arXiv subjects

Benjamin Nachman

Publications and source records attributed to Benjamin Nachman.

At least 19 recordsLinked to original sources

arXiv2026

Differentiable Parametric Simulation and Reconstruction Models in Parnassus

Parnassus is a framework for fast detector simulation and reconstruction, directly mapping truth-level particles onto reconstructed objects. Such models can be built from deep generative networks trained on paired samples, which are fit automatically to a target detector, or from parametric prescriptions of the kind used by Delphes, which are constructed by hand. We remove this asymmetry by making the parametric models fully differentiable so that their parameters can be fit to a target sample by gradient descent. We demonstrate closure by fitting a parametric model to samples from a known configuration of itself, recovering the generating parameters and characterizing the degeneracies among them, and we present a first fit to CMS full simulation. The resulting models are interpretable, inexpensive, and run in the standard Parnassus pipeline which is fully Python based and GPU enabled.

hep-ex↗

arXiv2026

The Living Guide of Machine Learning for Particle Physics

We started the Living Review of Machine Learning for Particle Physics (HEP-ML Living Review) in 2020 as a community-maintained, near-comprehensive bibliography of machine learning in particle physics. The field was then growing faster than any single researcher could follow, finding the relevant papers was hard, and a structured, continuously updated reference paid off immediately. Since then the literature has grown by more than an order of magnitude, the methods reach far beyond the classification and generation tasks of the early years, and the community has built its own ecosystem of topic-specific reviews, benchmark papers, and software frameworks. The original model no longer serves this field well, and we can no longer sustain it. We therefore change direction. We freeze the Living Review as an archival reference covering the literature up to 1 June 2026, where it remains a stable record of the first phase of HEP-ML. A new resource, the HEP-ML Living Guide, replaces it. It does not list everything. It curates, it annotates, and it points readers to foundational and representative work, so that researchers can find their way into a mature and rapidly diversifying field. In this article we explain why we make this change and how the new resource works.

hep-ph↗

arXiv2026

Predict before you train: Scaling Laws for particle physics foundation models

The largest machine learning models in particle physics are also the most expensive to train, yet the return on scaling a given architecture cannot be estimated before that compute is spent. Scaling laws have been fit for jets, but none has yet been shown to predict the performance of models it was not fit on. We show that, for a generic transformer pretrained on collider jets, it can be forecast. Fitting a joint model-and-data scaling law on small models alone, spanning three orders of magnitude of training compute, we predict the loss of models trained afterward with more than one hundred times more compute to within one percent. We then connect the forecast to downstream physics performance: across two standard tagging benchmarks, lower pretraining loss yields systematically lower fine-tuning loss and higher background rejection after fine-tuning. Within this model family and these tasks, a compute budget can therefore be translated into expected physics performance before any large model is trained. The final frontier model is consistent with the published numbers for current state-of-the-art physics-aware foundation models trained on the same corpus, on accuracy, AUC, and quark/gluon rejection, with a residual edge for the physics-aware model only in the high-purity tail of top tagging. We release five pretrained models spanning multiple sizes, together with the complete training recipe and code.

hep-ex↗

arXiv2026

The Well-Tempered Likelihood: Honest Confidence Intervals for Misspecified Models

Likelihood-based inference in particle physics, and in the physical sciences more broadly, relies on the assumption that the model accurately describes the data. When the model is misspecified, though, standard confidence intervals shrink to zero width with increasing data, producing overconfident and potentially misleading constraints. We propose the well-tempered likelihood, which divides the likelihood-ratio test statistic by a goodness-of-fit (GOF) statistic evaluated at the best-fit point. Under correct specification, the GOF is $\mathcal{O}(1)$ and standard inference is recovered. Under misspecification, both the likelihood and the GOF scale as $\mathcal{O}(N)$ for $N$ data points, so their ratio remains $\mathcal{O}(1)$ and the resulting well-tempered confidence interval self-limits at a floor determined by the model's inadequacy, essentially reducing the effective sample size. In other words, \textit{all models are correct, as long as your dataset is small enough}. We present binned and unbinned formulations of the well-tempered likelihood---the latter based on a classifier two-sample test---and demonstrate our method on a Gaussian example and on a measurement of the strong coupling constant using synthetic electron-positron collisions.

stat.ME↗

arXiv2026

Towards anomaly detection searches for new physics signatures including Higgs bosons with weakly supervised machine learning

The Higgs boson, with its universal coupling to mass, provides a broadly applicable portal to sectors beyond the Standard Model and is therefore a natural anchor for anomaly detection (AD) at collider experiments. The Higgs And X Anomaly Detection (HAXAD) strategy offers a principled approach to searching for such anomalies occurring in association with a Higgs boson by combining machine-learning-based feature embedding, background estimation, and weakly supervised classification. This work extends the previous HAXAD approach towards the level of maturity required for application to recorded collider data. A major addition is the introduction and comparison of two new embedding strategies, which in turn shape the background estimation and classification. In addition, a new inference framework is developed, yielding signal-agnostic and signal-specific cross section limits and thereby completing the statistical machinery needed for future AD analyses built on HAXAD. The set of investigated signal models is also significantly expanded, allowing for the evaluation of sensitivity on a much broader phase space. Improvements to the method increase signal sensitivity with respect to the original method, and when benchmarked against an example cut-based search on the same final state, HAXAD matches or exceeds the best individual cut-based limits for a wide variety of considered signal models. These developments strengthen the case for HAXAD as a viable and compelling AD-based search strategy with novel discovery potential at colliders.

hep-ex↗

arXiv2026

Parnassus: A GPU-enabled, Python-based Package for Fast Particle Detector Simulation and Reconstruction

We present the public software release of Parnassus, a Python/PyTorch, GPU-compatible framework for fast detector simulation and reconstruction in particle and nuclear physics. Parnassus provides a user-friendly framework with interchangeable detector models: neural models can emulate computationally expensive Geant4-based detector simulation and reconstruction chains, while parametric models provide PyTorch implementations of selected Delphes-style detector responses. This initial release includes two models of the CMS detector: one based on a flow-matching neural network architecture and one based on a PyTorch implementation of the Delphes CMS card (parametric bias and smearing). PyTorch versions of the ATLAS and ALEPH Delphes cards are also available, together with a flow-matching neural model of the ALEPH detector that extends the framework to the e+e- LEP environment. All detector-specific backends share the same process-agnostic and detector-agnostic API: users select a detector card - analogous to choosing a detector card in Delphes - and the same tool can be applied to new physics processes without retraining the released detector model. There are native interfaces to the event generator Pythia and the event clustering package FastJet. Unlike previous C++/ROOT-based tools, Parnassus provides GPU-capable PyTorch detector-response backends and requires no ROOT installation. We describe the installation, command-line and Python API, configuration system, and demonstrate the framework on Standard Model and BSM processes.

hep-ex↗

arXiv2026

Towards Engineering Scaling Laws with Pretraining Data Composition

Neural scaling laws describe how model performance improves as a power law in compute, model size, and dataset size. While well-established for large language models, these relationships are emerging for large models in particle physics. As with language, empirical studies show that the performance scales as a power law. However, unlike natural language or image domains, fundamental physics has high-fidelity simulators that produce synthetic data cheaply. This favors scaling regimes where additional data is cheaper than additional parameters, and allows the pretraining dataset itself to be engineered to influence the scaling. For the task of classifying hadronic jets produced in collisions of high-energy particle beams, we show that the scaling behavior can be engineered towards requiring more data rather than larger models by inclusion of pretraining data which is more diverse and better aligned with the downstream classification task.

hep-ex↗

arXiv2026

Pre-Training for Simulation-Based Science: A Study on Jet Foundation Model Training Objectives

Foundation models (FMs) trained on large datasets and fine-tuned on downstream tasks have emerged as a powerful paradigm in AI for science. Industrial FMs are typically trained using self-supervision with masking due to the lack of labels. In many scientific domains, accurate simulations are plentiful and facilitate large, labeled datasets. This opens up new possibilities for pre-training. We present a systematic comparison of pre-training methods using the OmniLearned High Energy Physics FM framework. We test supervised classification, flow-matching generation, and self-supervised masked particle modeling. All models are pre-trained on the JetClass dataset and fine-tuned on two representative downstream tasks, top jet classification and JetNet conditional generation. Among other observations, for classification tasks, we find that pure classifier pre-training is optimal when downstream labels and model capacity are plentiful, but combining it with self-supervised masked particle modeling (MPM) is uniquely powerful in the low-finetuning label regime. Flow matching-based generative pre-training seems to provide little benefit for downstream classification, and interestingly, for downstream generation, we find that flow matching must be in the pre-training objective to see a significant finetuning advantage, hinting at the orthogonality of classification and generation tasks. That is, for a model to transfer to both generative and classification downstream tasks, it must be pre-trained on both. This study provides a template for controlled scaling analysis of pre-training objectives for foundation models in simulation-based sciences.

hep-ph↗

arXiv2026

Reweighting Adversarial Networks for Unbinned Unfolding

Differential cross sections are the currency of scientific exchange in particle and nuclear physics. Recently, machine learning methods have enabled unbinned and high-dimensional cross section measurements through new approaches to unfolding. A key challenge with unfolding is that it is a bi-level optimization problem where constraints are available at the detector level while the target is at the particle level, linked by a stochastic detector response. Further complications arise when the particle-level and detector-level distributions have non-overlapping or only partially overlapping support, which can destabilize training and degrade unfolding performance. In this paper, we introduce a new unbinned unfolding technique called the Reweighting Adversarial Network (RAN), which can be viewed as a generalization of the Moment Unfolding protocol to accommodate full phase-space unfolding. RANs address the bi-level optimization problem through a particle-level reweighting function steered by a Wasserstein critic at the detector level. RANs do not require overlapping support at the detector level, nor multiple iterations of training. We evaluate the performance of RANs with Gaussian data and jet substructure studies, including cases specifically designed to stress test the method under vanishing support overlap. We demonstrate that RANs outperform state-of-the-art methods in accuracy and have a lower computational overhead.

hep-ph↗

arXiv2026

An AI-ready, Polarized Electron-Positron Collision Dataset

We present a modernized, AI-ready release of reconstructed data from the SLD experiment at the SLAC Linear Collider (SLC). The dataset comprises approximately 660{,}000 reconstructed events collected at $\sqrt{s}\approx 91.2$~GeV with a highly polarized electron beam from 1996--1998. The data have been translated from legacy formats into modern, widely-used file formats with the help of AI agents. The release also includes a corpus of newly digitized SLD internal documentation. We describe the contents of both components and provide physics validation demonstrations along with illustrations of their utility for physics and machine learning research in particle physics.

hep-ex↗

arXiv2026

A Scientific Human-Agent Reproduction Pipeline

Reproducing scientific analyses is essential for preserving knowledge, building extensible codebases, and deepening researcher understanding - yet the effort often outweighs its academic recognition. We argue that the reproduction of scientific data analyses is fundamentally a translation task: converting human-readable knowledge (papers, documentation) into machine-readable analysis code. This makes it uniquely well-suited for AI agents. We present SHARP (Scientific Human-Agent Reproduction Pipeline), a structured framework for reproducing scientific analyses through human-agent collaboration. SHARP decomposes a reproduction task into discrete steps, which an AI agent executes autonomously using specialized subagents for code generation, testing, and quality assurance. At defined checkpoints, the researcher reviews progress, provides feedback, and steers the analysis - keeping the human firmly in control of scientific judgment while the agent handles implementation. We demonstrate SHARP by reproducing a jet classification task in particle physics from a published paper. We evaluate the reproduction along three axes: analysis performance against the original results, code quality and faithfulness, and the nature of the human-agent conversation. The latter is evaluated with a novel framework for characterizing human-agent interactions. Our work highlights a practical model for AI-assisted scientific reproduction where the researcher's role shifts from writing code to understanding, evaluating, and directing - elevating human understanding rather than replacing it.

