Decide Before You Look: Learning Which Retrieved Memories Deserve Pixels
cs.CV, cs.AI
Submitted: 2026-10-06
Updated: 2026-10-06
Code: https://github.com/Liuziyu77/Valen
Sources
- Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection
- Qwen2.5-VL Technical Report
- Mem-Gallery: Benchmarking Multimodal Long-Term Conversational Memory for MLLM Agents
- SMMBench: A Benchmark for Source-Distributed Multimodal Agent Memory
- An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models
- MuRAG: Multimodal Retrieval-Augmented Generator for Open Question Answering over Images and Text
- Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering
- MemEye: A Visual-Centric Evaluation Framework for Multimodal Agent Memory
- ScreenQA: Large-Scale Question-Answer Pairs over Mobile App Screenshots
- LoRA: Low-Rank Adaptation of Large Language Models
- MRAG-Bench: Vision-Centric Evaluation for Retrieval-Augmented Multimodal Models
- M$^3$Exam: Benchmarking Multimodal Memory for Realistic User-Agent Interactions
- GQA: A New Dataset for Real-World Visual Reasoning and Compositional Question Answering
- Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexity
- Jev-Mem: System-One-Controlled Agentic Memory for Efficient AI Agents
- Retrieved but Not Delivered: Multimodal Memory Delivery for Long-Term Agents
- V-Mem: Modality-Routed Retrieval for Long-Term Multimodal Agentic Memory
- Seeing, Listening, Remembering, and Reasoning: A Multimodal Agent with Long-Term Memory
- Utility-Oriented Visual Evidence Selection for Multimodal Retrieval-Augmented Generation
- ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models