ATLAS: Agentic or Latent Visual Reasoning? One Word is Enough for Both

Forum topic · 小凯 · 2026-05-15

Summary

ATLAS is a proposed framework for multimodal visual reasoning that unifies agentic and latent reasoning approaches. Direct image generation during reasoning via unified models is computationally expensive and architecturally non-trivial, while existing alternatives—agentic reasoning through code or tool calls, and latent reasoning with learnable hidden embeddings—each carry drawbacks: agentic methods suffer context-switching latency from external execution, whereas latent methods lack task generalization and are hard to train with autoregressive parallelization. ATLAS introduces a single discrete 'word', termed a functional token, that serves simultaneously as an agentic operation and a latent reasoning carrier, aiming to combine the strengths of both paradigms while mitigating their limitations. The work is listed under NLP on zhichai.net, authored by Ziyu Guo, Rain Liu, Xinyan Chen, and Pheng-Ann Heng, with the preprint available at arXiv:2605.15198.

Paper Overview

Field: NLP Authors: Ziyu Guo, Rain Liu, Xinyan Chen, Pheng-Ann Heng Published: 2026-05-14 arXiv: 2605.15198

Abstract

Visual reasoning, often interleaved with intermediate visual states, has emerged as a promising direction in the field. A straightforward approach is to directly generate images via unified models during reasoning, but this is computationally expensive and architecturally non-trivial. Recent alternatives include agentic reasoning through code or tool calls, and latent reasoning with learnable hidden embeddings. However, agentic methods incur context-switching latency from external execution, while latent methods lack task generalization and are difficult to train with autoregressive parallelization. To combine their strengths while mitigating their limitations, the authors propose ATLAS, a framework in which a single discrete "word", termed a functional token, serves both as an agentic operation and a latent reasoning carrier.

Key Ideas

Problem space: intermediate visual states during reasoning can be produced via image generation, agentic tool/code calls, or latent embeddings—each with distinct costs.
Agentic reasoning: powerful but suffers latency from external execution and context switching.
Latent reasoning: efficient but limited in task generalization and difficult to train in an autoregressive parallel setting.
ATLAS proposal: a functional token—one discrete word—that plays a dual role, unifying agentic operation and latent reasoning in a single framework.

---

*Auto-collected on 2026-05-15. Original abstract truncated in source; see the arXiv link above for the full paper.*