hep-ph↗

arXiv2026

FAIR Universe Weak Lensing ML Uncertainty Challenge: Handling Uncertainties and Distribution Shifts for Precision Cosmology

Weak gravitational lensing, the correlated distortion of background galaxy shapes by foreground structures, is a powerful probe of the matter distribution in our universe and allows accurate constraints on the cosmological model. In recent years, high-order statistics and machine learning (ML) techniques have been applied to weak lensing data to extract the nonlinear information beyond traditional two-point analysis. However, these methods typically rely on cosmological simulations, which poses several challenges: simulations are computationally expensive, limiting most realistic setups to a low training data regime; inaccurate modeling of systematics in the simulations create distribution shifts that can bias cosmological parameter constraints; and varying simulation setups across studies make method comparison difficult. To address these difficulties, we present the first weak lensing benchmark dataset with several realistic systematics and launch the FAIR Universe Weak Lensing Machine Learning Uncertainty Challenge. The challenge focuses on measuring the fundamental properties of the universe from weak lensing data with limited training set and potential distribution shifts, while providing a standardized benchmark for rigorous comparison across methods. Organized in two phases, the challenge will bring together the physics and ML communities to advance the methodologies for handling systematic uncertainties, data efficiency, and distribution shifts in weak lensing analysis with ML, ultimately facilitating the deployment of ML approaches into upcoming weak lensing survey analysis.

astro-ph.CO↗

arXiv2026

Cross-Domain Transfer with Particle Physics Foundation Models: From Jets to Neutrino Interactions

Future AI-based studies in particle physics will likely start from a foundation model to accelerate training and enhance sensitivity. As a step toward a general-purpose foundation model for particle physics, we investigate whether the OmniLearned and ParticleViT foundation models pretrained on diverse high-$Q^2$ simulated and real $pp$ and $ep$ collisions retain useful knowledge to a few-GeV fixed-target neutrino experiment. We process MINERvA neutrino--nucleus scattering events and evaluate pretrained models on two types of tasks: regression of available energy and binary classification of charged-current pion final states ($\mathrm{CC1\pi^{\pm}}$, $\mathrm{CCN\pi^{\pm}}$, and $\mathrm{CC1\pi^{0}}$). Pretrained OmniLearned and ParticleViT models outperform similarly sized models trained from scratch at the same compute budget, with the largest gains for OmniLearned on regression and for ParticleViT on classification. When the same transformer architecture is instead initialized from unrelated text pretraining (BERT), this advantage appears only marginally for classification in terms of compute efficiency and not in any way for regression. These results suggest that particle-level foundation models acquire inductive biases that generalize across large differences in energy scale, detector technology, and underlying physics processes, pointing toward detector-agnostic inference in particle physics.

hep-ex↗

arXiv2026

An AI-based Detector Simulation and Reconstruction Model for the ALEPH Experiment at LEP

We present the application of Parnassus, a generative model for full detector simulation and reconstruction, to the ALEPH detector at the Large Electron-Positron Collider (LEP). Training on simulated $e^+e^-$ to Z to qqbar events processed through the ALEPH detector simulation and reconstruction, we demonstrate that Parnassus faithfully reproduces the detector response at the event, jet, and particle levels, with substantially better agreement than the Delphes fast simulation. The clean $e^+e^-$ environment, free of pileup and characterized by simple event topologies, provides a well-controlled benchmark for evaluating the generative model's fidelity. Our results demonstrate that modern neural-network-based generative simulation approaches, developed primarily for LHC experiments, generalize naturally to historical collider experiments with distinct detector geometries and physics environments. This work shows that Parnassus can be applied beyond the LHC context and serves as an important tool for legacy data analysis where archival software tools are challenging to resurrect.

physics.ins-det↗

arXiv2026

Generalizable Foundation Models for Calorimetry via Mixtures-of-Experts and Parameter Efficient Fine Tuning

