Cover Image

KVPO: ODE-Native GRPO for Autoregressive Video Alignment via KV Semantic Exploration

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Zhang, Ruicheng, Cong, Kaixi, Zhou, Jun, Zhong, Zhizhou, Xu, Zunnan, Mao, Shuiyang, Liu, Wei, Li, Xiu
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Forcing-KV: Hybrid KV Cache Compression for Efficient Autoregressive Video Diffusion Models
by: Ji, Yicheng, et al.
Published: (2026)

RoboStereo: Dual-Tower 4D Embodied World Models for Unified Policy Optimization
by: Zhang, Ruicheng, et al.
Published: (2026)

MIND-V: Hierarchical World Model for Long-Horizon Robotic Manipulation with RL-based Physical Alignment
by: Zhang, Ruicheng, et al.
Published: (2025)

Zo3T: Zero-Shot 3D-Aware Trajectory-Guided Image-to-Video Generation via Test-Time Training
by: Zhang, Ruicheng, et al.
Published: (2025)

Identity-Consistent Video Generation under Large Facial-Angle Variations
by: Hu, Bin, et al.
Published: (2026)

MixGRPO: Unlocking Flow-based GRPO Efficiency with Mixed ODE-SDE
by: Li, Junzhe, et al.
Published: (2025)

Alignment is All You Need: A Training-free Augmentation Strategy for Pose-guided Video Generation
by: Jin, Xiaoyu, et al.
Published: (2024)

Audio-visual Controlled Video Diffusion with Masked Selective State Spaces Modeling for Natural Talking Head Generation
by: Hong, Fa-Ting, et al.
Published: (2025)

Future Forcing: Future-aware Training-free KV Cache Policy for Autoregressive Video Generation
by: Luo, Jiayi, et al.
Published: (2026)

RAVEN: Real-time Autoregressive Video Extrapolation with Consistency-model GRPO
by: Lu, Yanzuo, et al.
Published: (2026)

AnyTalker: Scaling Multi-Person Talking Video Generation with Interactivity Refinement
by: Zhong, Zhizhou, et al.
Published: (2025)

AR-GRPO: Training Autoregressive Image Generation Models via Reinforcement Learning
by: Yuan, Shihao, et al.
Published: (2025)

Past- and Future-Informed KV Cache Policy with Salience Estimation in Autoregressive Video Diffusion
by: Chen, Hanmo, et al.
Published: (2026)

Flash-GRPO: Efficient Alignment for Video Diffusion via One-Step Policy Optimization
by: He, Xiaoxuan, et al.
Published: (2026)

DiverseGRPO: Mitigating Mode Collapse in Image Generation via Diversity-Aware GRPO
by: Liu, Henglin, et al.
Published: (2025)

AMS-KV: Adaptive KV Caching in Multi-Scale Visual Autoregressive Transformers
by: Xu, Boxun, et al.
Published: (2025)

HeatKV: Head-tuned KV-cache Compression for Visual Autoregressive Modeling
by: Cederlund, Jonathan, et al.
Published: (2026)

STAGE: Stable and Generalizable GRPO for Autoregressive Image Generation
by: Ma, Xiaoxiao, et al.
Published: (2025)

SAM-R1: Leveraging SAM for Reward Feedback in Multimodal Segmentation via Reinforcement Learning
by: Huang, Jiaqi, et al.
Published: (2025)

Neighbor GRPO: Contrastive ODE Policy Optimization Aligns Flow Models
by: He, Dailan, et al.
Published: (2025)

Controllable Layer Decomposition for Reversible Multi-Layer Image Generation
by: Liu, Zihao, et al.
Published: (2025)

REPARO: Compositional 3D Assets Generation with Differentiable 3D Layout Alignment
by: Han, Haonan, et al.
Published: (2024)

SARA: Semantically Adaptive Relational Alignment for Video Diffusion Models
by: Lian, Jiesong, et al.
Published: (2026)

VideoMLA: Low-Rank Latent KV Cache for Minute-Scale Autoregressive Video Diffusion
by: Yesiltepe, Hidir, et al.
Published: (2026)

DenseGRPO: From Sparse to Dense Reward for Flow Matching Model Alignment
by: Deng, Haoyou, et al.
Published: (2026)

PackCache: A Training-Free Acceleration Method for Unified Autoregressive Video Generation via Compact KV-Cache
by: Li, Kunyang, et al.
Published: (2026)

Context Forcing: Consistent Autoregressive Video Generation with Long Context
by: Chen, Shuo, et al.
Published: (2026)

Consistent123: One Image to Highly Consistent 3D Asset Using Case-Aware Diffusion Priors
by: Lin, Yukang, et al.
Published: (2023)

Beyond Pixels: Introducing Geometric-Semantic World Priors for Video-based Embodied Models via Spatio-temporal Alignment
by: Tang, Jinzhou, et al.
Published: (2025)

QuadGPT: Native Quadrilateral Mesh Generation with Autoregressive Models
by: Liu, Jian, et al.
Published: (2025)

Fine-Grained GRPO for Precise Preference Alignment in Flow Models
by: Zhou, Yujie, et al.
Published: (2025)

SafeGRPO: Self-Rewarded Multimodal Safety Alignment via Rule-Governed Policy Optimization
by: Rong, Xuankun, et al.
Published: (2025)

HaploOmni: Unified Single Transformer for Multimodal Video Understanding and Generation
by: Xiao, Yicheng, et al.
Published: (2025)

Focused Forcing: Content-Aware Per-Frame KV Selection for Efficient Autoregressive Video Diffusion
by: Cai, Peiliang, et al.
Published: (2026)

Weakly Supervised LiDAR Semantic Segmentation via Scatter Image Annotation
by: Chen, Yilong, et al.
Published: (2024)

Sparse Forcing: Native Trainable Sparse Attention for Real-time Autoregressive Diffusion Video Generation
by: Xu, Boxun, et al.
Published: (2026)

GRPO-RM: Fine-Tuning Representation Models via GRPO-Driven Reinforcement Learning
by: Xu, Yanchen, et al.
Published: (2025)

Next Block Prediction: Video Generation via Semi-Autoregressive Modeling
by: Ren, Shuhuai, et al.
Published: (2025)

VISTA: Enhancing Vision-Text Alignment in MLLMs via Cross-Modal Mutual Information Maximization
by: Li, Mingxiao, et al.
Published: (2025)

Match Stereo Videos via Bidirectional Alignment
by: Jing, Junpeng, et al.
Published: (2024)