dailey,
C. Deaconu,
P. F. Dowkontt,
B. D. Fox,
J. W. H. Gordon,
C. Hast,
B. Hill,
S. Y. Hsu
, et al. (35 additional authors not shown)
Abstract:
The Antarctic Impulsive Transient Antenna (ANITA), a NASA long-duration balloon payload, searches for radio emission from interactions of ultra-high-energy neutrinos in polar ice. The third flight of ANITA (ANITA-III) was launched in December 2014 and completed a 22-day flight. We present the results of three analyses searching for Askaryan radio emission of neutrino origin. In the most sensitive…
▽ More
The Antarctic Impulsive Transient Antenna (ANITA), a NASA long-duration balloon payload, searches for radio emission from interactions of ultra-high-energy neutrinos in polar ice. The third flight of ANITA (ANITA-III) was launched in December 2014 and completed a 22-day flight. We present the results of three analyses searching for Askaryan radio emission of neutrino origin. In the most sensitive of the analyses, we find one event in the signal region on an expected a priori background of $0.7^{+0.5}_{-0.3}$. Though consistent with the background estimate, the candidate event remains compatible with a neutrino hypothesis even after additional post-unblinding scrutiny.
△ Less
Submitted 18 June, 2018; v1 submitted 7 March, 2018;
originally announced March 2018.
-
Antarctic Surface Reflectivity Calculations and Measurements from the ANITA-4 and HiCal-2 Experiments
Authors:
S. Prohira,
A. Novikov,
P. Dasgupta,
P. Jain,
S. Nande,
P. Allison,
O. Banerjee,
L. Batten,
J. J. Beatty,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. Chen,
P. Chen,
J. M. Clem,
A. Connolly,
L. Cremonesi,
B. Dailey,
C. Deaconu,
P. F. Dowkontt,
B. D. Fox,
J. Gordon,
P. W. Gorham
, et al. (32 additional authors not shown)
Abstract:
The balloon-borne HiCal radio-frequency (RF) transmitter, in concert with the ANITA radio-frequency receiver array, is designed to measure the Antarctic surface reflectivity in the RF wavelength regime. The amplitude of surface-reflected transmissions from HiCal, registered as triggered events by ANITA, can be compared with the direct transmissions preceding them by O(10) microseconds, to infer th…
▽ More
The balloon-borne HiCal radio-frequency (RF) transmitter, in concert with the ANITA radio-frequency receiver array, is designed to measure the Antarctic surface reflectivity in the RF wavelength regime. The amplitude of surface-reflected transmissions from HiCal, registered as triggered events by ANITA, can be compared with the direct transmissions preceding them by O(10) microseconds, to infer the surface power reflection coefficient $\cal{R}$. The first HiCal mission (HiCal-1, Jan. 2015) yielded a sample of 100 such pairs, resulting in estimates of $\cal{R}$ at highly-glancing angles (i.e., zenith angles approaching $90^\circ$), with measured reflectivity for those events which exceeded extant calculations. The HiCal-2 experiment, flying from Dec., 2016-Jan., 2017, provided an improvement by nearly two orders of magnitude in our event statistics, allowing a considerably more precise map** of the reflectivity over a wider range of incidence angles. We find general agreement between the HiCal-2 reflectivity results and those obtained with the earlier HiCal-1 mission, as well as estimates from Solar reflections in the radio-frequency regime. In parallel, our calculations of expected reflectivity have matured; herein, we use a plane-wave expansion to estimate the reflectivity R from both a flat, smooth surface (and, in so doing, recover the Fresnel reflectivity equations) and also a curved surface. Multiplying our flat-smooth reflectivity by improved Earth curvature and surface roughness corrections now provides significantly better agreement between theory and the HiCal 2a/2b measurements.
△ Less
Submitted 26 January, 2018;
originally announced January 2018.
