MemOCR: Layout-Aware Visual Memory for Efficient Long-Horizon Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shi, Yaorui, Liu, Shugui, Yang, Yu, Mao, Wenyu, Chen, Yuxin, GU, Qi, Su, Hui, Cai, Xunliang, Wang, Xiang, Zhang, An
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918508572442624
author Shi, Yaorui
Liu, Shugui
Yang, Yu
Mao, Wenyu
Chen, Yuxin
GU, Qi
Su, Hui
Cai, Xunliang
Wang, Xiang
Zhang, An
author_facet Shi, Yaorui
Liu, Shugui
Yang, Yu
Mao, Wenyu
Chen, Yuxin
GU, Qi
Su, Hui
Cai, Xunliang
Wang, Xiang
Zhang, An
contents Long-horizon agentic reasoning necessitates effectively compressing growing interaction histories into a limited context window. Most existing memory systems serialize history as text, where token-level cost is uniform and scales linearly with length, often spending scarce budget on low-value details. To this end, we introduce MemOCR, a multimodal memory agent that improves long-horizon reasoning under tight context budgets by allocating memory space with adaptive information density through visual layout. Concretely, MemOCR maintains a structured rich-text memory (e.g., headings, highlights) and renders it into an image that the agent consults for memory access, visually prioritizing crucial evidence while aggressively compressing auxiliary details. To ensure robustness across varying memory budgets, we train MemOCR with reinforcement learning under budget-aware objectives that expose the agent to diverse compression levels. Across long-context multi-hop and single-hop question-answering benchmarks, MemOCR outperforms strong text-based baselines and achieves more effective context utilization under extreme budgets.
format Preprint
id arxiv_https___arxiv_org_abs_2601_21468
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MemOCR: Layout-Aware Visual Memory for Efficient Long-Horizon Reasoning
Shi, Yaorui
Liu, Shugui
Yang, Yu
Mao, Wenyu
Chen, Yuxin
GU, Qi
Su, Hui
Cai, Xunliang
Wang, Xiang
Zhang, An
Artificial Intelligence
Long-horizon agentic reasoning necessitates effectively compressing growing interaction histories into a limited context window. Most existing memory systems serialize history as text, where token-level cost is uniform and scales linearly with length, often spending scarce budget on low-value details. To this end, we introduce MemOCR, a multimodal memory agent that improves long-horizon reasoning under tight context budgets by allocating memory space with adaptive information density through visual layout. Concretely, MemOCR maintains a structured rich-text memory (e.g., headings, highlights) and renders it into an image that the agent consults for memory access, visually prioritizing crucial evidence while aggressively compressing auxiliary details. To ensure robustness across varying memory budgets, we train MemOCR with reinforcement learning under budget-aware objectives that expose the agent to diverse compression levels. Across long-context multi-hop and single-hop question-answering benchmarks, MemOCR outperforms strong text-based baselines and achieves more effective context utilization under extreme budgets.
title MemOCR: Layout-Aware Visual Memory for Efficient Long-Horizon Reasoning
topic Artificial Intelligence
url https://arxiv.org/abs/2601.21468