# MMDuet2: Enhancing Proactive Interaction of Video MLLMs with Multi-Turn Reinforcement Learning

arXiv: https://arxiv.org/abs/2512.06810 · HTML: https://mlsanity.com/paper/2512.06810/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Placed in Video understanding with MLLMs, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 30 of its 36 references that are in the map.
- First posted on arXiv 7 Dec 2025.

## Placed in

- Multimodal LLMs / vision-language models / [Video understanding with MLLMs](https://mlsanity.com/mllm/video-understanding-with-mllms/)

## Builds on (most cited first, 30 of 30)

- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling](https://mlsanity.com/paper/2412.05271/) (2024)
- [Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis](https://mlsanity.com/paper/2405.21075/) (2024)
- [MVBench: A Comprehensive Multi-modal Video Understanding Benchmark](https://mlsanity.com/paper/2311.17005/) (2023)
- [Qwen2.5-Omni Technical Report](https://mlsanity.com/paper/2503.20215/) (2025)
- [LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding](https://mlsanity.com/paper/2407.15754/) (2024)
- [LLaVA-Video: Video Instruction Tuning With Synthetic Data](https://mlsanity.com/paper/2410.02713/) (2024)
- [Long Context Transfer from Language to Vision](https://mlsanity.com/paper/2406.16852/) (2024)
- [Video-R1: Reinforcing Video Reasoning in MLLMs](https://mlsanity.com/paper/2503.21776/) (2025)
- [LMMs-Eval: Reality Check on the Evaluation of Large Multimodal Models](https://mlsanity.com/paper/2407.12772/) (2024)
- [VideoChat-R1: Enhancing Spatio-Temporal Perception via Reinforcement Fine-Tuning](https://mlsanity.com/paper/2504.06958/) (2025)
- [VideoLLM-online: Online Video Large Language Model for Streaming Video](https://mlsanity.com/paper/2406.11816/) (2024)
- [InternLM-XComposer-2.5: A Versatile Large Vision Language Model Supporting Long-Contextual Input and Output](https://mlsanity.com/paper/2407.03320/) (2024)
- [StreamingBench: Assessing the Gap for MLLMs to Achieve Streaming Video Understanding](https://mlsanity.com/paper/2411.03628/) (2024)
- [Dispider: Enabling Video LLMs with Active Real-Time Interaction via Disentangled Perception, Decision, and Reaction](https://mlsanity.com/paper/2501.03218/) (2025)
- [R1-Omni: Explainable Omni-Multimodal Emotion Recognition with Reinforcement Learning](https://mlsanity.com/paper/2503.05379/) (2025)
- [TimeChat-Online: 80% Visual Tokens are Naturally Redundant in Streaming Videos](https://mlsanity.com/paper/2504.17343/) (2025)
- [EgoExoLearn: A Dataset for Bridging Asynchronous Ego- and Exo-centric View of Procedural Activities in Real World](https://mlsanity.com/paper/2403.16182/) (2024)
- [Scaling RL to Long Videos](https://mlsanity.com/paper/2507.07966/) (2025)
- [Tarsier2: Advancing Large Vision-Language Models from Detailed Video Description to Comprehensive Video Understanding](https://mlsanity.com/paper/2501.07888/) (2025)
- [VideoLLM Knows When to Speak: Enhancing Time-Sensitive Video Comprehension with Video-Text Duet Interaction Format](https://mlsanity.com/paper/2411.17991/) (2024)
- [ViSpeak: Visual Instruction Feedback in Streaming Videos](https://mlsanity.com/paper/2503.12769/) (2025)
- [Tuning Large Multimodal Models for Videos using Reinforcement Learning from AI Feedback](https://mlsanity.com/paper/2402.03746/) (2024)
- [ProactiveVideoQA: A Comprehensive Benchmark Evaluating Proactive Interactions in Video Large Language Models](https://mlsanity.com/paper/2507.09313/) (2025)
- [What to Say and When to Say it: Live Fitness Coaching as a Testbed for Situated Interaction](https://mlsanity.com/paper/2407.08101/) (2024)
- [LiveCC: Learning Video LLM with Streaming Speech Transcription at Scale](https://mlsanity.com/paper/2504.16030/) (2025)
- [EgoSpeak: Learning When to Speak for Egocentric Conversational Agents in the Wild](https://mlsanity.com/paper/2502.14892/) (2025)

