16 papers
The Right Future for Action: Learning Action-Relevant Predictive States in World Action Models
Qiwen Gu, Jifan Li, Bingjie Gao +4
Generation-free world action models (WAMs) retain future-video prediction during training but act from internal video features at inference, leaving unclear what these features sho…
R2M-Bench: Evaluating Revisit Memory via Relative Consistency in Interactive Video World Models
Qiwen Gu, Bingjie Gao, Rui Chen +7
High similarity between first-visit and return frames does not necessarily show that a video world model remembered the scene; the intervening rollout may simply have changed very…
Why Struggle with Continuous Latents? Interpretable Discrete Latent Reasoning via Rendered Compression
Shuochen Chang, Qingyang Liu, Shaobo Wang +8
Large language models achieve high reasoning performance via explicit chain-of-thought and reinforcement learning, but require long output sequences and extended inference time. La…
In-context Region-based Drag: Drag Any Region to Any Shape
Jiacheng Sui, Tianyu Hao, Bingjie Gao +2
Diffusion models have shown promise in drag-style editing. Previous works mainly focus on point-based drag, which is inherently ambiguous. This paper focuses on region-based drag a…
PermaVid: Consistent Video Generation Across Edits via Disentangled Context Memory
Shuai Yang, Bingjie Gao, Ziwei Liu +3
Consistent video generation under editing operations requires persistence: when edits modify scene appearance or layout, subsequent generations should remain coherent across time a…
DreamX-World 1.0: A General-Purpose Interactive World Model
DreamX Team, Yancheng Bai, Rui Chen +20
DreamX-World 1.0 is a general-purpose interactive text/image-to-video world model for controllable long-horizon generation. It supports camera navigation, revisits to previously ob…
Breaking Dual Bottlenecks: Evolving Unified Multimodal Models into Self-Adaptive Interleaved Visual Reasoners
Qingyang Liu, Bingjie Gao, Canmiao Fu +9
Recent unified models integrate multimodal understanding and generation within a single framework. However, an "understanding-generation gap" persists, where models can capture use…
When Policy Entropy Constraint Fails: Preserving Diversity in Flow-based RLHF via Perceptual Entropy
Xiaofeng Tan, Jun Liu, Bin-Bin Gao +5
RLHF is widely used to align flow-matching text-to-image models with human preferences, but often leads to severe diversity collapse after fine-tuning. In RL, diversity is often as…
Large-Scale Universal Defect Generation: Foundation Models and Datasets
Yuanting Fan, Jun Liu, Bin-Bin Gao +5
Existing defect/anomaly generation methods often rely on few-shot learning, which overfits to specific defect categories due to the lack of large-scale paired defect editing data.…
Uni-Classifier: Leveraging Video Diffusion Priors for Universal Guidance Classifier
Yujie Zhou, Pengyang Ling, Jiazi Bu +2
In practical AI workflows, complex tasks often involve chaining multiple generative models, such as using a video or 3D generation model after a 2D image generator. However, distri…
RAPO++: Cross-Stage Prompt Optimization for Text-to-Video Generation via Data Alignment and Test-Time Scaling
Bingjie Gao, Qianli Ma, Xiaoxue Wu +9
Prompt design plays a crucial role in text-to-video (T2V) generation, yet user-provided prompts are often short, unstructured, and misaligned with training data, limiting the gener…
Human-Agent Collaborative Paper-to-Page Crafting
Qianli Ma, Siyu Wang, Yilin Chen +7
In the quest for scientific progress, communicating research is as vital as the discovery itself. Yet, researchers are often sidetracked by the manual, repetitive chore of building…
CineTrans: Learning to Generate Videos with Cinematic Transitions via Masked Diffusion Models
Xiaoxue Wu, Bingjie Gao, Yu Qiao +2
Despite significant advances in video synthesis, research into multi-shot video generation remains in its infancy. Even with scaled-up models and massive datasets, the shot transit…
AnimateScene: Camera-controllable Animation in Any Scene
Qingyang Liu, Bingjie Gao, Weiheng Huang +10
Recent advances in 3D scene reconstruction and 4D human animation have broadened adoption, but integrating the two remains difficult. Key challenges include placing humans at plaus…
The Devil is in the Prompts: Retrieval-Augmented Prompt Optimization for Text-to-Video Generation
Bingjie Gao, Xinyu Gao, Xiaoxue Wu +5
The evolution of Text-to-video (T2V) generative models, trained on large-scale datasets, has been marked by significant progress. However, the sensitivity of T2V generative models…
Object Placement for Anything
Bingjie Gao, Bo Zhang, Li Niu
Object placement aims to determine the appropriate placement (\emph{e.g.}, location and size) of a foreground object when placing it on the background image. Most previous works ar…