返回首页
arXiv AI··论文与技术

The Trinity:Capability-Driven Recursive Self-Improvement for Multimodal Agents

中文摘要

本研究提出“The Trinity”框架,利用跨任务共享能力实现多模态智能体的递归自我改进,解决了环境构建与人工迭代的瓶颈。

English Summary

The paper introduces The Trinity, a framework for multimodal agents to achieve recursive self-improvement by leveraging shared capabilities across tasks, overcoming environment and iteration bottlenecks.

原文节选

arXiv:2610.36077v1 Announce Type: new Abstract: Multimodal intelligence can be viewed as human-guided co-evolution between worlds and agents through interaction. However, scaling this process faces two bottlenecks: constructing heterogeneous, verifiable environments is costly, and manual Harness iteration cannot keep pace with growing task diversity and complexity. A key insight is that tasks with different modalities and surface forms can share the same underlying capability requirements. We therefore introduce The Trinity, a recursive self-improvement framework coupling world generation with Harness evolution through three roles: Creator, Explorer, and Evolver. Drawing on verified execution trajectories, Evolver shifts the basis of task organization from surface categories to shared requirements, reorganizing the original task tree into a hierarchical capability tree. We introduce Capability-Driven Real2Sim2Real: for heterogeneous tasks at the same node, Creator abstracts their common requirements into simplified, verifiable simulated worlds for capability learning, while evaluation on original tasks measures transfer. Along this capability tree, Explorer progressively evolves re…