17 papers
MultiAlloc: Budget-Aware Precision Allocation for KV Cache Quantization
Yangchengyu Zhou, Zhiping Wu, Runchang Jiang +5
The Key-Value (KV) cache constitutes a dominant source of memory consumption during large language model (LLM) inference, particularly in long-context and complex reasoning scenari…
RGSQ: Riemannian Geometry-Sensitive Quantization for Large Vision-Language Models
Zhiping Wu, Dongdong Ren, Yangchengyu Zhou +4
Large vision-language models (VLMs) can be efficiently deployed under stringent memory and latency constraints through post training quantization (PTQ). However, most PTQ methods a…
BUS: Brain-Inspired Unsupervised Self-Reflection via Backward Prediction for Multimodal Reasoning
Jiacheng Yang, Tongying Xiao, Yunkai Dang +7
Current Vision-Language Models (VLMs) often struggle to handle complex visual tasks that require consistent and fine-grained reasoning. Recent methods aim to train models to facili…
The Shape of Addition: Geometric Structures of Arithmetic in Large Language Models
Liuyuan Wen, Xun Zhu, Lihao Huang +2
Large Language Models exhibit paradoxical fragility in fundamental arithmetic, implying a disconnect between internal computation and discrete output. By analyzing the residual str…
Selective, Regularized, and Calibrated: Harnessing Vision Foundation Models for Cross-Domain Few-Shot Semantic Segmentation
Junyuan Ma, Xunzhi Xiang, Wenbin Li +2
Vision foundation models (VFMs) have achieved strong performance across various vision tasks. However, it still remains challenging to apply VFMs for cross-domain few-shot segmenta…
Instinct vs. Reflection: Unifying Token and Verbalized Confidence in Multimodal Large Models
Yunkai Dang, Yifan Jiang, Yizhu Jiang +3
Multimodal Large Language Models (MLLMs) have demonstrated exceptional capabilities in various perception and reasoning tasks. Despite this success, ensuring their reliability in p…
UHR-BAT: Budget-Aware Token Compression Vision-Language model for Ultra-High-Resolution Remote Sensing
Yunkai Dang, Minxin Dai, Yuekun Yang +4
Ultra-high-resolution (UHR) remote sensing imagery couples kilometer-scale context with query-critical evidence that may occupy only a few pixels. Such vast spatial scale leads to…
CLASP: Class-Adaptive Layer Fusion and Dual-Stage Pruning for Multimodal Large Language Models
Yunkai Dang, Yizhu Jiang, Yifan Jiang +4
Multimodal Large Language Models (MLLMs) suffer from substantial computational overhead due to the high redundancy in visual token sequences. Existing approaches typically address…
Understanding and Enforcing Weight Disentanglement in Task Arithmetic
Shangge Liu, Yuehan Yin, Lei Wang +5
Task arithmetic provides an efficient, training-free way to edit pre-trained models, yet lacks a fundamental theoretical explanation for its success. The existing concept of ``weig…
Prompt-Free Universal Region Proposal Network
Qihong Tang, Changhan Liu, Shaofeng Zhang +3
Identifying potential objects is critical for object recognition and analysis across various computer vision applications. Existing methods typically localize potential objects by…
HART: High-Resolution Annotation-Free Reasoning Technique through a Closed-loop Framework
Jiacheng Yang, Anqi Chen, Yunkai Dang +5
Current Large Multimodal Models (LMMs) struggle with high-resolution visual inputs during the reasoning process, as the number of image tokens increases quadratically with resoluti…
FUSE-RSVLM: Feature Fusion Vision-Language Model for Remote Sensing
Yunkai Dang, Donghao Wang, Jiacheng Yang +7
Large vision-language models (VLMs) exhibit strong performance across various tasks. However, these VLMs encounter significant challenges when applied to the remote sensing domain…
A Benchmark for Ultra-High-Resolution Remote Sensing MLLMs
Yunkai Dang, Meiyi Zhu, Donghao Wang +7
Multimodal large language models (MLLMs) demonstrate strong perception and reasoning performance on existing remote sensing (RS) benchmarks. However, most prior benchmarks rely on…
SEA: Semantic Map Prediction for Active Exploration of Uncertain Areas
Hongyu Ding, Xinyue Liang, Yudong Fang +7
In this paper, we propose SEA, a novel approach for active robot exploration through semantic map prediction and a reinforcement learning-based hierarchical exploration policy. Unl…
Adapting In-Domain Few-Shot Segmentation to New Domains without Source Domain Retraining
Qi Fan, Kaiqi Liu, Nian Liu +4
Cross-domain few-shot segmentation (CD-FSS) aims to segment objects of novel classes in new domains, which is often challenging due to the diverse characteristics of target domains…
Robust Dataset Distillation by Matching Adversarial Trajectories
Wei Lai, Tianyu Ding, ren dongdong +4
Dataset distillation synthesizes compact datasets that enable models to achieve performance comparable to training on the original large-scale datasets. However, existing distillat…
Efficient Last-iterate Convergence Algorithms in Solving Games
Linjian Meng, Youzhi Zhang, Zhenxing Ge +6
To establish last-iterate convergence for Counterfactual Regret Minimization (CFR) algorithms in learning a Nash equilibrium (NE) of extensive-form games (EFGs), recent studies ref…