LiveTalk: Real-Time Multimodal Interactive Video Diffusion via Improved On-Policy Distillation

• URL: http://arxiv.org/abs/2512.23576v1
• Date: Mon, 29 Dec 2025 16:17:36 GMT
• Title: LiveTalk: Real-Time Multimodal Interactive Video Diffusion via Improved On-Policy Distillation
• Authors: Ethan Chern, Zhulin Hu, Bohao Tang, Jiadi Su, Steffi Chern, Zhijie Deng, Pengfei Liu,
• Abstract summary: Real-time video generation via diffusion is essential for building general-purpose multimodal interactive AI systems.<n>Existing distillation methods can make the model autoregressive and reduce sampling steps to mitigate this.<n>This paper targets real-time interactive video diffusion conditioned on a multimodal context, including text, image, and audio, to bridge the gap.
• Score: 35.01134463094784
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Real-time video generation via diffusion is essential for building general-purpose multimodal interactive AI systems. However, the simultaneous denoising of all video frames with bidirectional attention via an iterative process in diffusion models prevents real-time interaction. While existing distillation methods can make the model autoregressive and reduce sampling steps to mitigate this, they focus primarily on text-to-video generation, leaving the human-AI interaction unnatural and less efficient. This paper targets real-time interactive video diffusion conditioned on a multimodal context, including text, image, and audio, to bridge the gap. Given the observation that the leading on-policy distillation approach Self Forcing encounters challenges (visual artifacts like flickering, black frames, and quality degradation) with multimodal conditioning, we investigate an improved distillation recipe with emphasis on the quality of condition inputs as well as the initialization and schedule for the on-policy optimization. On benchmarks for multimodal-conditioned (audio, image, and text) avatar video generation including HDTF, AVSpeech, and CelebV-HQ, our distilled model matches the visual quality of the full-step, bidirectional baselines of similar or larger size with 20x less inference cost and latency. Further, we integrate our model with audio language models and long-form video inference technique Anchor-Heavy Identity Sinks to build LiveTalk, a real-time multimodal interactive avatar system. System-level evaluation on our curated multi-turn interaction benchmark shows LiveTalk outperforms state-of-the-art models (Sora2, Veo3) in multi-turn video coherence and content quality, while reducing response latency from 1 to 2 minutes to real-time generation, enabling seamless human-AI multimodal interaction.

