Anchored Tree Sampling converts horizon-compounding drift into anchor-bounded drift by organizing video generation as a sparse-to-dense tree of imputations instead of left-to-right autoregressive rollout.
hub Canonical reference
Reward Forcing: Efficient Streaming Video Generation with Rewarded Distribution Matching Distillation
Canonical reference. 91% of citing Pith papers cite this work as background.
23
Pith papers citing it
Background
91% of classified citations
abstract
Efficient streaming video generation is critical for simulating interactive and dynamic worlds. Existing methods distill few-step video diffusion models with sliding window attention, using initial frames as sink tokens to maintain attention performance and reduce error accumulation. However, video frames become overly dependent on these static tokens, resulting in copied initial frames and diminished motion dynamics. To address this, we introduce Reward Forcing, a novel framework with two key designs. First, we propose EMA-Sink, which maintains fixed-size tokens initialized from initial frames and continuously updated by fusing evicted tokens via exponential moving average as they exit the sliding window. Without additional computation cost, EMA-Sink tokens capture both long-term context and recent dynamics, preventing initial frame copying while maintaining long-horizon consistency. Second, to better distill motion dynamics from teacher models, we propose a novel Rewarded Distribution Matching Distillation (Re-DMD). Vanilla distribution matching treats every training sample equally, limiting the model's ability to prioritize dynamic content. Instead, Re-DMD biases the model's output distribution toward high-reward regions by prioritizing samples with greater dynamics rated by a vision-language model. Re-DMD significantly enhances motion quality while preserving data fidelity. We include both quantitative and qualitative experiments to show that Reward Forcing achieves state-of-the-art performance on standard benchmarks while enabling high-quality streaming video generation at 23.1 FPS on a single H100 GPU.
hub tools
citation-role summary
background 10
method 1
citation-polarity summary
years
2026 23representative citing papers
Echo-Forcing decouples stable anchors, compressed history, and recent dynamics in video diffusion KV caches using hierarchical memory, scene recall frames, and difference-aware decay to support interactive long video generation under bounded cache.
KVPO aligns streaming autoregressive video generators with human preferences via ODE-native GRPO, using KV cache for semantic exploration and TVE for velocity-based policy modeling, yielding gains in quality and alignment.
DMGD achieves better performance than fine-tuned SOTA methods in dataset distillation on ImageNet subsets by using semantic matching through conditional likelihood optimization and OT-based distribution matching in a training-free diffusion setup.
Stream-R1 improves distillation of autoregressive streaming video diffusion models by adaptively weighting supervision with a reward model at both rollout and per-pixel levels.
Hallo-Live achieves 20.38 FPS real-time text-to-audio-video avatar generation with 0.94s latency using asynchronous dual-stream diffusion and HP-DMD preference distillation, matching teacher model quality at 16x higher throughput.
A survey that groups efficient video diffusion methods into four paradigms—step distillation, efficient attention, model compression, and cache/trajectory optimization—and outlines open challenges for practical use.
The first survey on Attention Sink in Transformers structures the literature around fundamental utilization, mechanistic interpretation, and strategic mitigation.
Grounded Forcing introduces dual memory caching, reference-based positional embeddings, and proximity-weighted recaching to bridge stable semantics with local dynamics, improving long-range consistency in autoregressive video synthesis.
IAMFlow is a training-free identity-aware memory system that tracks entities via LLM global ID assignment and VLM frame verification to reduce identity drift in narrative long video generation from shifting prompts.
MIGA introduces two-stage alignment to close train-inference gaps and dual consistency enhancement via self-reflection and long-range guidance to achieve SOTA temporal consistency in infinite-frame video generation on VBench and NarrLV.
FashionChameleon achieves interactive multi-garment video customization in real time by training a teacher model with in-context learning on single-garment pairs, applying streaming distillation, and using training-free KV cache rescheduling.
Head Forcing assigns tailored KV cache strategies to local, anchor, and memory attention heads plus head-wise RoPE re-encoding to extend autoregressive video generation from seconds to minutes without training.
