arXiv Machine Learning August 31, 2026 By Yuan-Chen Shu, Zhiwei Lin, Xiaoyu Zhou, Yongtao Wang

RegCL: Compact Continual SAM Adaptation for Visual Grounding in Multi-Sensorial Media

Read the original on arXiv Machine Learning →

RegCL is a non‑replay continual learning framework that adapts the Segment Anything Model (SAM) for visual grounding across evolving multi‑sensorial media domains. It consolidates domain‑specific segmentation knowledge into a single lightweight SAM adapter by incrementally merging LoRA‑style AugModules and preserving compact historical feature statistics. Experiments on five heterogeneous datasets demonstrate that RegCL retains performance while adapting to new domains, outperforming other non‑replay continual learning and merging baselines.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv Computer Vision

Aug 24

ProtoFlow: Mitigating Forgetting in Class-Incremental Remote Sensing Segmentation via Low-Curvature Prototype Flow

arXiv:2604.03212v4 Announce Type: replace Abstract: Remote sensing segmentation in real deployment is inherently continual: new semantic categories emerge, and acquisition conditions shift across sea...

By Jiekai Wu, Rong Fu, Chuangqi Li, Zijian Zhang, Guangxin Wu, Hao Zhang, Shiyin Lin, Yang Li, Dongxu Zhang, Amir H. Gandomi, Simon Fong, Pengbin Feng

computer-vision benchmarks

arXiv Computer Vision

Aug 25

Sa2VA: Marrying SAM2 with MLLM for Dense Grounded Understanding of Images and Videos

arXiv:2501.04001v4 Announce Type: replace Abstract: This work presents Sa2VA, the first comprehensive, unified model for dense grounded understanding of both images and videos. Unlike existing multi-...

By Haobo Yuan, Xiangtai Li, Tao Zhang, Yueyi Sun, Zilong Huang, Shilin Xu, Shunping Ji, Yunhai Tong, Lu Qi, Jiashi Feng, Ming-Hsuan Yang

llms computer-vision fine-tuning multimodal benchmarks

arXiv Machine Learning

Jul 31

Continual Learning with Vision-Language Models via Semantic-Geometry Preservation

arXiv:2603. 12055v3 Announce Type: replace-cross Abstract: Continual learning of pretrained vision-language models (VLMs) is prone to catastrophic forgetting, yet current approaches adapt to new tasks without explicitly preserving the cross-modal semantic geometry inherited from pretraining and previous stages, allowing new-task supervision to induce geometric distortion.

By Chiyuan He, Zihuan Qiu, Fanman Meng, Runtong Zhang, Linfeng Xu, Qingbo Wu, Hongliang Li

llms efficiency multimodal benchmarks safety

arXiv AI

Jul 14

Lifelong Representations: A Survey on Continual Self-Supervised Learning for Vision Models

arXiv:2607. 09785v1 Announce Type: cross Abstract: Traditionally, continual learning has assumed access to labeled data, yet many real-world applications -- such as lifelong robotics -- require models to adapt continuously from unlabeled streams.

By Sergi Masip, Alicja Dobrzeniecka, Jonathan Swinnen, Joachim Collin, Bart{\l}omiej Twardowski, Szymon {\L}ukasik, Tinne Tuytelaars

robotics efficiency multimodal benchmarks

arXiv Computer Vision

1d ago

TempoGround: State-Aware Streaming Visual Grounding with Vision-Language Models

arXiv:2609.02359v1 Announce Type: new Abstract: Visual grounding maps language referents to spatial targets and is central to open-vocabulary perception with vision-language models. Existing methods...

By Leqian Ding, Junning Qiu, Manwen Yang, Yu Guo, Fei Wang

llms multimodal benchmarks

arXiv Machine Learning

Jun 4

Stateful Visual Encoders for Vision-Language Models

arXiv:2606. 04433v1 Announce Type: cross Abstract: Vision-language models (VLMs) are increasingly used in multi-image, multi-turn agentic settings where decisions depend on visual changes.

By Zirui Wang, Junwei Yu, Adam Yala, David M. Chan, Joseph E. Gonzalez, Trevor Darrell

llms agents fine-tuning multimodal

One email a morning, machine-written

Related stories

ProtoFlow: Mitigating Forgetting in Class-Incremental Remote Sensing Segmentation via Low-Curvature Prototype Flow

Sa2VA: Marrying SAM2 with MLLM for Dense Grounded Understanding of Images and Videos

Continual Learning with Vision-Language Models via Semantic-Geometry Preservation

Lifelong Representations: A Survey on Continual Self-Supervised Learning for Vision Models

TempoGround: State-Aware Streaming Visual Grounding with Vision-Language Models

Stateful Visual Encoders for Vision-Language Models