-
HiCal 2: An instrument designed for calibration of the ANITA experiment and for Antarctic surface reflectivity measurements
Authors:
S. Prohira,
A. Novikov,
D. Z. Besson,
K. Ratzlaff,
J. Stockham,
M. Stockham,
J. M. Clem,
R. Young,
P. W. Gorham,
P. Allison,
O. Banerjee,
L. Batten,
J. J. Beatty,
K. Belov,
W. R. Binns,
V. Bugaev,
P. Cao,
C. Chen,
P. Chen,
A. Connolly,
L. Cremonesi,
B. Dailey,
C. Deaconu,
P. F. Dowkontt,
B. D. Fox
, et al. (33 additional authors not shown)
Abstract:
The NASA supported High-Altitude Calibration (HiCal)-2 instrument flew as a companion balloon to the ANITA-4 experiment in December 2016. Based on a HV discharge pulser producing radio-frequency (RF) calibration pulses, HiCal-2 comprised two payloads, which flew for a combined 18 days, covering 1.5 revolutions of the Antarctic continent. ANITA-4 captured over 10,000 pulses from HiCal, both direct…
▽ More
The NASA supported High-Altitude Calibration (HiCal)-2 instrument flew as a companion balloon to the ANITA-4 experiment in December 2016. Based on a HV discharge pulser producing radio-frequency (RF) calibration pulses, HiCal-2 comprised two payloads, which flew for a combined 18 days, covering 1.5 revolutions of the Antarctic continent. ANITA-4 captured over 10,000 pulses from HiCal, both direct and reflected from the surface, at distances varying from 100-800 km, providing a large dataset for surface reflectivity measurements. Herein we present details on the design, construction and performance of HiCal-2.
△ Less
Submitted 17 September, 2020; v1 submitted 30 October, 2017;
originally announced October 2017.
-
Dynamic tunable notch filters for the Antarctic Impulsive Transient Antenna (ANITA)
Authors:
P. Allison,
O. Banerjee,
J. J. Beatty,
A. Connolly,
C. Deaconu,
J. Gordon,
P. W. Gorham,
M. Kovacevich,
C. Miki,
E. Oberla,
J. Roberts,
B. Rotter,
S. Stafford,
K. Tatem,
L. Batten,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. Chen,
P. Chen,
Y. Chen,
J. M. Clem,
L. Cremonesi
, et al. (34 additional authors not shown)
Abstract:
The Antarctic Impulsive Transient Antenna (ANITA) is a NASA long-duration balloon experiment with the primary goal of detecting ultra-high-energy ($>10^{18}\,\mbox{eV}$) neutrinos via the Askaryan Effect. The fourth ANITA mission, ANITA-IV, recently flew from Dec 2 to Dec 29, 2016. For the first time, the Tunable Universal Filter Frontend (TUFF) boards were deployed for mitigation of narrow-band,…
▽ More
The Antarctic Impulsive Transient Antenna (ANITA) is a NASA long-duration balloon experiment with the primary goal of detecting ultra-high-energy ($>10^{18}\,\mbox{eV}$) neutrinos via the Askaryan Effect. The fourth ANITA mission, ANITA-IV, recently flew from Dec 2 to Dec 29, 2016. For the first time, the Tunable Universal Filter Frontend (TUFF) boards were deployed for mitigation of narrow-band, anthropogenic noise with tunable, switchable notch filters. The TUFF boards also performed second-stage amplification by approximately 45 dB to boost the $\sim\,μ\mbox{V-level}$ radio frequency (RF) signals to $\sim$ mV-level for digitization, and supplied power via bias tees to the first-stage, antenna-mounted amplifiers. The other major change in signal processing in ANITA-IV is the resurrection of the $90^{\circ}$ hybrids deployed previously in ANITA-I, in the trigger system, although in this paper we focus on the TUFF boards. During the ANITA-IV mission, the TUFF boards were successfully operated throughout the flight. They contributed to a factor of 2.8 higher total instrument livetime on average in ANITA-IV compared to ANITA-III due to reduction of narrow-band, anthropogenic noise before a trigger decision is made.
△ Less
Submitted 13 September, 2017;
originally announced September 2017.
