arXiv subjects

David Shih

Publications and source records attributed to David Shih.

At least 19 recordsLinked to original sources

arXiv2026

Anomaly detection for multijet scenarios

Signals of physics beyond the Standard Model continue to resist discovery at the LHC. Recent years have seen the proliferation of new anomaly detection techniques, promising discovery with significantly fewer model assumptions than traditional approaches. Strategies based on weak supervision have been especially successful, but so far were largely limited in scope by their reliance on a resonance manifesting a decay into a pair of jets. In this work, we demonstrate that a well-established idea from jet substructure physics - recursive soft drop - in combination with the CATHODE technique for anomaly detection can be used to simultaneously perform anomaly detection for signals with an arbitrary number of jets in the final state, greatly increasing the scope of such searches.

hep-ph↗

arXiv2026

A Universal Distribution of Dark Matter in Milky Way-like galaxies and How to Infer It

The phase-space density of dark matter within the Milky Way is a key quantity that encodes information about the nature of the dark sector. The local phase-space density is also required to properly interpret the results of dark matter direct detection experiments. However, there are at present few observational constraints. In this paper, we show that a simple coordinate transformation reveals a near-universal DM phase-space distribution function among three independent suites of cosmological simulations of Milky Way-mass galaxies. We provide evidence for this with plots of kinematic features as well as machine learning-based classifiers that are sensitive to all of the correlations in the full multivariate phase space. Deviations from universality are found only at extremes of galactic radius and/or velocity, and in one simulation that has a prominent accreted dark disc. We further show that the parameters for the coordinate transformation can be inferred from metal poor stars ($\log_{10}[{\rm Fe}/{\rm H}]< -2$). These stars also contain signatures of the dark disk, allowing the existence of such a structure to be inferred from observation. Finally, we construct a model of this universal phase-space distribution using a normalizing flow, trained on the standardized phase-space across simulations. We will apply our method to survey data from Gaia and SDSS in a forthcoming work.

astro-ph.GA↗

arXiv2026

Hiding in the Shadow of the Upsilon: Ditaus from a Light Pseudoscalar

The CMS collaboration has reported a measurement of $\Upsilon$ decays to ditaus using $61.9~{\rm fb}^{-1}$ of scouting data. If interpreted as the decay of $\Upsilon(1S,2S,3S)$, the measured ditau rate is more than ten times that seen in the dimuon final states at the $\sim 3 \sigma$ level, and is likewise inconsistent with the branching ratios measured at $B$-factories. If confirmed with more data and at higher significance, such a violation of lepton flavor universality would necessitate new physics. In this Letter, we present a simple model with a light pseudoscalar coincidentally near the $\Upsilon(1S)$ mass, which mixes with two Higgs doublets in the alignment limit. Such a particle naturally decays primarily to taus and evades all existing experimental constraints, while implying a number of predictions that can be tested in the near future.

hep-ph↗

arXiv2026

Neural Scaling Laws for Jet Generation

Recently observed empirical scaling laws describe the performance of foundation-type models as three independent key quantities -- dataset size, compute, and model parameters -- are modified. Extracting these scaling laws informs the training of large complex models for which the tuning of hyperparameters in traditional ways is not feasible. This work for the first time explores if scaling laws can also be observed for the task of particle jet generation -- both relevant as a pre-training objective for foundation models and as in-situ simulation by itself. We indeed replicate the key logarithmic scaling law behavior for model-size scaling. Beyond studying the next token prediction validation loss of the generative model, we also study the sliced Wasserstein distance of five physical quantities that are not immediately available to the model during training. Our study shows that this quantity is monotonically related to the next token prediction validation loss, meaning that this loss is indeed a good proxy for the physics performance. For the scaling with dataset size and compute, we observe substantially weaker scaling behavior of both the loss and the sliced Wasserstein distance. We analyze this behavior by introducing the concept of a learnable window, and argue that autoregressive next token prediction on jet constituents exhibits comparatively rapid saturation relative to language-model studies. We discuss possible origins of this behavior, including the stochastic nature of QCD radiation and differences between generative and supervised learning tasks in collider physics.

