arXiv subjects

Brandon Anderson

Publications and source records attributed to Brandon Anderson.

At least 19 recordsLinked to original sources

arXiv2022

Integrating Reward Maximization and Population Estimation: Sequential Decision-Making for Internal Revenue Service Audit Selection

We introduce a new setting, optimize-and-estimate structured bandits. Here, a policy must select a batch of arms, each characterized by its own context, that would allow it to both maximize reward and maintain an accurate (ideally unbiased) population estimate of the reward. This setting is inherent to many public and private sector applications and often requires handling delayed feedback, small data, and distribution shifts. We demonstrate its importance on real data from the United States Internal Revenue Service (IRS). The IRS performs yearly audits of the tax base. Two of its most important objectives are to identify suspected misreporting and to estimate the "tax gap" -- the global difference between the amount paid and true amount owed. Based on a unique collaboration with the IRS, we cast these two processes as a unified optimize-and-estimate structured bandit. We analyze optimize-and-estimate approaches to the IRS problem and propose a novel mechanism for unbiased population estimation that achieves rewards comparable to baseline approaches. This approach has the potential to improve audit efficacy, while maintaining policy-relevant estimates of the tax gap. This has important social consequences given that the current tax gap is estimated at nearly half a trillion dollars. We suggest that this problem setting is fertile ground for further research and we highlight its interesting challenges. The results of this and related research are currently being incorporated into the continual improvement of the IRS audit selection methods.

cs.LG↗

arXiv2021

Mapping industrial poultry operations at scale with deep learning and aerial imagery

Concentrated Animal Feeding Operations (CAFOs) pose serious risks to air, water, and public health, but have proven to be challenging to regulate. The U.S. Government Accountability Office notes that a basic challenge is the lack of comprehensive location information on CAFOs. We use the USDA's National Agricultural Imagery Program (NAIP) 1m/pixel aerial imagery to detect poultry CAFOs across the continental United States. We train convolutional neural network (CNN) models to identify individual poultry barns and apply the best performing model to over 42 TB of imagery to create the first national, open-source dataset of poultry CAFOs. We validate the model predictions against held-out validation set on poultry CAFO facility locations from 10 hand-labeled counties in California and demonstrate that this approach has significant potential to fill gaps in environmental monitoring.

cs.CV↗

arXiv2021

Beyond Ads: Sequential Decision-Making Algorithms in Law and Public Policy

We explore the promises and challenges of employing sequential decision-making algorithms -- such as bandits, reinforcement learning, and active learning -- in law and public policy. While such algorithms have well-characterized performance in the private sector (e.g., online advertising), the tendency to naively apply algorithms motivated by one domain, often online advertisements, can be called the "advertisement fallacy." Our main thesis is that law and public policy pose distinct methodological challenges that the machine learning community has not yet addressed. Machine learning will need to address these methodological problems to move "beyond ads." Public law, for instance, can pose multiple objectives, necessitate batched and delayed feedback, and require systems to learn rational, causal decision-making policies, each of which presents novel questions at the research frontier. We discuss a wide range of potential applications of sequential decision-making algorithms in regulation and governance, including public health, environmental protection, tax administration, occupational safety, and benefits adjudication. We use these examples to highlight research needed to render sequential decision making policy-compliant, adaptable, and effective in the public sector. We also note the potential risks of such deployments and describe how sequential decision systems can also facilitate the discovery of harms. We hope our work inspires more investigation of sequential decision making in law and public policy, which provide unique challenges for machine learning researchers with potential for significant social benefit.

cs.CY↗

arXiv2021

Learning physics confers pose-sensitivity in structure-based virtual screening

In drug discovery, structure-based virtual high-throughput screening (vHTS) campaigns aim to identify bioactive ligands or "hits" for therapeutic protein targets from docked poses at specific binding sites. However, while generally successful at this task, many deep learning methods are known to be insensitive to protein-ligand interactions, decreasing the reliability of hit detection and hindering discovery at novel binding sites. Here, we overcome this limitation by introducing a class of models with two key features: 1) we condition bioactivity on pose quality score, and 2) we present poor poses of true binders to the model as negative examples. The conditioning forces the model to learn details of physical interactions. We evaluate these models on a new benchmark designed to detect pose-sensitivity.

q-bio.QM↗

arXiv2021

Enhancing Environmental Enforcement with Near Real-Time Monitoring: Likelihood-Based Detection of Structural Expansion of Intensive Livestock Farms

