跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 4 天前AI 評分42

AVTR:以視覺想像實現自適應多模態推理

AVTR: Adaptive Multimodal Reasoning with Visual Imagination

AI 導讀

AVTR 結合生成式視覺反思與文字推理,並按輸入內容學習選擇推理方式,以適應不同的感知與推理需求。在 16 項基準測試中,AVTR 較 Qwen3.5 基線平均提升 4.5 分,同時將推理延遲降低 33%。擴展至較小規模的骨幹模型時,AVTR 亦勝過受評估的 2B 規模基線。

正文

Published on Nov 18

Authors:

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

Adaptive Visual-Textual Reasoning combines generative visual reflection with textual reasoning and learns to route between them per input, improving multimodal task accuracy while reducing inference latency.

Vision-Language Models increasingly rely on textual reasoning to solve advanced multimodal tasks. Despite its effectiveness, this paradigm faces two fundamental limitations. First, relying on textual reasoning alone can limit perceptual ability in vision-centric questions, where accurate answers depend on interpreting task-relevant visual evidence. Second, applying textual reasoning uniformly across inputs overlooks question-dependent variation in both reasoning modality and depth, often producing verbose and unnecessary traces when little logical inference is required. To this end, we propose Adaptive Visual-Textual Reasoning (AVTR), which introduces visual imagination as a complementary reasoning modality to textual thinking, enabling the model to revisit its visual observations and distill the salient evidence through generative visual reflection. Furthermore, we formulate reasoning-modality selection as a learned routing problem, allowing a single policy to adaptively harness the complementary strengths of visual reflection and textual reasoning in response to the perceptual and inferential demands of the input. Across 16 benchmarks, AVTR improves the average score by 4.5 points over Qwen3.5 baseline, while reducing inference latency by 33\%. It also scales effectively into a smaller size backbone, outperforming the evaluated 2B-scale baselines.

View arXiv page View PDF Project page GitHub Add to collection

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2611.24161 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2611.24161 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2611.24161 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.

來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co