-
Antarctic Surface Reflectivity Measurements from the ANITA-3 and HiCal-1 Experiments
Authors:
P. W. Gorham,
P. Allison,
O. Banerjee,
J. J. Beatty,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. Chen,
P. Chen,
J. M. Clem,
A. Connolly,
B. Dailey,
P. Dasgupta,
C. Deaconu,
L. Cremonesi,
P. F. Dowkontt,
B. D. Fox,
J. Gordon,
B. Hill,
R. Hupe,
M. H. Israel,
P. Jain,
J. Kowalski
, et al. (31 additional authors not shown)
Abstract:
The primary science goal of the NASA-sponsored ANITA project is measurement of ultra-high energy neutrinos and cosmic rays, observed via radio-frequency signals resulting from a neutrino- or cosmic ray- interaction with terrestrial matter (atmospheric or ice molecules, e.g.). Accurate inference of the energies of these cosmic rays requires understanding the transmission/reflection of radio wave si…
▽ More
The primary science goal of the NASA-sponsored ANITA project is measurement of ultra-high energy neutrinos and cosmic rays, observed via radio-frequency signals resulting from a neutrino- or cosmic ray- interaction with terrestrial matter (atmospheric or ice molecules, e.g.). Accurate inference of the energies of these cosmic rays requires understanding the transmission/reflection of radio wave signals across the ice-air boundary. Satellite-based measurements of Antarctic surface reflectivity, using a co-located transmitter and receiver, have been performed more-or-less continuously for the last few decades. Satellite-based reflectivity surveys, at frequencies ranging from 2--45 GHz and at near-normal incidence, yield generally consistent reflectivity maps across Antarctica. Using the Sun as an RF source, and the ANITA-3 balloon borne radio-frequency antenna array as the RF receiver, we have also measured the surface reflectivity over the interval 200-1000 MHz, at elevation angles of 12-30 degrees, finding agreement with the Fresnel equations within systematic errors. To probe low incidence angles, inaccessible to the Antarctic Solar technique and not probed by previous satellite surveys, a novel experimental approach ("HiCal-1") was devised. Unlike previous measurements, HiCal-ANITA constitute a bi-static transmitter-receiver pair separated by hundreds of kilometers. Data taken with HiCal, between 200--600 MHz shows a significant departure from the Fresnel equations, constant with frequency over that band, with the deficit increasing with obliquity of incidence, which we attribute to the combined effects of possible surface roughness, surface grain effects, radar clutter and/or shadowing of the reflection zone due to Earth curvature effects.
△ Less
Submitted 21 June, 2017; v1 submitted 1 March, 2017;
originally announced March 2017.
-
Characteristics of Four Upward-pointing Cosmic-ray-like Events Observed with ANITA
Authors:
P. W. Gorham,
J. Nam,
A. Romero-Wolf,
S. Hoover,
P. Allison,
O. Banerjee,
J. J. Beatty,
K. Belov,
D. Z. Besson,
W. R. Binns,
V. Bugaev,
P. Cao,
C. Chen,
P. Chen,
J. M. Clem,
A. Connolly,
B. Dailey,
C. Deaconu,
L. Cremonesi,
P. F. Dowkonnt,
M. A. Duvernois,
R. C. Field,
B. D. Fox,
D. Goldstein,
J. Gordon
, et al. (44 additional authors not shown)
Abstract:
We report on four radio-detected cosmic-ray (CR) or CR-like events observed with the Antarctic Impulsive Transient Antenna (ANITA), a NASA-sponsored long-duration balloon payload. Two of the four were previously identified as stratospheric CR air showers during the ANITA-I flight. A third stratospheric CR was detected during the ANITA-II flight. Here we report on characteristics these three unusua…
▽ More
We report on four radio-detected cosmic-ray (CR) or CR-like events observed with the Antarctic Impulsive Transient Antenna (ANITA), a NASA-sponsored long-duration balloon payload. Two of the four were previously identified as stratospheric CR air showers during the ANITA-I flight. A third stratospheric CR was detected during the ANITA-II flight. Here we report on characteristics these three unusual CR events, which develop nearly horizontally, 20-30~km above the surface of the Earth. In addition, we report on a fourth steeply upward-pointing ANITA-I CR-like radio event which has characteristics consistent with a primary that emerged from the surface of the ice. This suggests a possible $τ$-lepton decay as the origin of this event, but such an interpretation would require significant suppression of the Standard Model $τ$-neutrino cross section.
△ Less
Submitted 29 June, 2016; v1 submitted 16 March, 2016;
originally announced March 2016.
-
An empirical comparative study of approximate methods for binary graphical models; application to the search of associations among causes of death in French death certificates
Authors:
Vivian Viallon,
Onureena Banerjee,
Gregoire Rey,
Eric Jougla,
Joel Coste
Abstract:
Looking for associations among multiple variables is a topical issue in statistics due to the increasing amount of data encountered in biology, medicine and many other domains involving statistical applications. Graphical models have recently gained popularity for this purpose in the statistical literature. Following the ideas of the LASSO procedure designed for the linear regression framework, re…
▽ More
Looking for associations among multiple variables is a topical issue in statistics due to the increasing amount of data encountered in biology, medicine and many other domains involving statistical applications. Graphical models have recently gained popularity for this purpose in the statistical literature. Following the ideas of the LASSO procedure designed for the linear regression framework, recent developments dealing with graphical model selection have been based on $\ell_1$-penalization. In the binary case, however, exact inference is generally very slow or even intractable because of the form of the so-called log-partition function. Various approximate methods have recently been proposed in the literature and the main objective of this paper is to compare them. Through an extensive simulation study, we show that a simple modification of a method relying on a Gaussian approximation achieves good performance and is very fast. We present a real application in which we search for associations among causes of death recorded on French death certificates.