Much environmental enforcement in the United States has historically relied on either self-reported data or physical, resource-intensive, infrequent inspections. Advances in remote sensing and computer vision, however, have the potential to augment compliance monitoring by detecting early warning signs of noncompliance. We demonstrate a process for rapid identification of significant structural expansion using Planet's 3m/pixel satellite imagery products and focusing on Concentrated Animal Feeding Operations (CAFOs) in the US as a test case. Unpermitted building expansion has been a particular challenge with CAFOs, which pose significant health and environmental risks. Using new hand-labeled dataset of 145,053 images of 1,513 CAFOs, we combine state-of-the-art building segmentation with a likelihood-based change-point detection model to provide a robust signal of building expansion (AUC = 0.86). A major advantage of this approach is that it can work with higher cadence (daily to weekly), but lower resolution (3m/pixel), satellite imagery than previously used in similar environmental settings. It is also highly generalizable and thus provides a near real-time monitoring tool to prioritize enforcement resources in other settings where unpermitted construction poses environmental risk, e.g. zoning, habitat modification, or wetland protection.

cs.CV↗

arXiv2021

Temporal Cluster Matching for Change Detection of Structures from Satellite Imagery

Longitudinal studies are vital to understanding dynamic changes of the planet, but labels (e.g., buildings, facilities, roads) are often available only for a single point in time. We propose a general model, Temporal Cluster Matching (TCM), for detecting building changes in time series of remotely sensed imagery when footprint labels are observed only once. The intuition behind the model is that the relationship between spectral values inside and outside of building's footprint will change when a building is constructed (or demolished). For instance, in rural settings, the pre-construction area may look similar to the surrounding environment until the building is constructed. Similarly, in urban settings, the pre-construction areas will look different from the surrounding environment until construction. We further propose a heuristic method for selecting the parameters of our model which allows it to be applied in novel settings without requiring data labeling efforts (to fit the parameters). We apply our model over a dataset of poultry barns from 2016/2017 high-resolution aerial imagery in the Delmarva Peninsula and a dataset of solar farms from a 2020 mosaic of Sentinel 2 imagery in India. Our results show that our model performs as well when fit using the proposed heuristic as it does when fit with labeled data, and further, that supervised versions of our model perform the best among all the baselines we test against. Finally, we show that our proposed approach can act as an effective data augmentation strategy -- it enables researchers to augment existing structure footprint labels along the time dimension and thus use imagery from multiple points in time to train deep learning models. We show that this improves the spatial generalization of such models when evaluated on the same change detection task.

cs.CV↗

arXiv2020

ATOM3D: Tasks On Molecules in Three Dimensions

Computational methods that operate on three-dimensional molecular structure have the potential to solve important questions in biology and chemistry. In particular, deep neural networks have gained significant attention, but their widespread adoption in the biomolecular domain has been limited by a lack of either systematic performance benchmarks or a unified toolkit for interacting with molecular data. To address this, we present ATOM3D, a collection of both novel and existing benchmark datasets spanning several key classes of biomolecules. We implement several classes of three-dimensional molecular learning methods for each of these tasks and show that they consistently improve performance relative to methods based on one- and two-dimensional representations. The specific choice of architecture proves to be critical for performance, with three-dimensional convolutional networks excelling at tasks involving complex geometries, graph networks performing well on systems requiring detailed positional information, and the more recently developed equivariant networks showing significant promise. Our results indicate that many molecular problems stand to gain from three-dimensional molecular learning, and that there is potential for improvement on many tasks which remain underexplored. To lower the barrier to entry and facilitate further developments in the field, we also provide a comprehensive suite of tools for dataset processing, model training, and evaluation in our open-source atom3d Python package. All datasets are available for download from https://www.atom3d.ai .

cs.LG↗

arXiv2020

A community-powered search of machine learning strategy space to find NMR property prediction models

The rise of machine learning (ML) has created an explosion in the potential strategies for using data to make scientific predictions. For physical scientists wishing to apply ML strategies to a particular domain, it can be difficult to assess in advance what strategy to adopt within a vast space of possibilities. Here we outline the results of an online community-powered effort to swarm search the space of ML strategies and develop algorithms for predicting atomic-pairwise nuclear magnetic resonance (NMR) properties in molecules. Using an open-source dataset, we worked with Kaggle to design and host a 3-month competition which received 47,800 ML model predictions from 2,700 teams in 84 countries. Within 3 weeks, the Kaggle community produced models with comparable accuracy to our best previously published "in-house" efforts. A meta-ensemble model constructed as a linear combination of the top predictions has a prediction accuracy which exceeds that of any individual model, 7-19x better than our previous state-of-the-art. The results highlight the potential of transformer architectures for predicting quantum mechanical (QM) molecular properties.