Related papers

• JavisDiT++: Unified Modeling and Optimization for Joint Audio-Video Generation [112.614973927778]
  Joint audio-video generation (JAVG) produces synchronized and semantically aligned sound and vision from textual descriptions.<n>This paper presents JavisDiT++, a framework for unified modeling and optimization of JAVG.<n>Our model achieves state-of-the-art performance merely with around 1M public training entries.
  arXiv  Detail & Related papers  (2026-02-22T12:44:28Z)
• GMS-CAVP: Improving Audio-Video Correspondence with Multi-Scale Contrastive and Generative Pretraining [64.72014392166625]
  GMS-CAVP is a novel framework that combines Multi-Scale Video-Audio Alignment and Multi-Scale Spatial-Temporal Diffusion-based pretraining objectives.<n>First, GMS-CAVP introduces a multi-scale contrastive learning strategy that captures semantic and temporal relations across varying granularities.<n>Second, we go beyond traditional contrastive learning by incorporating a diffusion-based generative objective, enabling modality translation and synthesis between video and audio.
  arXiv  Detail & Related papers  (2026-01-27T13:43:32Z)
• FlowAct-R1: Towards Interactive Humanoid Video Generation [37.04996721172613]
  FlowAct-R1 is a framework specifically designed for real-time interactive humanoid video generation.<n>Our framework achieves a stable 25 fps at 480p resolution with a time-to-first-frame (F) of only around 1.5 seconds.
  arXiv  Detail & Related papers  (2026-01-15T06:16:22Z)
• BindWeave: Subject-Consistent Video Generation via Cross-Modal Integration [56.98981194478512]
  We propose a unified framework that handles a broad range of subject-to-video scenarios.<n>We introduce an MLLM-DiT framework in which a pretrained multimodal large language model performs deep cross-modal reasoning to ground entities.<n>Experiments on the OpenS2V benchmark demonstrate that our method achieves superior performance across subject consistency, naturalness, and text relevance in generated videos.
  arXiv  Detail & Related papers  (2025-10-01T02:41:11Z)
• Hyper-Bagel: A Unified Acceleration Framework for Multimodal Understanding and Generation [19.010105652612616]
  Hyper-Bagel is designed to simultaneously speed up both multimodal understanding and generation tasks.<n>For generative tasks, our resulting 6-NFE model yields a 16.67x speedup in text-to-image generation and a 22x speedup in image editing.
  arXiv  Detail & Related papers  (2025-09-23T09:12:46Z)
• MIDAS: Multimodal Interactive Digital-humAn Synthesis via Real-time Autoregressive Video Generation [23.343080324521434]
  We introduce an autoregressive video generation framework that enables interactive multimodal control and low-latency extrapolation in a streaming manner.<n>Our framework accepts multimodal condition encodings including audio, pose, and text, and outputs spatially and semantically coherent representations.<n>To support this, we construct a large-scale dialogue dataset of approximately 20,000 hours from multiple sources.
  arXiv  Detail & Related papers  (2025-08-26T14:00:16Z)
• AlignDiT: Multimodal Aligned Diffusion Transformer for Synchronized Speech Generation [65.06374691172061]
  multimodal-to-speech task has gained increasing attention due to its wide range of applications, such as film production, dubbing, and virtual avatars.<n>Existing methods still suffer from limitations in speech intelligibility, audio-video synchronization, speech naturalness, and voice similarity to the reference speaker.<n>We propose AlignDiT, a multimodal Aligned Diffusion Transformer that generates accurate, synchronized, and natural-sounding speech from aligned multimodal inputs.
  arXiv  Detail & Related papers  (2025-04-29T10:56:24Z)
• DiVE: Efficient Multi-View Driving Scenes Generation Based on Video Diffusion Transformer [56.98400572837792]
  DiVE produces high-fidelity, temporally coherent, and cross-view consistent multi-view videos.<n>These innovations collectively achieve a 2.62x speedup with minimal quality degradation.
  arXiv  Detail & Related papers  (2025-04-28T09:20:50Z)
• AVadCLIP: Audio-Visual Collaboration for Robust Video Anomaly Detection [57.649223695021114]
  We present a novel weakly supervised framework that leverages audio-visual collaboration for robust video anomaly detection.<n>Our framework demonstrates superior performance across multiple benchmarks, with audio integration significantly boosting anomaly detection accuracy.
  arXiv  Detail & Related papers  (2025-04-06T13:59:16Z)
• VideoFusion: A Spatio-Temporal Collaborative Network for Multi-modal Video Fusion and Restoration [26.59510171451438]
  Existing multi-sensor fusion research predominantly integrates complementary from multiple images rather than videos.<n>VideoFusion exploits cross-modal complementarity and temporal dynamics to generate context-temporally coherent videos.<n>Extensive experiments reveal that VideoFusion outperforms existing image-oriented fusion paradigms in sequential scenarios.
  arXiv  Detail & Related papers  (2025-03-30T08:27:18Z)
• Efficient Long-duration Talking Video Synthesis with Linear Diffusion Transformer under Multimodal Guidance [39.94595889521696]
  LetsTalk is a diffusion transformer framework equipped with multimodal guidance and a novel memory bank mechanism.<n>In particular, LetsTalk introduces a noise-regularized memory bank to alleviate error accumulation and sampling artifacts during extended video generation.<n>We show that LetsTalk establishes new state-of-the-art in generation quality, producing temporally coherent and realistic talking videos.
  arXiv  Detail & Related papers  (2024-11-24T04:46:00Z)
• Live2Diff: Live Stream Translation via Uni-directional Attention in Video Diffusion Models [64.2445487645478]
  Large Language Models have shown remarkable efficacy in generating streaming data such as text and audio. We present Live2Diff, the first attempt at designing a video diffusion model with uni-directional temporal attention, specifically targeting live streaming video translation.
  arXiv  Detail & Related papers  (2024-07-11T17:34:51Z)
• Joint Multimodal Transformer for Emotion Recognition in the Wild [49.735299182004404]
  Multimodal emotion recognition (MMER) systems typically outperform unimodal systems. This paper proposes an MMER method that relies on a joint multimodal transformer (JMT) for fusion with key-based cross-attention.
  arXiv  Detail & Related papers  (2024-03-15T17:23:38Z)
• CMMD: Contrastive Multi-Modal Diffusion for Video-Audio Conditional Modeling [21.380988939240844]
  We introduce a multi-modal diffusion model tailored for the bi-directional conditional generation of video and audio. We propose a joint contrastive training loss to improve the synchronization between visual and auditory occurrences.
  arXiv  Detail & Related papers  (2023-12-08T23:55:19Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.