# HALO: A Physics-Aware LLM Agent Framework for Nanophotonic Design

arXiv: https://arxiv.org/abs/2608.28877 · HTML: https://mlsanity.com/paper/2608.28877/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Medical and scientific domain agents, one of 16 directions in LLM agents and tool use.
- Builds on 16 of its 42 references that are in the map.
- First posted on arXiv 28 Aug 2026.

## Placed in

- LLM agents and tool use / [Medical and scientific domain agents](https://mlsanity.com/agents/medical-and-scientific-domain-agents/)

## Builds on (most cited first, 16 of 16)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Toolformer: Language Models Can Teach Themselves to Use Tools](https://mlsanity.com/paper/2302.04761/) (2023)
- [The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery](https://mlsanity.com/paper/2408.06292/) (2024)
- [PaperBench: Evaluating AI's Ability to Replicate AI Research](https://mlsanity.com/paper/2504.01848/) (2025)
- [ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery](https://mlsanity.com/paper/2410.05080/) (2024)
- [MatSci-NLP: Evaluating Scientific Language Models on Materials Science Language Tasks Using Text-to-Schema Modeling](https://mlsanity.com/paper/2305.08264/) (2023)
- [Global optimization of dielectric metasurfaces using a physics-driven neural network](https://mlsanity.com/paper/1906.04157/) (2019)
- [A multi-agentic framework for real-time, autonomous freeform metasurface design](https://mlsanity.com/paper/2503.20479/) (2025)
- [GeoGramBench: Benchmarking the Geometric Program Reasoning in Modern LLMs](https://mlsanity.com/paper/2505.17653/) (2025)
- [FloorplanQA: A Benchmark for Spatial Reasoning in LLMs using Structured Representations](https://mlsanity.com/paper/2507.07644/) (2025)
- [A Generative Machine Learning-Based Approach for Inverse Design of Multilayer Metasurfaces](https://mlsanity.com/paper/2008.02074/) (2020)
- [Toward Greater Autonomy in Materials Discovery Agents: Unifying Planning, Physics, and Scientists](https://mlsanity.com/paper/2506.05616/) (2025)
- [An Agentic Framework for Autonomous Metamaterial Modeling and Inverse Design](https://mlsanity.com/paper/2506.06935/) (2025)
- [Chat to Chip: Large Language Model Based Design of Arbitrarily Shaped Metasurfaces](https://mlsanity.com/paper/2509.24196/) (2025)
- [OPTIAGENT: A Physics-Driven Agentic Framework for Automated Optical Design](https://mlsanity.com/paper/2602.23761/) (2026)
- [Re4: Scientific Computing Agent with Rewriting, Resolution, Review and Revision](https://mlsanity.com/paper/2508.20729/) (2025)

## Abstract (arXiv)

Language models have recently been applied to nanophotonic design, but it remains unclear whether they can reliably translate optical objectives into simulation-ready designs, execute electromagnetic analysis, and revise decisions from numerical feedback. We introduce HALO, a physics-aware framework that couples language-model planners with typed design specifications, electromagnetic simulation, diagnostic evaluation, and optional reuse of prior failure trajectories in an iterative design loop. We further introduce HALO-Bench, a 52-task benchmark spanning lab-derived, paper-derived, and open-ended nanophotonic design tasks under a shared evaluation protocol. We compare three planner configurations: a Fixed Structured Workflow, an Autonomous Structured Agent using the same simulation interface, and an Autonomous Coding Agent that directly writes and executes simulation code. The Fixed Structured Workflow is the most token-efficient and exhibits no observed code- or path-level failures, while autonomous coding can achieve higher task success with stronger models at the cost of additional operational failures. We also study reuse of prior failed trajectories. On targeted multi-round tasks, retrieved failure feedback reduces both iterations to first success and total token use. These results clarify the tradeoffs between explicit interfaces, autonomous execution, and reusable design experience in scientific agents.