NovelSeek: When Agent Becomes the Scientist -- Building Closed-Loop System from Hypothesis to Verification

2025-05-22

AI Summary

Addressing interdisciplinary, complex scientific challenges remains a bottleneck in autonomous scientific research (ASR). Method: We propose NovelSeek—the first closed-loop, multi-agent framework for ASR—integrating large language model–driven scientific reasoning, automated code generation and execution, collaborative multi-agent decision-making, and real-time human expert intervention to form an end-to-end research pipeline spanning hypothesis generation, experimental design, and empirical validation. Contribution/Results: NovelSeek introduces a scalable, interactive, and computationally efficient architecture unified across 12 distinct scientific tasks. Empirical evaluation demonstrates substantial acceleration in high-quality discovery: +7.8 percentage points in reaction yield prediction within 12 hours; +0.27 in enhancer activity prediction accuracy within 4 hours; and +2.2% in 2D semantic segmentation precision within 30 hours. These results underscore NovelSeek’s capacity to significantly expedite rigorous, reproducible scientific inquiry.

Technology Category

• Search and Optimization: Learning to Search
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Artificial Intelligence (AI) is accelerating the transformation of scientific research paradigms, not only enhancing research efficiency but also driving innovation. We introduce NovelSeek, a unified closed-loop multi-agent framework to conduct Autonomous Scientific Research (ASR) across various scientific research fields, enabling researchers to tackle complicated problems in these fields with unprecedented speed and precision. NovelSeek highlights three key advantages: 1) Scalability: NovelSeek has demonstrated its versatility across 12 scientific research tasks, capable of generating innovative ideas to enhance the performance of baseline code. 2) Interactivity: NovelSeek provides an interface for human expert feedback and multi-agent interaction in automated end-to-end processes, allowing for the seamless integration of domain expert knowledge. 3) Efficiency: NovelSeek has achieved promising performance gains in several scientific fields with significantly less time cost compared to human efforts. For instance, in reaction yield prediction, it increased from 27.6% to 35.4% in just 12 hours; in enhancer activity prediction, accuracy rose from 0.52 to 0.79 with only 4 hours of processing; and in 2D semantic segmentation, precision advanced from 78.8% to 81.0% in a mere 30 hours.

Problem

• Developing a closed-loop multi-agent system for autonomous scientific research
• Enhancing research efficiency and innovation across diverse scientific fields
• Integrating human expert feedback with AI-driven hypothesis verification

Innovation

• Closed-loop multi-agent framework for ASR
• Human expert feedback integration interface
• Efficiency gains with minimal time cost