Modern particle physics experiments face an increasing demand for high-fidelity detector simulation as luminosities rise and computational requirements approach the limits of available resources. Deep generative models have emerged as promising surrogates for traditional Monte Carlo simulation, with recent advances drawing inspiration from large language models (LLM) and next-token prediction paradigms. In this work, we introduce a generalizable foundation model for calorimetry built on next-token transformer backbones, designed to support modular adaptation across materials, particle species, and detector configurations. Our approach combines Mixture-of-Experts pre-training with parameter-efficient fine-tuning strategies to enable controlled, additive model expansion without catastrophic forgetting. A pre-trained backbone is trained to generate electromagnetic showers across multiple absorber materials, while new materials are incorporated through the addition and tuning of lightweight expert modules. Extensions to new particle types are achieved via parameter-efficient fine-tuning and modular vocabularies, preserving the integrity of the base model. This design enables efficient, incremental knowledge integration as new simulation datasets become available, a critical requirement in realistic detector-development workflows. In addition, we demonstrate that next-token calorimeter models are computationally competitive with standard generative approaches under established LLM optimization procedures. These results establish next-token architectures as a viable path toward extensible, physics-aware foundation models for calorimetry and future high-energy physics experiments.

physics.ins-det↗

arXiv2026

Signal-Aware Contrastive Latent Spaces for Anomaly Detection

High-dimensional feature spaces in particle physics events pose a fundamental challenge to density-estimation-based weakly supervised anomaly detection, whose fidelity degrades rapidly with an increasing number of dimensions. We propose a signal-aware latent space construction using supervised contrastive learning trained on simulated Standard Model backgrounds and a diverse set of hypothesized Beyond the Standard Model (BSM) signals. The resulting latent space is low-dimensional, regularized, and signal-sensitive, enabling high-fidelity density estimation for downstream weakly supervised anomaly detection. We demonstrate the approach in a diphoton final state, testing sensitivity across a broad range of BSM scenarios including supersymmetry models, extended Higgs sectors, heavy neutral resonances, and flavor-changing neutral currents. For signals represented in the contrastive training data, the method can elevate discovery sensitivity from previously inaccessible levels to the discovery regime. Critically, the approach retains sensitivity to BSM models not present during training: interpolation and extrapolation to unseen signal topologies yield substantial improvements in expected significance compared to a background-only baseline. By bridging supervised latent space embedding with weakly supervised anomaly detection, this strategy offers a viable path toward anomaly detection in high-dimensional feature spaces at the LHC and beyond.

hep-ph↗

arXiv2026

Searching for Anomalies with Foundation Models

Foundation models have the potential to extend the discovery reach for anomaly detection searches. When studying the large OmniLearned foundation model on data from the CMS experiment, unexpected behavior was observed in a mass sideband. The purpose of this paper is to perform a full analysis, including a complete background estimate, on the phase space picked out by the large model. We find that the background estimation describes the data well in validation regions, but is unable to accurately model the signal region. We invite further scrutiny of these events and our methods.

hep-ex↗

arXiv2026

Unfolding with a Wasserstein Loss

Data unfolding -- the removal of noise or artifacts from measurements -- is a fundamental task across the experimental sciences. Of particular interest are applications in physics, where the dominant approach is Richardson-Lucy (RL) deconvolution. The classical RL approach aims to find denoised data that, once passed through the noise model, is as close as possible to the measured data in terms of Kullback-Leibler (KL) divergence. This requires that the support of the measured data overlaps with the output of the noise model, a hypothesis typically enforced by binning, which introduces numerical error. As a counterpoint, the present work studies an alternative formulation using a Wasserstein loss. We establish sharp conditions for existence and uniqueness of optimizers, answering open questions of Li, et al., regarding necessary conditions for uniqueness in the case of transport map noise models. We then develop a provably convergent generalized Sinkhorn algorithm to compute approximate optimizers. Our algorithm requires only empirical observations of the noise model and measured data and scales with the size of the data, rather than the ambient dimension. Numerical experiments on one- and two-dimensional problems inspired by jet mass unfolding in particle physics demonstrate that the optimal transport approach offers robust, accurate performance compared to classical RL deconvolution, particularly when binning artifacts are significant.

math.OC↗