physics.comp-ph↗

arXiv2020

Lorentz Group Equivariant Neural Network for Particle Physics

We present a neural network architecture that is fully equivariant with respect to transformations under the Lorentz group, a fundamental symmetry of space and time in physics. The architecture is based on the theory of the finite-dimensional representations of the Lorentz group and the equivariant nonlinearity involves the tensor product. For classification tasks in particle physics, we demonstrate that such an equivariant architecture leads to drastically simpler models that have relatively few learnable parameters and are much more physically interpretable than leading approaches that use CNNs and point cloud approaches. The competitive performance of the network is demonstrated on a public classification dataset [27] for tagging top quark decays given energy-momenta of jet constituents produced in proton-proton collisions.

hep-ph↗

arXiv2019

Cormorant: Covariant Molecular Neural Networks

We propose Cormorant, a rotationally covariant neural network architecture for learning the behavior and properties of complex many-body physical systems. We apply these networks to molecular systems with two goals: learning atomic potential energy surfaces for use in Molecular Dynamics simulations, and learning ground state properties of molecules calculated by Density Functional Theory. Some of the key features of our network are that (a) each neuron explicitly corresponds to a subset of atoms; (b) the activation of each neuron is covariant to rotations, ensuring that overall the network is fully rotationally invariant. Furthermore, the non-linearity in our network is based upon tensor products and the Clebsch-Gordan decomposition, allowing the network to operate entirely in Fourier space. Cormorant significantly outperforms competing algorithms in learning molecular Potential Energy Surfaces from conformational geometries in the MD-17 dataset, and is competitive with other methods at learning geometric, energetic, electronic, and thermodynamic properties of molecules on the GDB-9 dataset.

physics.comp-ph↗

arXiv2018

Covariant Compositional Networks For Learning Graphs

Most existing neural networks for learning graphs address permutation invariance by conceiving of the network as a message passing scheme, where each node sums the feature vectors coming from its neighbors. We argue that this imposes a limitation on their representation power, and instead propose a new general architecture for representing objects consisting of a hierarchy of parts, which we call Covariant Compositional Networks (CCNs). Here, covariance means that the activation of each neuron must transform in a specific way under permutations, similarly to steerability in CNNs. We achieve covariance by making each activation transform according to a tensor representation of the permutation group, and derive the corresponding tensor aggregation rules that each neuron must implement. Experiments show that CCNs can outperform competing methods on standard graph learning benchmarks.

cs.LG↗

arXiv2016

Sensitivity Projections for Dark Matter Searches with the Fermi Large Area Telescope

The nature of dark matter is a longstanding enigma of physics; it may consist of particles beyond the Standard Model that are still elusive to experiments. Among indirect search techniques, which look for stable products from the annihilation or decay of dark matter particles, or from axions coupling to high-energy photons, observations of the $\gamma$-ray sky have come to prominence over the last few years, because of the excellent sensitivity of the Large Area Telescope (LAT) on the Fermi Gamma-ray Space Telescope mission. The LAT energy range from 20 MeV to above 300 GeV is particularly well suited for searching for products of the interactions of dark matter particles. In this report we describe methods used to search for evidence of dark matter with the LAT, and review the status of searches performed with up to six years of LAT data. We also discuss the factors that determine the sensitivities of these searches, including the magnitudes of the signals and the relevant backgrounds, considering both statistical and systematic uncertainties. We project the expected sensitivities of each search method for 10 and 15 years of LAT data taking. In particular, we find that the sensitivity of searches targeting dwarf galaxies, which provide the best limits currently, will improve faster than the square root of observing time. Current LAT limits for dwarf galaxies using six years of data reach the thermal relic level for masses up to 120 GeV for the $b\bar{b}$ annihilation channel for reasonable dark matter density profiles. With projected discoveries of additional dwarfs, these limits could extend to about 250 GeV. With as much as 15 years of LAT data these searches would be sensitive to dark matter annihilations at the thermal relic cross section for masses to greater than 400 GeV (200 GeV) in the $b\bar{b}$ ($\tau^+ \tau^-$) annihilation channels.

astro-ph.HE↗

arXiv2016

On methods for correcting for the look-elsewhere effect in searches for new physics

The search for new significant peaks over a energy spectrum often involves a statistical multiple hypothesis testing problem. Separate tests of hypothesis are conducted at different locations producing an ensemble of local p-values, the smallest of which is reported as evidence for the new resonance. Unfortunately, controlling the false detection rate (type I error rate) of such procedures may lead to excessively stringent acceptance criteria. In the recent physics literature, two promising statistical tools have been proposed to overcome these limitations. In 2005, a method to "find needles in haystacks" was introduced by Pilla et al. [1], and a second method was later proposed by Gross and Vitells [2] in the context of the "look elsewhere effect" and trial factors. We show that, for relatively small sample sizes, the former leads to an artificial inflation of statistical power that stems from an increase in the false detection rate, whereas the two methods exhibit similar performance for large sample sizes. We apply the methods to realistic simulations of the Fermi Large Area Telescope data, in particular the search for dark matter annihilation lines. Further, we discuss the counter-intutive scenario where the look-elsewhere corrections are more conservative than much more computationally efficient corrections for multiple hypothesis testing. Finally, we provide general guidelines for navigating the tradeoffs between statistical and computational efficiency when selecting a statistical procedure for signal detection.

