AI-for-Science Low-code Platform with Bayesian Adversarial Multi-Agent Framework

2026-03-03

AI Summary

This work proposes a low-code platform for AI for Science to address key challenges in scientific code generation, including the unreliability of large language models, severe error propagation in multi-agent workflows, and the lack of well-defined evaluation metrics for scientific tasks. The platform employs a three-agent Bayesian adversarial framework that jointly optimizes task management, code generation, and evaluation by dynamically updating prompt distributions and generating adversarial test cases. This co-optimization of tests and code reduces dependence on model reliability. Integrating multidimensional evaluation—encompassing functional correctness, structural alignment, and static analysis—with a human-in-the-loop interface, the system effectively suppresses error propagation and generates more robust and reliable scientific code, outperforming existing methods on interdisciplinary Earth science tasks.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Multiagent Systems: Adversarial Agents
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Large Language Models (LLMs) demonstrate potentials for automating scientific code generation but face challenges in reliability, error propagation in multi-agent workflows, and evaluation in domains with ill-defined success metrics. We present a Bayesian adversarial multi-agent framework specifically designed for AI for Science (AI4S) tasks in the form of a Low-code Platform (LCP). Three LLM-based agents are coordinated under the Bayesian framework: a Task Manager that structures user inputs into actionable plans and adaptive test cases, a Code Generator that produces candidate solutions, and an Evaluator providing comprehensive feedback. The framework employs an adversarial loop where the Task Manager iteratively refines test cases to challenge the Code Generator, while prompt distributions are dynamically updated using Bayesian principles by integrating code quality metrics: functional correctness, structural alignment, and static analysis. This co-optimization of tests and code reduces dependence on LLM reliability and addresses evaluation uncertainty inherent to scientific tasks. LCP also streamlines human-AI collaboration by translating non-expert prompts into domain-specific requirements, bypassing the need for manual prompt engineering by practitioners without coding backgrounds. Benchmark evaluations demonstrate LCP's effectiveness in generating robust code while minimizing error propagation. The proposed platform is also tested on an Earth Science cross-disciplinary task and demonstrates strong reliability, outperforming competing models.

Problem

• AI for Science
• code generation reliability
• error propagation
• evaluation uncertainty
• multi-agent workflows

Innovation

• Bayesian adversarial framework
• multi-agent LLM
• low-code platform
• AI for Science
• code generation reliability