Paper page - Automating the Design of Embodied Agent Architectures

Source: https://huggingface.co/papers/2606.30111 Published on Jul 3

·

Submitted byhttps://huggingface.co/JianZhou0420

Zhouon Jul 9

Abstract

Automated agent architecture search demonstrates potential for improving embodied agent performance while revealing challenges related to optimization signals, local optima, and credit assignment in simulation-based training.

Embodied agentsare typically built as hand-designed compositions of perception, memory, planning, and action modules. This modularity exposes a large architectural design space, but current systems still rely on researcher intuition to choose where information is stored, how observations are processed, and how model calls are connected.Agent Architecture Search(AAS) automates such design for text-domain agents, but has not been systematically evaluated onperceptual embodied agentsthroughsimulator rollouts. We study this transfer. We introduceAgentCanvas, atyped-graph runtimethat hosts embodied executors as editable node-and-wire programs with simulator-aware execution and episode-level logs, andKDLoop, a coding-agent search procedure that cycles through proposal, critique, experiment, and distillation, with triggered reflection after stalls. We evaluate three AAS variants across four embodied executors spanningvision-language navigation,embodied question answering, andlanguage-conditioned manipulation. The resulting 3x4 matrix shows that architecture-level search can produce deployable and directional success-rate gains on embodied tasks, while one apparent high-scoring candidate is rejected as leak-bearing. At the same time, the experiments expose constraints that are muted in text-domain AAS:optimization signalscan be masked byrollout noise, search can become trapped inlocal edit basins, andepisode-level credit assignmentonly partially emerges even when detailed logs are available. These results characterize both the promise and the current limits of automated architecture search forembodied agents.

View arXiv pageView PDFProject pageGitHub33Add to collection

Get this paper in your agent:

hf papers read 2606\.30111

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.30111 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.30111 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.30111 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.