## Built on by (most cited first, 22 of 22)

- [AURA: Always-On Understanding and Real-Time Assistance via Video Streams](https://mlsanity.com/paper/2604.04184/) (2026)
- [Proact-VL: A Proactive VideoLLM for Real-Time AI Companions](https://mlsanity.com/paper/2603.03447/) (2026)
- [PhoStream: Benchmarking Real-World Streaming for Omnimodal Assistants in Mobile Scenarios](https://mlsanity.com/paper/2601.22575/) (2026)
- [VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding](https://mlsanity.com/paper/2607.14935/) (2026)
- [Em-Garde: A Propose-Match Framework for Proactive Streaming Video Understanding](https://mlsanity.com/paper/2603.19054/) (2026)
- [IPIBench: Evaluating Interactive Proactive Intelligence of MLLMs under Continuous Streams](https://mlsanity.com/paper/2605.27074/) (2026)
- [MOSS-VL Technical Report](https://mlsanity.com/paper/2608.15045/) (2026)
- [Don't Pause! Every prediction matters in a streaming video](https://mlsanity.com/paper/2604.24317/) (2026)
- [StreamPro: From Reactive Perception to Proactive Decision-Making in Streaming Video](https://mlsanity.com/paper/2605.16381/) (2026)
- [OmniInteract: Benchmarking Real-World Streaming Interaction for Real-Time Omnimodal Assistants](https://mlsanity.com/paper/2605.26485/) (2026)
- [Realtime-Venus: A full-duplex interaction system with asynchronous delegation](https://mlsanity.com/paper/2609.13814/) (2026)
- [From Static Inference to Dynamic Interaction: A Survey of Streaming Large Language Models](https://mlsanity.com/paper/2603.04592/) (2026)
- [Omni-DuplexEval: Evaluating Real-time Duplex Omni-modal Interaction](https://mlsanity.com/paper/2605.17360/) (2026)
- [Don't Pause: Streaming Video-Language Synchrony for Online Video Understanding](https://mlsanity.com/paper/2606.06991/) (2026)
- [GuideMe: Multi-Domain Task Guidance and Intervention in Streaming Video](https://mlsanity.com/paper/2607.02991/) (2026)
- [OneStreamer: Unifying Perception, Memory, and Proactive Response in Streaming Video Interaction](https://mlsanity.com/paper/2610.01762/) (2026)
- [ProactiveLLM: Learning Active Interaction for Streaming Large Language Models](https://mlsanity.com/paper/2606.00523/) (2026)
- [X-Stream: Exploring MLLMs as Multiplexers for Multi-Stream Understanding](https://mlsanity.com/paper/2606.02482/) (2026)
- [Live Assistant: Learning Whether, When, and Whom to Assist in Real-World Live Social Streams](https://mlsanity.com/paper/2609.27303/) (2026)
- [Foresight: planning future perception in streaming VLMs without retraining](https://mlsanity.com/paper/2610.03123/) (2026)
- [InteractionBench: A Real-Time Interaction Benchmark for Streaming Video Systems](https://mlsanity.com/paper/2610.05775/) (2026)
- [Proactive Physical AI through Continuous Observation and Timely Intervention](https://mlsanity.com/paper/2610.16967/) (2026)

## Abstract (arXiv)

Recent advances in video multimodal large language models (Video MLLMs) have significantly enhanced video understanding and multi-modal interaction capabilities. While most existing systems operate in a turn-based manner where the model can only reply after user turns, proactively deciding when to reply during video playback presents a promising yet challenging direction for real-time applications. In this work, we propose a novel text-to-text approach to proactive interaction, where the model autonomously determines whether to respond or remain silent at each turn based on dialogue history and visual context up to current frame of an streaming video. To overcome difficulties in previous methods such as manually tuning response decision thresholds and annotating precise reply times, we introduce a multi-turn RL based training method that encourages timely and accurate responses without requiring precise response time annotations. We train our model MMDuet2 on a dataset of 52k videos with two types of dialogues via SFT and RL. Experimental results demonstrate that MMDuet2 outperforms existing proactive Video MLLM baselines in response timing and quality, achieving state-of-the-art performance on the ProactiveVideoQA benchmark.