MIDAS: Multimodal Interactive Digital-humAn Synthesis via Real-time Autoregressive Video Generation

Recently, interactive digital human video generation has attracted widespread attention and achieved remarkable progress. However, building such a practical system that can interact with diverse input signals in real time remains challenging to existing methods, which often struggle with heavy computational cost and limited controllability. In this work, we introduce an autoregressive video generation framework that enables interactive multimodal control and low-latency extrapolation in a streaming manner. With minimal modifications to a standard large language model (LLM), our framework accepts multimodal condition encodings including audio, pose, and text, and outputs spatially and semantically coherent representations to guide the denoising process of a diffusion head. To support this, we construct a large-scale dialogue dataset of approximately 20,000 hours from multiple sources, providing rich conversational scenarios for training. We further introduce a deep compression autoencoder with up to 64 reduction ratio, which effectively alleviates the long-horizon inference burden of the autoregres-sive model. Condition-driven human video generation transforms static portraits into dynamic, interactive virtual avatars that synchronize speech with natural facial expressions, body movements, and emotional cues (Qi et al., 2025; Lin et al., 2025a; Xu et al., 2024). Such technologies enhance digital communication by making human-AI interactions more engaging and natural, and opens promising avenues for future applications such as virtual education and creative media.