hep-ph↗

arXiv2026

Collider-Bench: Benchmarking AI Agents with Particle Physics Analysis Reproduction

Autonomous language-model agents are increasingly evaluated on long-horizon tool-use tasks, but existing benchmarks rarely capture the complexity and nuance of real scientific work. To address this gap, we introduce Collider-Bench, a benchmark for evaluating whether LLM agents can reproduce experimental analyses from the Large Hadron Collider (LHC) using only public papers and open scientific software. Such analyses are often difficult to reproduce because the public toolchain only approximates the software used internally by the experimental collaborations, while the published papers inevitably omit implementation details needed for a faithful reconstruction. Agents must therefore rely on physical reasoning, domain knowledge, and trial-and-error to fill these gaps. Each task requires the agent to turn a published analysis into an executable simulation-and-selection pipeline and submit predicted collision event yields in specified signal regions. These predictions are evaluated with standard histogram metrics that provide continuous fidelity scores without a hand-written rubric. We also report the computational cost incurred by each agent per task. Finally, we evaluate the codebase and full session trace using an LLM judge to catch qualitative failure modes such as fabrications, hallucinations and duplications. We release an initial set of tasks drawn from LHC searches, together with a containerized sandbox and event simulation tools. We evaluate across a capability ladder of general purpose coding agents. Our results show that on average no agent reliably beats the physicist-in-the-loop solution.

cs.LG↗

arXiv2026

Kitchen Sink Anomaly Detection

An enormous amount of R&D effort has resulted in many new resonant anomaly detection methods being proposed in recent years. However, the vast majority of previous R&D studies have suffered from two limitations: they have focused on a very small set of simulated signal benchmark models; and they have either used small sets of carefully crafted high-level jet substructure observables, which can be highly performant but are prone to model dependence, or the full collider event phase space, which is more agnostic but suffers from reduced sensitivity. In this work, we address both limitations: we formulate a number of new simulated signal benchmarks, which we make publicly available in a format fully compatible with the LHCO R&D benchmark; and we explore a high-level, yet highly agnostic, observable set consisting of Energy Flow Polynomials in addition to the usual subjettiness variables. We evaluate this "kitchen sink" observable set for both an idealized anomaly detector and the CWoLa hunting task, along with three baseline observable sets (the Baseline LHC Olympics set, subjettiness observables, and Energy Flow Polynomials). We find that our kitchen sink approach is the most sensitive to a broad range of signal types. Furthermore, we show that an attribute bagging variant, in which each ensemble member is trained on a random subset of substructure observables, yields comparable anomaly detection performance while significantly reducing training cost.

hep-ph↗

arXiv2026

Learning to Unscramble Feynman Loop Integrals with SAILIR

Integration-by-parts (IBP) reduction of Feynman integrals to master integrals is a key computational bottleneck in precision calculations in high-energy physics. Traditional approaches based on the Laporta algorithm require solving large systems of equations, leading to memory consumption that grows rapidly with integral complexity. We present SAILIR (Self-supervised AI for Loop Integral Reduction), a new machine learning approach in which a transformer-based classifier guides the reduction of integrals one step at a time in a fully online fashion. The classifier is trained in an entirely self-supervised manner on synthetic data generated by a scramble/unscramble procedure: known reduction identities are applied in reverse to build expressions of increasing complexity, and the classifier learns to undo these steps. When combined with beam search and a highly parallelized, asynchronous, single-episode reduction strategy, SAILIR can reduce integrals of arbitrarily high weight with bounded memory. We benchmark SAILIR on the two-loop triangle-box topology, comparing against the state-of-the-art IBP reduction code Kira across 16 integrals of varying complexity. While SAILIR is slower in wall-clock time, its per-worker memory consumption remains approximately flat regardless of integral complexity, in contrast to Kira whose memory grows rapidly with complexity. For the most complex integrals considered here, SAILIR uses only 40\% of the memory of Kira while achieving comparable reduction times. This demonstrates a fundamentally new paradigm for IBP reduction in which the memory bottleneck of Laporta-based approaches could be entirely overcome, potentially opening the door to precision calculations that are currently intractable.

