TT-VidT is a video pretraining method that decouples the temporal axis by combining a per‑frame ViT-B/16 spatial encoder with a compact Temporal Transfer Layer trained via Diff Compression. The authors conduct a systematic 24‑configuration study to isolate architecture, objective, and decoder effects, showing that the full TT-VidT design yields the strongest motion‑sensitive representations. In downstream fine‑tuning, TT‑VidT outperforms state‑of‑the‑art baselines on Jester, Something‑Something V2, ARID, and Diving48 while using significantly fewer encoder FLOPs.
By Shih-Ying Yeh, Daniel Z. Kaplan, Xuehai Wang, Fu-En Yang, Min-Hung Chen, Shang-Hong Lai
arXiv:2609.40347v1 Announce Type: new
Abstract: We introduce VideoMSN, a Masked Siamese Network framework for efficient self-supervised spatio-temporal representation learning in videos. Instead of r...
By Owais Iqbal, Sudipta Sarkar, Shyam Marjit, Omprakash Chakraborty, Anirban Chakraborty, Abir Das
arXiv:2609.22674v1 Announce Type: cross
Abstract: Joint Embedding Predictive Architectures (JEPAs) are becoming a core representation-learning primitive and a building block for latent world models a...
By Md Musfiqur Rahman Sanim, Zhihao Shu, Bahram Afsharmanesh, Amirali Mirian, Wei Niu, Gagan Agrawal
arXiv:2609.40333v1 Announce Type: new
Abstract: Self-supervised learning draws inspiration from infant visual development, yet standard training pipelines bear little resemblance to it: images are in...
By Ivan Martinovi\'c, Lukas Knobel, Yuki M. Asano
ViRDM is a new post‑training method for few‑step causal video generation that eliminates the need for a large teacher model and an online critic. By applying representation distribution matching (RDM) with a precomputed target distribution, a lightweight VAE decoder, and staged vector–Jacobian products, ViRDM overcomes memory, optimization, and temporal dynamics challenges. The approach reduces GPU memory usage and training time, achieving state‑of‑the‑art VBench performance with only 20 generator updates and 16 A100 GPU‑hours.
By Zichong Meng, Chongjian Ge, Chun-Hao P. Huang, Yang Zhou, Huaizu Jiang
The paper reports a controlled study of self‑supervised learning (SSL) objectives for image and video pretraining under limited data, architecture, and compute budgets. It compares contrastive, reconstruction, feature‑prediction, and diffusion methods, finding that DINOv2‑style pretraining delivers the best overall performance. Combining DINOv2 with video SSL objectives such as VideoMAE improves image classification and segmentation but harms video tracking and camera‑pose estimation, highlighting a trade‑off between semantic and geometric learning.
By Brun\'o B. Englert, Gijs Dubbelman
Latent video generation relies on autoencoders to define a compact space in which generative models operate. Although video autoencoder architectures have evolved substantially, their latent spaces are still optimized primarily for pixel-level reconstruction and provide limited high-level semantic organization.
The paper introduces a zero‑shot subject‑driven video generation framework that eliminates the need for per‑subject tuning and large subject‑video datasets. It achieves this by separating identity injection—learned from subject‑image pairs—and motion‑awareness preservation—maintained with a small set of arbitrary videos, and optimizes both with stochastic switching and dropout techniques. Using CogVideoX‑5B, the method adapts a single model with only 200K subject‑image pairs and 4,000 arbitrary videos in 288 A100 GPU hours, representing roughly 1% of the compute required by previous zero‑shot baselines while preserving subject fidelity and motion quality.
By Daneul Kim, Jingxu Zhang, Wonjoon Jin, Sunghyun Cho, Qi Dai, Jaesik Park, Chong Luo
arXiv:2609.37250v1 Announce Type: cross
Abstract: World-action models (WAMs) couple future visual-state prediction with action generation. By adapting video generators or image-editing models pretrai...
By Yang Zhang, Jiangyuan Zhao, Chenyou Fan, Jiayu Hu, Xiu Yuan, Chenjia Bai, Xiu Li
Few-step autoregressive (AR) video diffusion enables low-latency streaming generation, but existing post-training methods predominantly rely on Distribution Matching Distillation (DMD), requiring both...
arXiv:2605. 23045v2 Announce Type: replace-cross Abstract: Video representation learning has seen tremendous progress in recent years.
By Mantas Skackauskas, Xinyue Hao, Laura Sevilla-Lara
The paper revisits the use of contrastive loss for post‑training CLIP models, showing that the reported catastrophic forgetting is mainly due to an overly large temperature parameter in the InfoNCE objective. By setting a smaller temperature, contrastive post‑training can actually improve CLIP. The authors introduce ComCLIP, a lightweight single‑epoch recipe that freezes the text encoder, trains the vision encoder with a properly tempered contrastive loss, an MSE anchoring loss, and a relational distillation loss, achieving comparable zero‑shot classification to CLIP‑Refine while significantly enhancing visual feature transferability and maintaining downstream compatibility with LLaVA.
By Zidan Wang, Yaqian Li, Xiaokai Zhang, Kaiwen Long, Kun He, Hanpeng Liu