{"ID":2846119,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.02360","arxiv_id":"2511.02360","title":"LaRe: Latent Refocusing for Multimodal Reasoning","abstract":"Chain of Thought (CoT) reasoning enhances logical performance by decomposing complex tasks, yet its multimodal extension faces a trade-off. The prevailing Thinking with Images paradigm achieves visual refocusing by explicitly cropping image regions, yet incurs rapidly growing computational overhead. The emerging line of latent-space reasoning reduces token consumption, but lacks the capacity for dynamic refocusing. We argue that this trade-off stems from a tacitly accepted premise that effective visual refocusing must occur in the form of explicit tokens. Building on this, we propose Latent Refocusing (LaRe), a new multimodal reasoning paradigm in which visual refocusing takes place entirely within the latent space. We further design a semantic augmentation training strategy that ensures the semantic structure of the latent space through visual reconstruction objective. Experimental evaluations demonstrate that LaRe improves average accuracy by 7.6% compared to existing baselines while reducing the number of tokens required for inference by 59.7%. When scaled to a 8B-parameter Vision-Language Model backbone, LaRe achieves performance comparable to state-of-the-art methods, demonstrating the efficacy of our proposed latent refocusing paradigm for multimodal reasoning.","short_abstract":"Chain of Thought (CoT) reasoning enhances logical performance by decomposing complex tasks, yet its multimodal extension faces a trade-off. The prevailing Thinking with Images paradigm achieves visual refocusing by explicitly cropping image regions, yet incurs rapidly growing computational overhead. The emerging line o...","url_abs":"https://arxiv.org/abs/2511.02360","url_pdf":"https://arxiv.org/pdf/2511.02360v4","authors":"[\"Jizheng Ma\",\"Xiaofei Zhou\",\"Geyuan Zhang\",\"Yanlong Song\",\"Han Yan\"]","published":"2025-11-04T08:28:46Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.CL\"]","methods":"[\"Language Model\"]","has_code":false}