PhyMotion scores generated human videos by grounding recovered 3D poses in a physics simulator across kinematic, contact, and dynamic axes, yielding stronger human correlation and larger RL post-training gains than prior 2D rewards.
Stream-T1 is a test-time scaling framework for streaming video generation using scaled noise propagation from history, reward pruning across short and long windows, and feedback-guided memory sinking to improve temporal consistency and visual quality.
The paper introduces a unified framework for world models that fully incorporates all cognitive functions from Cognitive Architecture Theory, highlights under-researched areas in motivation and meta-cognition, and proposes Epistemic World Models as a new category for scientific discovery agents.
Hybrid Forcing combines linear temporal attention for long-range retention, block-sparse attention for efficiency, and decoupled distillation to achieve real-time unbounded 832x480 streaming video generation at 29.5 FPS.
Salt improves low-step video generation quality by adding endpoint-consistent regularization to distribution matching distillation and using cache-conditioned feature alignment for autoregressive models.
Video generation models can function as world simulators if efficiency gaps in spatiotemporal modeling are bridged via organized paradigms, architectures, and algorithms.
Rolling Sink is a training-free cache adjustment technique that maintains visual consistency in autoregressive video diffusion models for ultra-long open-ended generation beyond training horizons.
One-Forcing augments DMD with a GAN loss to enable stable one-step causal autoregressive video generation, reporting a VBench score of 83.76 as SOTA among one-step methods.
EchoTorrent combines multi-teacher distillation, adaptive CFG calibration, hybrid long-tail forcing, and VAE decoder refinement to enable few-pass autoregressive streaming video generation with improved temporal consistency and audio-lip sync.
LingBot-World is presented as an open-source world model that delivers high-fidelity simulation, minute-level contextual consistency, and real-time interactivity under one second latency.
citing papers explorer
-
Goodbye Drift: Anchored Tree Sampling for Long-Horizon Video-to-Video Generation
Anchored Tree Sampling converts horizon-compounding drift into anchor-bounded drift by organizing video generation as a sparse-to-dense tree of imputations instead of left-to-right autoregressive rollout.
-
Echo-Forcing: A Scene Memory Framework for Interactive Long Video Generation
Echo-Forcing decouples stable anchors, compressed history, and recent dynamics in video diffusion KV caches using hierarchical memory, scene recall frames, and difference-aware decay to support interactive long video generation under bounded cache.
-
KVPO: ODE-Native GRPO for Autoregressive Video Alignment via KV Semantic Exploration
KVPO aligns streaming autoregressive video generators with human preferences via ODE-native GRPO, using KV cache for semantic exploration and TVE for velocity-based policy modeling, yielding gains in quality and alignment.
-
DMGD: Train-Free Dataset Distillation with Semantic-Distribution Matching in Diffusion Models
DMGD achieves better performance than fine-tuned SOTA methods in dataset distillation on ImageNet subsets by using semantic matching through conditional likelihood optimization and OT-based distribution matching in a training-free diffusion setup.
-
Stream-R1: Reliability-Perplexity Aware Reward Distillation for Streaming Video Generation
Stream-R1 improves distillation of autoregressive streaming video diffusion models by adaptively weighting supervision with a reward model at both rollout and per-pixel levels.
-
Hallo-Live: Real-Time Streaming Joint Audio-Video Avatar Generation with Asynchronous Dual-Stream and Human-Centric Preference Distillation
Hallo-Live achieves 20.38 FPS real-time text-to-audio-video avatar generation with 0.94s latency using asynchronous dual-stream diffusion and HP-DMD preference distillation, matching teacher model quality at 16x higher throughput.
-
Efficient Video Diffusion Models: Advancements and Challenges
A survey that groups efficient video diffusion methods into four paradigms—step distillation, efficient attention, model compression, and cache/trajectory optimization—and outlines open challenges for practical use.
-
Attention Sink in Transformers: A Survey on Utilization, Interpretation, and Mitigation
The first survey on Attention Sink in Transformers structures the literature around fundamental utilization, mechanistic interpretation, and strategic mitigation.