hep-ph↗

arXiv2026

Learning to Unscramble: Simplifying Symbolic Expressions via Self-Supervised Oracle Trajectories

We present a new self-supervised machine learning approach for symbolic simplification of complex mathematical expressions. Training data is generated by scrambling simple expressions and recording the inverse operations, creating oracle trajectories that provide both goal states and explicit paths to reach them. A permutation-equivariant, transformer-based policy network is then trained on this data step-wise to predict the oracle action given the input expression. We demonstrate this approach on two problems in high-energy physics: dilogarithm reduction and spinor-helicity scattering amplitude simplification. In both cases, our trained policy network achieves near perfect solve rates across a wide range of difficulty levels, substantially outperforming prior approaches based on reinforcement learning and end-to-end regression. When combined with contrastive grouping and beam search, our model achieves a 100\% full simplification rate on a representative selection of 5-point gluon tree-level amplitudes in Yang-Mills theory, including expressions with over 200 initial terms.

hep-th↗

arXiv2026

A Quenched and Relatively Isolated Dwarf Galaxy in the Local Volume

An increasing number of discoveries of isolated and quenched dwarf galaxies are challenging the idea that the present-day local environment of low-mass systems is the main determinant of their quenching. We present new Hubble Space Telescope (HST) data of one such system, the dwarf galaxy Canes Venatici C (CVn C). CVn C is a low-mass (3.4(+4.2-2.6)*10^6 M_sun) galaxy with a Tip of the Red Giant Branch distance of 8.43(+0.47-0.32) Mpc determined from the resolved stars in the HST imaging, which we also use to derive CVn C's structural parameters. CVn C's distance places CVn C in the Local Volume and in an isolated environment with the most tidally influential L* galaxy > 5Rvir away. Additional constraints from the HST color-magnitude diagram, archival Far-Ultraviolet (FUV), and neutral hydrogen (HI) data show that CVn C is quenched, with no evidence of star formation in the last 100 Myr and no detectable gas (MHI < 1.5*10^6 M_sun). Circumstantial evidence suggests that CVn C may have quenched via past interactions with the L* galaxy NGC 4631 (L_K = 10^10.4 L_sun), and was possibly sent on an extreme backsplash orbit by the tidal dissolution of a subhalo group. However, other quenching mechanisms-such as stripping via the cosmic web-cannot be ruled out. CVn C adds to the growing number of quenched dwarf galaxies in under-dense environments, a population that will be critical to defining the mass and environment regimes in which different quenching mechanisms operate.

astro-ph.GA↗

arXiv2025

Look everywhere effects in anomaly detection

Machine learning-based anomaly detection methods are able to search high-dimensional spaces for hints of new physics with much less theory bias than traditional searches. However, by searching in many directions all at once, the statistical power of these search strategies is diluted by a variant of the look elsewhere effect. We examine this challenge in detail, focusing on weakly supervised methods. We find that training and testing on the same data results in badly miscalibrated $p$-values due to the anomaly detector searching everywhere in the data and overfitting on statistical fluctuations. However, if these $p$-values can be calibrated, they may offer the best sensitivity to anomalies, since this approach uses all of the data. Conversely, training on half of the data and testing on the other half results in perfectly calibrated $p$-values, but at the cost of reduced sensitivity to anomalies. Similarly, regularization methods such as early stopping can help with $p$-value calibration but also possibly at the expense of sensitivity. Finally, we find that k-folding strikes an effective balance between calibration and sensitivity. Our findings are supported by numerical studies with Gaussian random variables as well as from collider physics using the LHC Olympics benchmark anomaly detection dataset.

