A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Unknown
Tune-A-Video enables one-shot text-to-video generation by tuning a pretrained image diffusion model on a single video.
Unknown
EchoWM is an omnimodal world model that generates 720p video and environmental sound in response to continuous navigation for enterable generative media.
Unknown
Molmo2 introduces open weights and data for vision-language models with video understanding and grounding, addressing efficient packing of crops and tokens.
Keming Wu, Baoyi Wang, Kaichen Zhang, et al.
StreamOPD is a post-training recipe combining verifiable streaming-video data, thinking-mode on-policy distillation, and instruct-mode deployment, improving streaming video understanding without inference-time memory or retrieval.
Xiaoyu Zhu, Xinke Deng, Suresh Taddewadikar, et al.
Introduces Internalized Visual Thinking (IVT), a post-training framework that predicts latent future frames during training to enable efficient, direct-answer video reasoning at inference, outperforming Visual CoT with 5x lower latency.
Hmrishav Bandyopadhyay, Xuanchi Ren, Zijian Huang, et al.
Introduces Context-Matched Distillation (CMD), a causal DMD framework that aligns teacher supervision with the information available during autoregressive video generation, improving control adherence and state-of-the-art performance.
Haodong Li, Shaoteng Liu, Tianyu Wang, et al.
Introduces Latent Dynamics Reasoning (LDR), a video world model that explicitly models latent dynamics via kinematic integration, achieving far better out-of-distribution extrapolation than video diffusion baselines.
Mukhtiar Ali, Harsh Dubey, Sugam Mishra, et al.
Introduces CLIP-CC-Bench, an evaluation suite for long-form video description using 90-second movie clips and expert-written paragraph references, with an ensemble of five LLM embeddings and coarse/fine-grained semantic matching to rank 17 video-lang
Unknown
DreamGen introduces a pipeline that uses state-of-the-art video generative models to create synthetic videos and extract pseudo-actions, enabling generalization in robot learning.
Unknown
This paper introduces text-to-video generation, a conditional video generation approach that produces videos from textual descriptions.
Unknown
Long Context Tuning (LCT) adapts a pretrained single-shot video generation model for scene-level video generation by extending its context length.
Unknown
Streamingt2v introduces a method for generating consistent, dynamic, and extendable long videos from text by conditioning on an initial image and adapting techniques for enhanced quality and resolution.