The paper introduces PuRF, a cache‑based method for multi‑label test‑time adaptation of vision‑language models. PuRF first performs region purification to reliably identify object‑specific regions, then uses these purified regions to build a discriminative, temporally refreshed cache that mitigates dominant‑label bias. Experiments show that PuRF outperforms existing methods, achieving a 4.05% mAP gain on ViT‑B/32 across five datasets.
By Yiwen Liang, Hui Chen, Yizhe Xiong, Mengyao Lyu, Yuhan Cao, Zijia Lin, Shuaicheng Niu, Sicheng Zhao, Jungong Han, Guiguang Ding
arXiv:2603. 00198v2 Announce Type: replace-cross Abstract: Token reduction accelerates long-video vision--language models (VLMs), but existing methods target Transformers, where reduction is treated as token pruning.
By Jindong Jiang, Amala Sanjay Deshmukh, Kateryna Chumachenko, Karan Sapra, Zhiding Yu, Guilin Liu, Andrew Tao, Pavlo Molchanov, Jan Kautz, Wonmin Byeon
LongVU‑TTT is a causal test‑time training method for long‑video multimodal large language models that inserts a convolutional resampler with fast‑weight updates between the vision encoder and the LLM. The fast weights adapt per video and contextualize frame features before compression, while a hybrid selector keeps explicit visual evidence for downstream reasoning. Experiments show that TTT‑Conv outperforms TTT‑MLP and bidirectional Mamba2 on MLVU, and beats attention‑ and fixed‑state recurrent resamplers on three benchmarks, achieving competitive results on five video‑understanding tasks after reducing 512 frames to 128 LLM frames.
By Mahmoud Ahmed, Sameh Abdulah, Olatunji Ruwase, Sam Ade Jacobs, Mathis Bode, Mohamed Elhoseiny
arXiv:2608.29395v1 Announce Type: new
Abstract: Vision-language models such as CLIP and SigLIP provide strong zero-shot recognition, but their predictions can degrade when deployed on target data tha...
By Pedram MohajerAnsari, Amir Salarpour, Run Wang, Mert D. Pes\'e
arXiv:2606. 13035v1 Announce Type: cross Abstract: Autoregressive video diffusion models provide a natural formulation for streaming and variable-length video generation by conditioning newly generated frames on previously generated content.
By Yu Meng, Xiangyang Luo, Letian Li, Wenyuan Jiang, Chen Gao, Xinlei Chen, Yong Li, Xiao-Ping Zhang
Token-Budget Distillation (TBD) is a parameter‑efficient fine‑tuning framework that adapts video vision‑language models to a fixed token budget. It freezes the pretrained backbone, updates only LoRA adapters, and incorporates FlashVID visual token compression. TBD uses a dual‑path teacher‑student design with full‑token supervision and compressed student optimization, enabling the student to recover full‑token semantics while remaining efficient under aggressive token reduction.
By Xiaoyang Guo, Guoping Luo, Jusheng Zhang, Keze Wang, Wenhao Wang
The paper introduces TTSD‑FAR, a test‑time self‑distillation framework that adapts large video‑language models to missing‑modality scenarios in emotion recognition. A frozen teacher trained on complete modalities guides a low‑rank student, while Fisher‑Anchored Restoration monitors Fisher information to prevent drift and restore the student when distribution shifts occur. Experiments on MELD, DFEW, and BAH with up to 50% missing modalities show TTSD‑FAR consistently outperforms entropy‑based adaptation, retrieval‑augmented generation, and perplexity‑based generation, maintaining performance over long adaptation horizons.
By Muhammad Haseeb Aslam, Alessandro Koerich, Marco Pedersoli, Ali Etemad, Eric Granger
Chehre is an emoji‑prompted video dataset designed to study perceptual flexibility in video language models. It contains 2,111 videos of 203 participants expressing 40 facial emojis, with each video annotated by about 30 perceivers, yielding 1,242 annotators in total. The dataset introduces a new task—distributional expression recognition—that evaluates a model’s ability to reproduce the variation seen in human annotations, and shows that persona prompting can shift model perception to better match human variability.
By Bita Azari, Zoe Stanley, Avneet Batra, Poorvi Bhatia, Hali Kil, Manolis Savva, Angelica Lim
arXiv:2606. 14299v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) such as CLIP have become a standard backbone for open-vocabulary recognition, yet their zero-shot predictions remain vulnerable to distribution shifts encountered at deployment.
By Jiazhen Huang, Xiao Chen, Zhiming Liu, Yaru Sun, Jingyan Jiang, Zhi Wang
The paper introduces MiRA, a plug‑in framework that reweights framewise attention in Vision Transformer video models to better capture subtle facial dynamics for expression recognition. MiRA computes frame‑level confidence and intra‑frame concentration from self‑attention maps, redistributing attention toward localized facial cues without adding trainable parameters. Two modes—an exact post‑softmax redistribution and a lightweight flashLite pre‑softmax approximation—are proposed, and experiments on facial expression recognition benchmarks show consistent gains over strong ViT baselines.
By Seongro Yoon, Donghyeon Cho, Jinsun Park, Fran\c{c}ois Br\'emond
Long-video MLLMs must model temporal change before a limited visual-token budget removes most frame evidence. We introduce LongVU-TTT, which inserts a convolutional Test-Time Training (TTT) resampler...
Autoregressive video diffusion models provide a natural formulation for streaming and variable-length video generation by conditioning newly generated frames on previously generated content. However, extending these models to minute-level generation remains challenging: the limited KV-cache budget prevents the model from retaining the full history, while repeatedly conditioning on self-generated frames induces a context distribution shift that accumulates over time, leading to visual artifacts, quality degradation, and temporal drift.