arXiv subjects

Daniel Schiller

Publications and source records attributed to Daniel Schiller.

5 recordsLinked to original sources

arXiv2026

Agentic Re-Casting using Agentic Re-Simulations

Analysis re-casting at the LHC is highly standardized and nevertheless requires resources, time, and physics input. Building on the new MadAgents.v3, we show how a global SFitter analysis can be updated by an agentic system with a physicist in the loop. The agentic interface allows us to make the advanced SFitter methodology available to a wider audience. All physical and technical aspects of this agentic re-casting study can be trivially generalized beyond SFitter.

hep-ph↗

arXiv2026

One Generator, Any Process: LLM-Conditioning for the LHC

Neural network training for LHC event generation should, ideally, benefit from common high-level patterns in different processes. We propose novel conditioning schemes for continuous parameters, process labels, and Feynman diagrams. We employ pre-trained LLMs as multi-modal foundation models to provide descriptive embeddings for an autoregressive transformer. With such high-level physics-inductive bias the generative networks converge faster, provide better result, and generalize to unseen processes.

hep-ph↗

arXiv2026

MadAgents

We uncover an effective and communicative set of agents working with MadGraph. Agentic installation, learning-by-doing training, and user support provide easy access to state-of-the-art simulations and accelerate LHC research. We show in detail how MadAgents interact with inexperienced and advanced users, support a range of simulation tasks, and analyze results. In a second step, we illustrate how MadAgents automatize event generation and run an autonomous simulation campaign, starting from a pdf file of a paper. The updated Claude Code implementation includes a self-improvement loop.

hep-ph↗

arXiv2025

Large Language Models -- the Future of Fundamental Physics?

For many fundamental physics applications, transformers, as the state of the art in learning complex correlations, benefit from pretraining on quasi-out-of-domain data. The obvious question is whether we can exploit Large Language Models, requiring proper out-of-domain transfer learning. We show how the Qwen2.5 LLM can be used to analyze and generate SKA data, specifically 3D maps of the cosmological large-scale structure for a large part of the observable Universe. We combine the LLM with connector networks and show, for cosmological parameter regression and lightcone generation, that this Lightcone LLM (L3M) with Qwen2.5 weights outperforms standard initialization and compares favorably with dedicated networks of matching size.

astro-ph.CO↗

arXiv2025

Localized Gravity, de Sitter, and the Horizon Criterion

Realizing de Sitter-like solutions in string theory remains challenging, prompting speculation about which specific feature might be responsible for their inconsistency in quantum gravity. In this work, we focus on the `Horizon Criterion', which identifies spacetimes as problematic if they exhibit cosmological horizons. In particular, we study the implications for spacetimes with dynamical boundaries. We argue that requiring inertial observers localized on an end-of-the-world (ETW) brane to be in causal contact with every other observer is too restrictive as there exist string-theoretic solutions without this property. Hence, if one does not want to abandon the idea of cosmological horizons being the fundamental issue with de Sitter, a refined condition is needed. The requirement that inertial, boundary-localized observers should be in causal contact with all other observers on the same ETW brane is such an appropriate refinement. We explore the consequences of this criterion for ETW branes whose energy density is governed by a scalar field, considering two cases: First, with a scalar field confined to the ETW brane, and second, with a bulk modulus subject to a brane-localized potential.

hep-th↗