hep-ph↗

arXiv2025

ClearPotential: Revealing Local Dark Matter in Three Dimensions

We present ClearPotential, a data-driven, three-dimensional measurement of the gravitational potential of the local Milky Way using unsupervised machine learning, without the symmetry assumptions, specific functional forms, and binning required in previous work. The potential is modeled as a neural network, optimized to solve the equilibrium collisionless Boltzmann equation for the observed phase space density of Gaia DR3 Red Clump stars within 4 kpc of the Sun. This density is obtained from data using normalizing flows, and our unsupervised solution to the Boltzmann equation automatically corrects for selection effects from crowding and the dust-driven extinction of starlight. Our fully-differentiable model of the gravitational potential allows us to map the acceleration and mass density of the Galaxy in the volume around the Sun, including in the dust-obscured disk towards the Galactic Center. We determine the dark matter density at the Solar radius to be $(0.84 \pm 0.08)\times 10^{-2}\,{M}_\odot/{\rm pc}^3$, and analyze the structure of the dark matter halo. We find strong evidence for a tilted oblate halo, weak preference for a cored inner profile, and the strongest constraints to date on a possible dark matter disk. We place a bound on the timescale of disequilibrium in the local Milky Way, and find mild evidence for disequilibrium using independent acceleration measurements from timings of binary pulsar systems. This work provides the clearest map of the local Galactic potential to date and marks an important step in the era of data-driven astrometry.

astro-ph.GA↗

arXiv2025

Enhancing next token prediction based pre-training for jet foundation models

Next token prediction is an attractive pre-training task for jet foundation models, in that it is simulation free and enables excellent generative capabilities that can transfer across datasets. Here we study multiple improvements to next token prediction, building on the initial work of OmniJet-$\alpha$. Instead of tokenizing particles and subsequently only using the token-ID as the model input for both the generative and the classification task, we adopt a hybrid setup, which allows us to use continuous feature vectors as model input while only using token-IDs in the next token prediction target. Secondly, we explore a combined pre-training strategy that combines masked particle modeling and generative learning objectives. Taken together, these changes greatly improve the performance in downstream classification tasks without any loss in generative performance.

hep-ph↗

arXiv2025

SURFing to the Fundamental Limit of Jet Tagging

Beyond the practical goal of improving search and measurement sensitivity through better jet tagging algorithms, there is a deeper question: what are their upper performance limits? Generative surrogate models with learned likelihood functions offer a new approach to this problem, provided the surrogate correctly captures the underlying data distribution. In this work, we introduce the SUrrogate ReFerence (SURF) method, a new approach to validating generative models. This framework enables exact Neyman-Pearson tests by training the target model on samples from another tractable surrogate, which is itself trained on real data. We argue that the EPiC-FM generative model is a valid surrogate reference for JetClass jets and apply SURF to show that modern jet taggers may already be operating close to the true statistical limit. By contrast, we find that autoregressive GPT models unphysically exaggerate top vs. QCD separation power encoded in the surrogate reference, implying that they are giving a misleading picture of the fundamental limit.

hep-ph↗

arXiv2025

How to pick the best anomaly detector?

Anomaly detection has the potential to discover new physics in unexplored regions of the data. However, choosing the best anomaly detector for a given data set in a model-agnostic way is an important challenge which has hitherto largely been neglected. In this paper, we introduce the data-driven ARGOS metric, which has a sound theoretical foundation and is empirically shown to robustly select the most sensitive anomaly detection model given the data. Focusing on weakly-supervised, classifier-based anomaly detection methods, we show that the ARGOS metric outperforms other model selection metrics previously used in the literature, in particular the binary cross-entropy loss. We explore several realistic applications, including hyperparameter tuning as well as architecture and feature selection, and in all cases we demonstrate that ARGOS is robust to the noisy conditions of anomaly detection.