-
Grounded Forcing: Bridging Time-Independent Semantics and Proximal Dynamics in Autoregressive Video Synthesis
Grounded Forcing introduces dual memory caching, reference-based positional embeddings, and proximity-weighted recaching to bridge stable semantics with local dynamics, improving long-range consistency in autoregressive video synthesis.
-
Advancing Narrative Long Video Generation via Training-Free Identity-Aware Memory
IAMFlow is a training-free identity-aware memory system that tracks entities via LLM global ID assignment and VLM frame verification to reduce identity drift in narrative long video generation from shifting prompts.
-
Enhancing Train-Free Infinite-Frame Generation for Consistent Long Videos
MIGA introduces two-stage alignment to close train-inference gaps and dual consistency enhancement via self-reflection and long-range guidance to achieve SOTA temporal consistency in infinite-frame video generation on VBench and NarrLV.
-
FashionChameleon: Towards Real-Time and Interactive Human-Garment Video Customization
FashionChameleon achieves interactive multi-garment video customization in real time by training a teacher model with in-context learning on single-garment pairs, applying streaming distillation, and using training-free KV cache rescheduling.
-
Head Forcing: Long Autoregressive Video Generation via Head Heterogeneity
Head Forcing assigns tailored KV cache strategies to local, anchor, and memory attention heads plus head-wise RoPE re-encoding to extend autoregressive video generation from seconds to minutes without training.
-
PhyMotion: Structured 3D Motion Reward for Physics-Grounded Human Video Generation
PhyMotion scores generated human videos by grounding recovered 3D poses in a physics simulator across kinematic, contact, and dynamic axes, yielding stronger human correlation and larger RL post-training gains than prior 2D rewards.
-
Stream-T1: Test-Time Scaling for Streaming Video Generation
Stream-T1 is a test-time scaling framework for streaming video generation using scaled noise propagation from history, reward pruning across short and long windows, and feedback-guided memory sinking to improve temporal consistency and visual quality.
-
Human Cognition in Machines: A Unified Perspective of World Models
The paper introduces a unified framework for world models that fully incorporates all cognitive functions from Cognitive Architecture Theory, highlights under-researched areas in motivation and meta-cognition, and proposes Epistemic World Models as a new category for scientific discovery agents.
-
Long-Horizon Streaming Video Generation via Hybrid Attention with Decoupled Distillation
Hybrid Forcing combines linear temporal attention for long-range retention, block-sparse attention for efficiency, and decoupled distillation to achieve real-time unbounded 832x480 streaming video generation at 29.5 FPS.
-
Salt: Self-Consistent Distribution Matching with Cache-Aware Training for Fast Video Generation
Salt improves low-step video generation quality by adding endpoint-consistent regularization to distribution matching distillation and using cache-conditioned feature alignment for autoregressive models.
-
Video Generation Models as World Models: Efficient Paradigms, Architectures and Algorithms
Video generation models can function as world simulators if efficiency gaps in spatiotemporal modeling are bridged via organized paradigms, architectures, and algorithms.
-
Rolling Sink: Bridging Limited-Horizon Training and Open-Ended Testing in Autoregressive Video Diffusion
Rolling Sink is a training-free cache adjustment technique that maintains visual consistency in autoregressive video diffusion models for ultra-long open-ended generation beyond training horizons.
-
One-Forcing: Towards Stable One-Step Autoregressive Video Generation
One-Forcing augments DMD with a GAN loss to enable stable one-step causal autoregressive video generation, reporting a VBench score of 83.76 as SOTA among one-step methods.
-
EchoTorrent: Towards Swift, Sustained, and Streaming Multi-Modal Video Generation
EchoTorrent combines multi-teacher distillation, adaptive CFG calibration, hybrid long-tail forcing, and VAE decoder refinement to enable few-pass autoregressive streaming video generation with improved temporal consistency and audio-lip sync.
-
Advancing Open-source World Models
LingBot-World is presented as an open-source world model that delivers high-fidelity simulation, minute-level contextual consistency, and real-time interactivity under one second latency.