Researchllm agentsllmhigh energy physicssnakemake
LLM Agents Automate Higgs Diphoton Analysis
9.0
Researchers present a proof-of-principle study (initial submission Aug 30, 2025) demonstrating LLM agents automating a Higgs boson diphoton cross-section measurement using ATLAS Open Data and Snakemake. They benchmark Gemini, GPT-5, Claude, and leading open-weight models, define quantitative metrics (success rates, error distributions, costs), and release baseline code, with acceptance as a poster at ML4PS NeurIPS 2025.
Key Points
- 1Implements LLM-agent-driven automation for ATLAS Higgs diphoton cross-section analysis using Snakemake
- 2Benchmarks Gemini, GPT-5, Claude, and open-weight models to evaluate performance and variability
- 3Provides quantitative metrics and code release enabling reproducible evaluation and practitioner adoption
Scoring Rationale
High novelty with reproducible code and benchmarks; limited by focus on HEP use case and model sampling nondeterminism.
Sources
Public references used for this report.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
High-Value Overnight OrdersEasyDelivered International ShipmentsMediumOn-Time Delivery Rate by CarrierHard
250 free problems · No credit card
See all Logistics & Shipping problems
