CORTEXA
← Browse
arxivcs.CVcs.CLcs.LG2026-07-01

StochasT: Learning with Stochastic Turn Depth for Visual Instruction Tuning

Yuan Qing, Chengzhi Mao, Boqing Gong

Large Vision-Language Models (LVLMs) rely extensively on Visual Instruction Tuning (VIT) to elicit their multimodal reasoning capabilities. However, we find a discrepancy: VIT often packs multiple language tasks about the same image for conversational, multi-turn training, whereas existing benchmarks evaluate LVLMs in isolated, single-turn scenarios. The models can suffer from visual attention decay and contextual overfitting during multi-turn training, making it hard for them to realize their full potential in the mismatched test phase. To close the gap, we propose learning with Stochastic Turn Depth (StochasT), which stochastically groups language tasks for the same image into clusters of varying sizes (turn depth) while preserving their organic order. Hence, while StochasT draws on Dropout and stochastic depth for ResNets, it does not actually drop anything to maximize the utility of the training data. Furthermore, we introduce a challenging, benchmark-agnostic evaluation mechanism based on the Balanced Latin Square to measure LVLMs' robustness under varying contextual dependencies. Extensive experiments demonstrate that StochasT effectively grants LVLMs strong, harmonized capabilities for both single-turn and multi-turn use cases.

View free PDFSource page

Related papers

arxivcs.LGcs.AIcs.CLcs.CV2026-07-05

DynaVieW: Schema-Guided World Modeling for Understanding Hierarchical Visual Dynamics

Silin Gao, Hao Zhao, Zeming Chen, Sepideh Mamooler, Antara Raaghavi Bhattacharya, Qiyu Wu, et al.

Multimodal LLMs struggle to systematically model the temporal evolution of visual scenes in videos or multi-image sequences. Such inputs require models to predict or simulate multiple levels of dynamic constituents, such as actions taken in the visual sequence, and the associated…

View free PDFSource page
arxivcs.CVcs.AIcs.CLcs.LG2026-07-09

Cognitive-structured Multimodal Agent for Multimodal Understanding, Generation, and Editing

Feng Wang, Canmiao Fu, Zhipeng Huang, Chen Li, Jing Lyu, Ge Li

Recent unified multimodal models show a single architecture can jointly perform vision/language understanding and image generation/editing. However, they repeatedly feed all historical visual and textual inputs into a shared context window, limiting long-horizon multimodal dialog…

View free PDFSource page
arxivcs.CVcs.CLcs.LG2026-06-26

DataComp-VLM: Improved Open Datasets for Vision-Language Models

Matteo Farina, Vishaal Udandarao, Thao Nguyen, Selim Kuzucu, Maximilian Böther, Andreas Hochlehnert, et al.

Building performant Vision-Language Models (VLMs) requires carefully curating large-scale training datasets, yet the community lacks systematic benchmarks for evaluating such curation strategies. We introduce DataComp for VLMs (DCVLM), a benchmark for controlled data-centric expe…

View free PDFSource page
arxivcs.LGcs.CLcs.CV2026-06-25

Just how sure are you? Improving Verbalized Uncertainty Calibration in Medical VQA

Eren Senoglu, Federico Toschi, Nicolo Brunello, Andrea Sassella, Mark James Carman

Multimodal large language models (MLLMs) applied to Medical Visual Question Answering (VQA) tend to produce overconfident outputs regardless of actual correctness, and existing verbalized confidence calibration methods, developed primarily for text only LLMs, do not account for t…

View free PDFSource page
arxivcs.ROcs.AIcs.CLcs.CVcs.LG2026-07-02

PhysMani: Physics-principled 3D World Model for Dynamic Object Manipulation

Peng Yun, Shouwang Huang, Hao Li, Jinxi Li, Jianan Wang, Bo Yang

Manipulating fast and dynamically moving targets in unstructured 3D environments remains challenging for embodied AI. Existing visual-language-action models and world models struggle with accurate 3D geometry and physically meaningful forecasting. We propose PhysMani, a framework…

View free PDFSource page
arxivcs.CVcs.AIcs.CLcs.CYcs.LG2026-06-30

Learning from Failure: Inference-Time Self-Improvement for Computer-Use Agents

Xueqiao Sun, Xiaohan Wang, Ludwig Schmidt, Serena Yeung-Levy, Yuhui Zhang

Computer-use agents, which leverage multimodal large language models (MLLMs) to operate computers and complete tasks, have attracted significant attention for their utility and versatility. A major challenge in developing these agents is collecting large-scale, high-quality traje…

View free PDFSource page