PEARL: Personalized Streaming Video Understanding with Runtime-Defined Concepts
cs.CV, cs.AI, cs.IR
Submitted: 2026-03-20
Updated: 2026-11-18
Code: https://github.com/Yuanhong-Zheng/PEARL
Terminology
Sources
- InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
- Qwen3-VL Technical Report
- LLaVA-OneVision: Easy Visual Task Transfer
- MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe
- Long Context Transfer from Language to Vision
- MC-LLaVA: Multi-Concept Personalized Vision-Language Model
- Jarvis: Towards Personalized AI Assistant via Personal KV-Cache Retrieval
- UniCTokens: Boosting Personalized Understanding and Generation via Unified Concept Tokens
- Small-Large Collaboration: Training-efficient Concept Personalization for Large VLM using a Meta Personalized Small VLM
- MMPB: It's Time for Multi-Modal Personalization
- LLaVA-Video: Video Instruction Tuning With Synthetic Data
- Personalized Multimodal Large Language Models: A Survey
- Contextualized Visual Personalization in Vision-Language Models
- SlowBA: An efficiency backdoor attack towards VLM-based GUI agents
- RePIC: Reinforced Post-Training for Personalizing Multi-Modal Language Models
- M2A: Multimodal Memory Agent with Dual-Layer Hybrid Memory for Long-Term Personalized Interactions
- Unified Personalized Understanding, Generating and Editing
- Understanding vs. Generation: Navigating Optimization Dilemma in Multimodal Models
- Distribution Matching Variational AutoEncoder
- Streaming Video Question-Answering with In-context Video KV-Cache Retrieval
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs