Paper
Seeing Farther and Smarter: Value-Guided Multi-Path Reflection for VLM Policy Optimization
Authors
Yanting Yang, Shenyuan Gao, Qingwen Bu, Li Chen, Dimitris N. Metaxas
Abstract
Solving complex, long-horizon robotic manipulation tasks requires a deep understanding of physical interactions, reasoning about their long-term consequences, and precise high-level planning. Vision-Language Models (VLMs) offer a general perceive-reason-act framework for this goal. However, previous approaches using reflective planning to guide VLMs in correcting actions encounter significant limitations. These methods rely on inefficient and often inaccurate implicit learning of state-values from noisy foresight predictions, evaluate only a single greedy future, and suffer from substantial inference latency. To address these limitations, we propose a novel test-time computation framework that decouples state evaluation from action generation. This provides a more direct and fine-grained supervisory signal for robust decision-making. Our method explicitly models the advantage of an action plan, quantified by its reduction in distance to the goal, and uses a scalable critic to estimate. To address the stochastic nature of single-trajectory evaluation, we employ beam search to explore multiple future paths and aggregate them during decoding to model their expected long-term returns, leading to more robust action generation. Additionally, we introduce a lightweight, confidence-based trigger that allows for early exit when direct predictions are reliable, invoking reflection only when necessary. Extensive experiments on diverse, unseen multi-stage robotic manipulation tasks demonstrate a 24.6% improvement in success rate over state-of-the-art baselines, while significantly reducing inference time by 56.5%.
Metadata
Related papers
Fractal universe and quantum gravity made simple
Fabio Briscese, Gianluca Calcagni • 2026-03-25
POLY-SIM: Polyglot Speaker Identification with Missing Modality Grand Challenge 2026 Evaluation Plan
Marta Moscati, Muhammad Saad Saeed, Marina Zanoni, Mubashir Noman, Rohan Kuma... • 2026-03-25
LensWalk: Agentic Video Understanding by Planning How You See in Videos
Keliang Li, Yansong Li, Hongze Shen, Mengdi Liu, Hong Chang, Shiguang Shan • 2026-03-25
Orientation Reconstruction of Proteins using Coulomb Explosions
Tomas André, Alfredo Bellisario, Nicusor Timneanu, Carl Caleman • 2026-03-25
The role of spatial context and multitask learning in the detection of organic and conventional farming systems based on Sentinel-2 time series
Jan Hemmerling, Marcel Schwieder, Philippe Rufin, Leon-Friedrich Thomas, Mire... • 2026-03-25
Raw Data (Debug)
{
"raw_xml": "<entry>\n <id>http://arxiv.org/abs/2602.19372v1</id>\n <title>Seeing Farther and Smarter: Value-Guided Multi-Path Reflection for VLM Policy Optimization</title>\n <updated>2026-02-22T22:53:16Z</updated>\n <link href='https://arxiv.org/abs/2602.19372v1' rel='alternate' type='text/html'/>\n <link href='https://arxiv.org/pdf/2602.19372v1' rel='related' title='pdf' type='application/pdf'/>\n <summary>Solving complex, long-horizon robotic manipulation tasks requires a deep understanding of physical interactions, reasoning about their long-term consequences, and precise high-level planning. Vision-Language Models (VLMs) offer a general perceive-reason-act framework for this goal. However, previous approaches using reflective planning to guide VLMs in correcting actions encounter significant limitations. These methods rely on inefficient and often inaccurate implicit learning of state-values from noisy foresight predictions, evaluate only a single greedy future, and suffer from substantial inference latency. To address these limitations, we propose a novel test-time computation framework that decouples state evaluation from action generation. This provides a more direct and fine-grained supervisory signal for robust decision-making. Our method explicitly models the advantage of an action plan, quantified by its reduction in distance to the goal, and uses a scalable critic to estimate. To address the stochastic nature of single-trajectory evaluation, we employ beam search to explore multiple future paths and aggregate them during decoding to model their expected long-term returns, leading to more robust action generation. Additionally, we introduce a lightweight, confidence-based trigger that allows for early exit when direct predictions are reliable, invoking reflection only when necessary. Extensive experiments on diverse, unseen multi-stage robotic manipulation tasks demonstrate a 24.6% improvement in success rate over state-of-the-art baselines, while significantly reducing inference time by 56.5%.</summary>\n <category scheme='http://arxiv.org/schemas/atom' term='cs.RO'/>\n <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n <published>2026-02-22T22:53:16Z</published>\n <arxiv:comment>ICRA 2026</arxiv:comment>\n <arxiv:primary_category term='cs.RO'/>\n <author>\n <name>Yanting Yang</name>\n </author>\n <author>\n <name>Shenyuan Gao</name>\n </author>\n <author>\n <name>Qingwen Bu</name>\n </author>\n <author>\n <name>Li Chen</name>\n </author>\n <author>\n <name>Dimitris N. Metaxas</name>\n </author>\n </entry>"
}