GRAFT-ATHENA: Self-Improving Agentic Teams for Autonomous Discovery and Evolutionary Numerical Algorithms
cs.LG, cs.MA, math.PR
Submitted: 2026-05-11
Updated: 2026-09-15
License: http://creativecommons.org/licenses/by-nc-sa/4.0/
The gist: Scientific methods are developed for classes of problems, so knowledge transfers across structurally related cases.
Terminology
Abstract
Scientific methods are developed for classes of problems, so knowledge transfers across structurally related cases. Language-model agents can execute scientific workflows, but their problem--method relationships remain implicit, so each new problem restarts the search and little of what worked transfers. We introduce GRAFT--ATHENA, which makes this problem-to-method map explicit as an expandable probabilistic structure of admissible problems, methods, and their dependencies. Graph factorization keeps the substrate tractable, and semantic fingerprints measure similarity, so experience guides related problems. As a result, the framework matched or exceeded expert baselines, attaining near-machine-precision losses in physics-informed learning, reproducing clinically consistent blood-rheology trends, and developing a high-order hypersonic-flow solver for the Apollo Command Module that matched experimental measurements within 1.8%. It also proposed a certified regularization for ill-posed in vivo brain-flow reconstruction, developed a spectrally convergent physics-informed architecture, and established machine-checked universal-approximation theorems for two widely used architectures. Scientific structure enables cumulative and verifiable agentic discovery.
Sources
- Accelerating scientific discovery with Co-Scientist
- The Denario project: Deep knowledge AI agents for scientific discovery
- GraphAgents: Knowledge Graph-Guided Agentic AI for Cross-Domain Materials Design
- Higher-Order Knowledge Representations for Agentic Scientific Reasoning
- ATHENA: Agentic Team for Hierarchical Evolutionary Numerical Algorithms
- Semi-Autonomous Mathematics Discovery with Gemini: A Case Study on the Erd\H{o}s Problems
- ATLAS: A Multi-LLM Training Framework for EvoDPO with Adaptive Reference Evolution
- Autonomous Agents Coordinating Distributed Discovery Through Emergent Artifact Exchange
- Multiagent Finetuning: Self Improvement with Diverse Reasoning Chains
- GPT vs Human for Scientific Reviews: A Dual Source Review on Applications of ChatGPT in Science
- Mathematical exploration and discovery at scale
- TRINITY: An Evolved LLM Coordinator
- Recursive Multi-Agent Systems
- Adaptive numerical simulations with Trixi.jl: A case study of Julia for scientific computing
- Curvature-Aware Optimization for High-Accuracy Physics-Informed Neural Networks
- Gradient Alignment in Physics-informed Neural Networks: A Second-Order Optimization Perspective
- PINNsAgent: Automated PDE Surrogation with Large Language Models
- Lang-PINN: From Language to Physics-Informed Neural Networks via a Multi-Agent Framework
- Unveiling the optimization process of Physics Informed Neural Networks: How accurate and competitive can PINNs be?
- A deep reinforcement learning model based on deterministic policy gradient for collective neural crest cell migration
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks