paper-with-me

홈 › Papers

Cognitive-structured Multimodal Agent for Multimodal Understanding, Generation, and Editing

2026-07-09 · Feng Wang, Canmiao Fu, Zhipeng Huang, Chen Li, Jing Lyu, Ge Li arxiv

Recent unified multimodal models show a single architecture can jointly perform vision/language understanding and image generation/editing. However, they repeatedly feed all historical visual and textual inputs into a shared context window, limiting long-horizon multimodal dialogue due to visual token explosion and unreliable cross-turn referencing. We propose a Cognitive-structured Multimodal Agent that externalizes visual information into an Episodic Visual Memory and selectively reactivates relevant episodes during reasoning. The agent consists of a Perceptual Abstraction Engine for structured visual abstraction, a Cognitive Retrieval Engine for cross-turn memory retrieval, and a Multimodal Executive Controller for autonomous task inference and action planning. To address the lack of turn-level retrieval supervision in existing datasets, we develop a Unified Scenario Engine that programmatically generates structured multi-turn conversations with fine-grained retrieval annotations, enabling reinforcement learning to optimize abstraction and retrieval policies. We also construct a long-horizon visual-dialogue benchmark stratified by difficulty to evaluate episodic visual recall. Our 8B agent achieves 91.4% retrieval accuracy over 20-turn sessions, surpassing 32B baselines by +8.2% while nearly halving per-turn inference time (23.1s -> 12.7s). We further present the Cognitive-structured Multimodal Agent Harness (CMA-Harness), a tool-augmented deployment of the same cognitive structure integrating persistent multimodal memory, web access, image generation/editing/composition tools, and OpenAI-compatible serving. Structured memory and modular decision-making offer a more scalable, efficient paradigm for long-horizon multimodal agents than monolithic parameter scaling. Code: https://github.com/caseclose/cma-harness ; Project page: https://caseclose.github.io/cma-harness/

📄 PDF Abstract BibTeX arXiv:2607.08497

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement LearningImage Generation

Similar Papers 제목 키워드 기반

AUGUSTUS: An LLM-Driven Multimodal Agent System with Contextualized User Memory

2025-10-17 · Jitesh Jain, Shubham Maheshwari, Ning Yu, Wen-mei Hwu 외 arxiv

Riding on the success of LLMs with retrieval-augmented generation (RAG), there has been a growing interest in augmenting agent systems with external memory databases. However, the existing systems focus on storing text i…

Cognitive Chain-of-Thought (CoCoT): Structured Multimodal Reasoning about Social Situations

2025-07-27 · Eunkyu Park, Wesley Hanwen Deng, Gunhee Kim, Motahhare Eslami 외 arxiv

Chain-of-Thought (CoT) prompting helps models think step by step. But naive CoT breaks down in visually grounded social tasks, where models must perceive, understand, and judge all at once; bridging perception with norm-…

Instruction FollowingMultimodal Reasoning

Web-CogReasoner: Towards Multimodal Knowledge-Induced Cognitive Reasoning for Web Agents

2025-08-03 · Yuhan Guo, Cong Guo, Aiwen Sun, Hongliang He 외 arxiv

Multimodal large-scale models have significantly advanced the development of web agents, enabling perception and interaction with digital environments akin to human cognition. In this paper, we argue that web agents must…

OGA-AID: Clinician-in-the-loop AI Report Drafting Assistant for Multimodal Observational Gait Analysis in Post-Stroke Rehabilitation

2026-04-07 · Khoi T. N. Nguyen, Nghia D. Nguyen, Hui Yu Koh, Patrick W. H. Kwong 외 arxiv

Gait analysis is essential in post-stroke rehabilitation but remains time-intensive and cognitively demanding, especially when clinicians must integrate gait videos and motion-capture data into structured reports. We pre…

From Where Things Are to What They Are For: Benchmarking Spatial-Functional Intelligence in Multimodal LLMs

2026-05-04 · Le Zhang, Jihan Yang, Soundarya Krishnan, Jimit Majmudar 외 arxiv

Human-level agentic intelligence extends beyond low-level geometric perception, evolving from recognizing where things are to understanding what they are for. While existing benchmarks effectively evaluate the geometric …

Relational ReasoningSpatial Reasoning