△ Less
Submitted 13 April, 2010;
originally announced April 2010.
-
Model Selection Through Sparse Maximum Likelihood Estimation
Authors:
Onureena Banerjee,
Laurent El Ghaoui,
Alexandre d'Aspremont
Abstract:
We consider the problem of estimating the parameters of a Gaussian or binary distribution in such a way that the resulting undirected graphical model is sparse. Our approach is to solve a maximum likelihood problem with an added l_1-norm penalty term. The problem as formulated is convex but the memory requirements and complexity of existing interior point methods are prohibitive for problems wit…
▽ More
We consider the problem of estimating the parameters of a Gaussian or binary distribution in such a way that the resulting undirected graphical model is sparse. Our approach is to solve a maximum likelihood problem with an added l_1-norm penalty term. The problem as formulated is convex but the memory requirements and complexity of existing interior point methods are prohibitive for problems with more than tens of nodes. We present two new algorithms for solving problems with at least a thousand nodes in the Gaussian case. Our first algorithm uses block coordinate descent, and can be interpreted as recursive l_1-norm penalized regression. Our second algorithm, based on Nesterov's first order method, yields a complexity estimate with a better dependence on problem size than existing interior point methods. Using a log determinant relaxation of the log partition function (Wainwright & Jordan (2006)), we show that these same algorithms can be used to solve an approximate sparse maximum likelihood problem for the binary case. We test our algorithms on synthetic data, as well as on gene expression and senate voting records data.
△ Less
Submitted 4 July, 2007;
originally announced July 2007.
-
First-order methods for sparse covariance selection
Authors:
Alexandre d'Aspremont,
Onureena Banerjee,
Laurent El Ghaoui
Abstract:
Given a sample covariance matrix, we solve a maximum likelihood problem penalized by the number of nonzero coefficients in the inverse covariance matrix. Our objective is to find a sparse representation of the sample data and to highlight conditional independence relationships between the sample variables. We first formulate a convex relaxation of this combinatorial problem, we then detail two e…
▽ More
Given a sample covariance matrix, we solve a maximum likelihood problem penalized by the number of nonzero coefficients in the inverse covariance matrix. Our objective is to find a sparse representation of the sample data and to highlight conditional independence relationships between the sample variables. We first formulate a convex relaxation of this combinatorial problem, we then detail two efficient first-order algorithms with low memory requirements to solve large-scale, dense problem instances.
△ Less
Submitted 28 September, 2006;
originally announced September 2006.
-
Sparse Covariance Selection via Robust Maximum Likelihood Estimation
Authors:
Onureena Banerjee,
Alexandre d'Aspremont,
Laurent El Ghaoui
Abstract:
We address a problem of covariance selection, where we seek a trade-off between a high likelihood against the number of non-zero elements in the inverse covariance matrix. We solve a maximum likelihood problem with a penalty term given by the sum of absolute values of the elements of the inverse covariance matrix, and allow for imposing bounds on the condition number of the solution. The problem…
▽ More
We address a problem of covariance selection, where we seek a trade-off between a high likelihood against the number of non-zero elements in the inverse covariance matrix. We solve a maximum likelihood problem with a penalty term given by the sum of absolute values of the elements of the inverse covariance matrix, and allow for imposing bounds on the condition number of the solution. The problem is directly amenable to now standard interior-point algorithms for convex optimization, but remains challenging due to its size. We first give some results on the theoretical computational complexity of the problem, by showing that a recent methodology for non-smooth convex optimization due to Nesterov can be applied to this problem, to greatly improve on the complexity estimate given by interior-point algorithms. We then examine two practical algorithms aimed at solving large-scale, noisy (hence dense) instances: one is based on a block-coordinate descent approach, where columns and rows are updated sequentially, another applies a dual version of Nesterov's method.
△ Less
Submitted 8 June, 2005;
originally announced June 2005.
