paper-with-me

Papers

Language Agents as Optimizable Graphs

2024-02-26 · Mingchen Zhuge, Wenyi Wang, Louis Kirsch, Francesco Faccio, Dmitrii Khizbullin, Jürgen Schmidhuber

Various human-designed prompt engineering techniques have been proposed to improve problem solvers based on Large Language Models (LLMs), yielding many disparate code bases. We unify these approaches by describing LLM-based agents as computational graphs. The nodes implement functions to process multimodal data or query LLMs, and the edges describe the information flow between operations. Graphs can be recursively combined into larger composite graphs representing hierarchies of inter-agent collaboration (where edges connect operations of different agents). Our novel automatic graph optimizers (1) refine node-level LLM prompts (node optimization) and (2) improve agent orchestration by changing graph connectivity (edge optimization). Experiments demonstrate that our framework can be used to efficiently develop, integrate, and automatically improve various LLM agents. The code can be found at https://github.com/metauto-ai/gptswarm.

📄 PDF Abstract BibTeX arXiv:2402.16823

Code (2)

metauto-ai/gptswarm 공식 구현
lukasvierling/dynamicgptswarm

Tasks

Prompt Engineering

Similar Papers 제목 키워드 기반

Advanced Situational Graphs for Robot Navigation in Structured Indoor Environments

2022-11-16 · Hriday Bavle, Jose Luis Sanchez-Lopez, Muhammad Shaheer, Javier Civera 외

Mobile robots extract information from its environment to understand their current situation to enable intelligent decision making and autonomous task execution. In our previous work, we introduced the concept of Situati…

Decision MakingRobot Navigation

Agint: Agentic Graph Compilation for Software Engineering Agents

2025-11-24 · Abhi Chivukula, Jay Somasundaram, Vijay Somasundaram arxiv

LLM-based coding agents are increasingly common but still face challenges in context management, latency, reliability, reproducibility, and scalability. We present Agint, an agentic graph compiler, interpreter, and runti…

vS-Graphs: Integrating Visual SLAM and Situational Graphs through Multi-level Scene Understanding

2025-03-03 · Ali Tourani, Saad Ejaz, Hriday Bavle, David Morilla-Cabello 외

Current Visual Simultaneous Localization and Mapping (VSLAM) systems often struggle to create maps that are both semantically rich and easily interpretable. While incorporating semantic scene knowledge aids in building r…

Scene UnderstandingSimultaneous Localization and Mapping

Generalizing Test-time Compute-optimal Scaling as an Optimizable Graph

2025-10-29 · Fali Wang, Jihai Chen, Shuhua Yang, Runxue Bao 외 arxiv

Test-Time Scaling (TTS) improves large language models (LLMs) by allocating additional computation during inference, typically through parallel, sequential, or hybrid scaling. However, prior studies often assume fixed co…

VLA Grounder: Language-Conditioning Space Optimization for Black-Box VLA Models

2026-07-05 · Damir Shodiev, Aleksei Staroverov, Nikita Kachaev, Alexey K. Kovalev 외 arxiv

Vision-Language-Action (VLA) models are commonly treated as end-to-end action policies conditioned on natural-language task descriptions. In practice, however, their behavior often depends sharply on how the instruction …

Reinforcement Learning