hep-ph↗

arXiv2025

Via Machinae 3.0: A search for stellar streams in Gaia with the CATHODE algorithm

We apply the model-agnostic anomaly detection method Cathode - originally developed for particle physics - to search for stellar streams in Gaia data. We combine Cathode with Via Machinae 3.0: a re-optimized version of the stellar stream detection method that was previously applied to Gaia data together with the related anomaly detection technique Anode. We demonstrate that the combination of Via Machinae 3.0 with Cathode, called VM3-C, not only re-discovers previously known streams, but also confirms many candidate streams identified in combination with Anode (denoted VM3-A). Compared to VM3-A, the number of stream candidates detected by VM3-C increases by around 10%. Moreover, both of the methods discover the same two large clusters of stream candidates in the Northern Galactic hemisphere. We dub these highly significant anomalous structures the Raritan stream and the Passaic stream. These two structures may indicate the presence of larger objects, such as dwarf galaxy streams, or non-trivial orbital dynamics resulting in bifurcation or fanning, and are promising and high-priority targets for further analysis.

astro-ph.GA↗

arXiv2025

The Future of Artificial Intelligence and the Mathematical and Physical Sciences (AI+MPS)

This community paper developed out of the NSF Workshop on the Future of Artificial Intelligence (AI) and the Mathematical and Physics Sciences (MPS), which was held in March 2025 with the goal of understanding how the MPS domains (Astronomy, Chemistry, Materials Research, Mathematical Sciences, and Physics) can best capitalize on, and contribute to, the future of AI. We present here a summary and snapshot of the MPS community's perspective, as of Spring/Summer 2025, in a rapidly developing field. The link between AI and MPS is becoming increasingly inextricable; now is a crucial moment to strengthen the link between AI and Science by pursuing a strategy that proactively and thoughtfully leverages the potential of AI for scientific discovery and optimizes opportunities to impact the development of AI by applying concepts from fundamental science. To achieve this, we propose activities and strategic priorities that: (1) enable AI+MPS research in both directions; (2) build up an interdisciplinary community of AI+MPS researchers; and (3) foster education and workforce development in AI for MPS researchers and students. We conclude with a summary of suggested priorities for funding agencies, educational institutions, and individual researchers to help position the MPS community to be a leader in, and take full advantage of, the transformative potential of AI+MPS.

cs.AI↗

arXiv2025

Quirk SUEP

We propose searching for physics beyond the Standard Model in the low-transverse-momentum tracks accompanying hard-scatter events at the LHC. TeV-scale resonances connected to a dark QCD sector could be enhanced by selecting events with anomalies in the track distributions. As a benchmark, a quirk model with microscopic string lengths is developed, including a setup for event simulation. For this model, strategies are presented to enhance the sensitivity compared to inclusive resonance searches: a simple cut-based selection, a supervised search, and a model-agnostic weakly supervised anomaly search with the CATHODE method. Expected discovery potentials and exclusion limits are shown for 140 fb$^{-1}$ of 13 TeV proton-proton collisions at the LHC.

hep-ph↗

arXiv2025

Generator Based Inference (GBI)

Statistical inference in physics is often based on samples from a generator (sometimes referred to as a ``forward model") that emulate experimental data and depend on parameters of the underlying theory. Modern machine learning has supercharged this workflow to enable high-dimensional and unbinned analyses to utilize much more information than ever before. We propose a general framework for describing the integration of machine learning with generators called Generator Based Inference (GBI). A well-studied special case of this setup is Simulation Based Inference (SBI) where the generator is a physics-based simulator. In this work, we examine other methods within the GBI toolkit that use data-driven methods to build the generator. In particular, we focus on resonant anomaly detection, where the generator describing the background is learned from sidebands. We show how to perform machine learning-based parameter estimation in this context with data-derived generators. This transforms the statistical outputs of anomaly detection to be directly interpretable and the performance on the LHCO community benchmark dataset establishes a new state-of-the-art for anomaly detection sensitivity.

hep-ph↗