English

PRISM: Perception Reasoning Interleaved for Sequential Decision Making

Artificial Intelligence 2026-05-08 v1

Abstract

Scaling LLM-based embodied agents from text-only environments to complex multimodal settings remains a major challenge. Recent work identifies a perception-reasoning-decision gap in standalone Vision-Language Models (VLMs), which often overlook task-critical information. In this paper, we introduce PRISM, a framework that tightly couples perception (VLM) and decision (LLM) through a dynamic question-answer (DQA) pipeline. Instead of passively accepting the VLM's description, the LLM critiques it, probes the VLM with goal-oriented questions, and synthesizes a compact image description. This closed-loop interaction yields a sharp, task-driven understanding of the scene. We evaluate PRISM on the ALFWorld and Room-to-Room (R2R) benchmarks. We show that: (1) PRISM significantly outperforms state-of-the-art image-based models, (2) our Interactive goal-oriented perception pipeline yields systematic and substantial gains, and (3) PRISM is fully automatic, eliminating the need for handcrafted questions or answers.

Keywords

Cite

@article{arxiv.2605.05407,
  title  = {PRISM: Perception Reasoning Interleaved for Sequential Decision Making},
  author = {Mohamed Salim Aissi and Clemence Grislain and Clement Romac and Laure Soulier and Mohamed Chetouani and Olivier Sigaud and Nicolas Thome},
  journal= {arXiv preprint arXiv:2605.05407},
  year   = {2026}
}
R2 v1 2026-07-01T12:53:37.228Z