physics.data-an↗

arXiv2015

Dark Matter Searches with the Fermi-LAT in the Direction of Dwarf Spheroidals

The dwarf spheroidal satellite galaxies of the Milky Way are some of the most dark-matter-dominated objects known. Due to their proximity, high dark matter content, and lack of astrophysical backgrounds, dwarf spheroidal galaxies are widely considered to be among the most promising targets for the indirect detection of dark matter via gamma rays. Here we report on gamma-ray observations of Milky Way dwarf spheroidal satellite galaxies based on 6 years of Fermi Large Area Telescope data processed with the new Pass 8 reconstruction and event-level analysis. None of the dwarf galaxies are significantly detected in gamma rays, and we present upper limits on the dark matter annihilation cross section from a combined analysis of the 15 most promising dwarf galaxies. The constraints derived are among the strongest to date using gamma rays and lie below the canonical thermal relic cross section for WIMPs of mass $\lesssim 100~GeV$ annihilating via the $b \bar b$ and $\tau^{+}\tau^{-}$ channels.

astro-ph.HE↗

arXiv2015

Using Likelihood for Combined Data Set Analysis

The joint likelihood is a simple extension of the standard likelihood formalism that enables the estimation of common parameters across disjoint datasets. Joining the likelihood, rather than the data itself, means nuisance parameters can be dealt with independently. Application of this technique, particularly to Fermi-LAT dwarf spheroidal analyses, has already been met with great success. We present a description of the method's general implementation along with a toy Monte-Carlo study of its properties and limitations.

astro-ph.HE↗

arXiv2012

Chaos-driven dynamics in spin-orbit coupled atomic gases

The dynamics, appearing after a quantum quench, of a trapped, spin-orbit coupled, dilute atomic gas is studied. The characteristics of the evolution is greatly influenced by the symmetries of the system, and we especially compare evolution for an isotropic Rashba coupling and for an anisotropic spin-orbit coupling. As we make the spin-orbit coupling anisotropic, we break the rotational symmetry and the underlying classical model becomes chaotic; the quantum dynamics is affected accordingly. Within experimentally relevant time-scales and parameters, the system thermalizes in a quantum sense. The corresponding equilibration time is found to agree with the Ehrenfest time, i.e. we numerically verify a ~log(1/h) scaling. Upon thermalization, we find the equilibrated distributions show examples of quantum scars distinguished by accumulation of atomic density for certain energies. At shorter time-scales we discuss non-adiabatic effects deriving from the spin-orbit coupled induced Dirac point. In the vicinity of the Dirac point, spin fluctuations are large and, even at short times, a semi-classical analysis fails.

quant-ph↗

arXiv2010

Fermi-LAT constraints on diffuse Dark Matter annihilation from the Galactic Halo

Our Galaxy resides in the center of a vast "Halo" of Dark Matter (DM). This concentration produces, in many viable particle physics models, an indirect Weakly Interacting Massive Particle (WIMP) annihilation signal that peaks in the Fermi-LAT's energy range. Our knowledge of the diffuse background is essential to placing reasonable limits on the DM mass and cross-section. We incorporate a systematic variation of the GALPROP galactic diffuse background model, constrained by current cosmic-ray measurements, into a profile likelihood analysis and present preliminary upper limits on the DM annihilation cross-section using the Fermi-LAT data.

hep-ph↗

arXiv2010

Fermi-LAT Sensitivity to Dark Matter Annihilation in Via Lactea II Substructure

We present a study of the ability of the Fermi Gamma-ray Space Telescope to detect dark-matter annihilation signals from the Galactic subhalos predicted by the Via Lactea II N-body simulation. We implement an improved formalism for estimating the boost factor needed to account for the effect of dark-matter clumping on scales below the resolution of the simulation, and we incorporate a detailed Monte Carlo simulation of the response of the Fermi-LAT telescope, including a simulation of its all-sky observing mode integrated over a ten year mission. We find that for WIMP masses up to about 150 GeV in standard supersymmetric models with velocity-averaged cross section 3*10^-26 cm^3 s^-1, a few subhalos could be detectable with >5 standard deviations significance and would likely deviate significantly from the appearance of a point source.

astro-ph.HE↗