Hefei University of Technology（合肥工业大学）；Jianghuai Advance Technology Center（江淮先进技术中心）；Anhui Provincial Key Laboratory of Humanoid Robots（安徽省人形机器人重点实验室）；Kaiyang Laboratory, Chery（凯阳实验室，中国奇瑞）；Zhejiang University of Technology（浙江工业大学）
Dataset distillation (DD) condenses large corpora into compact, information-rich subsets for efficient training and reuse. However, under noisy supervision, DD risks condensing corrupted associations together with useful signals, degrading robustness. Conventional noisy-label remedies (sample selection, loss weighting, label correction) tightly couple noise estimation with model optimization, often require clean anchors, and can amplify confirmation bias-assumptions that are misaligned with DD's goal of compact, plug-and-play supervision. We therefore propose a trajectory-based DD framework that jointly suppresses noise and preserves transferable knowledge without relabeling or clean subsets. It comprises two complementary components: Selective Guidance Reweighting (SGR), which fuses global forgetting patterns (second-split forgetting) with local neighborhood consistency into a progressive reweighting scheme that prioritizes clean supervision along the teacher trajectory; and Teacher-Inspired Auxiliary Targets (TIAT), which inject auxiliary residual guidance distilled from intermediate teacher dynamics to reinforce informative signals while remaining internally consistent. Together, SGR and TIAT produce distilled datasets with cleaner and richer representations under noisy supervision. The framework is robust, label-preserving, computationally lightweight, and broadly applicable, yielding consistent gains over state-of-the-art DD baselines across symmetric, asymmetric, and real-world noise.
The Chinese University of Hong Kong（香港中文大学）；Nanyang Technological University（南洋理工大学）
Knowledge distillation, which involves extracting the "dark knowledge" from a teacher network to guide the learning of a student network, has emerged as an important technique for model compression and transfer learning. Unlike previous works that exploit architecture-specific cues such as activation and attention for distillation, here we wish to explore a more general and model-agnostic approach for extracting "richer dark knowledge" from the pre-trained teacher model. We show that the seemingly different self-supervision task can serve as a simple yet powerful solution. For example, when pe…
State Grid Smart Grid Research Institute Co., Ltd.（国网智能电网研究院有限公司）；Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）；University of Chinese Academy of Sciences（中国科学院大学）
Recently, many semi-supervised object detection (SSOD) methods adopt teacher-student framework and have achieved state-of-the-art results. However, the teacher network is tightly coupled with the student network since the teacher is an exponential moving average (EMA) of the student, which causes a performance bottleneck. To address the coupling problem, we propose a Cycle Self-Training (CST) framework for SSOD, which consists of two teachers T1 and T2, two students S1 and S2. Based on these networks, a cycle self-training mechanism is built, i.e., S1${\rightarrow}$T1${\rightarrow}$S2${\righta…
When Data-Free Knowledge Distillation Meets Non-Transferable Teacher: Escaping Out-of-Distribution Trap is All You Need
Sydney AI Centre, The University of Sydney（悉尼人工智能中心，悉尼大学）；School of Computer Science, Wuhan University（武汉大学计算机学院）；Department of Computer Science, Hong Kong Baptist University（香港 Baptist 大学计算机系）
This paper tackles the problem of semi-supervised video object segmentation on resource-constrained devices, such as mobile phones. We formulate this problem as a distillation task, whereby we demonstrate that small space-time-memory networks with finite memory can achieve competitive results with state of the art, but at a fraction of the computational cost (32 milliseconds per frame on a Samsung Galaxy S22). Specifically, we provide a theoretically grounded framework that unifies knowledge distillation with supervised contrastive representation learning. These models are able to jointly bene…
Queen Mary University of London（伦敦玛丽女王大学）；Vision Semantics Ltd（视觉语义有限公司）
Knowledge distillation is effective to train small and generalisable network models for meeting the low-memory and fast running requirements. Existing offline distillation methods rely on a strong pre-trained teacher, which enables favourable knowledge discovery and transfer but requires a complex two-phase training procedure. Online counterparts address this limitation at the price of lacking a highcapacity teacher. In this work, we present an On-the-fly Native Ensemble (ONE) strategy for one-stage online distillation. Specifically, ONE trains only a single multi-branch network while simultan…
Seoul National University（首尔国立大学）；Qualcomm AI Research（高通人工智能研究院）；Qualcomm Technologies, Inc.（高通技术公司）
Quantization and Knowledge distillation (KD) methods are widely used to reduce memory and power consumption of deep neural networks (DNNs), especially for resource-constrained edge devices. Although their combination is quite promising to meet these requirements, it may not work as desired. It is mainly because the regularization effect of KD further diminishes the already reduced representation power of a quantized model. To address this short-coming, we propose Quantization-aware Knowledge Distillation (QKD) wherein quantization and KD are care-fully coordinated in three phases. First, Self-…
Overcoming Language Priors for Visual Question Answering Based on Knowledge Distillation
School of Computer Science and Technology, Huazhong University of Science and Technology（华中科技大学计算机科学与技术学院）；Joint Laboratory of HUST and Pingan Property & Casualty Research (HPL)（华中科技大学与平安财产保险研究联合实验室）
Transformer-Driven Modeling of Variable Frequency Features for Classifying Student Engagement in Online Learning
Malaviya National Institute of Technology Jaipur（斋浦尔马维亚国家理工学院）；Thapar Institute of Engineering and Technology（塔帕尔工程技术学院）；SRM University AP（SRM 大学阿普拉分校）
Beijing Electronic Science and Technology Institute（北京电子科技学院）；University of California, Davis（加州大学戴维斯分校）；National Institute of Technology Patna（巴特那国家理工学院）
With the continuous development of social software and multimedia technology, images have become a kind of important carrier for spreading information and socializing. How to evaluate an image comprehensively has become the focus of recent researches. The traditional image aesthetic assessment methods often adopt single numerical overall assessment scores, which has certain subjectivity and can no longer meet the higher aesthetic requirements. In this paper, we construct an new image attribute dataset called aesthetic mixed dataset with attributes(AMD-A) and design external attribute features …
Tsinghua Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院）；Tencent AI Lab（腾讯人工智能实验室）；The Chinese University of Hong Kong（香港中文大学）
Visual-Advantage On-Policy Distillation for Vision-Language Models
Institute of Automation, CAS（中国科学院自动化研究所）；School of Advanced Interdisciplinary Sciences, UCAS（中国科学院大学(UCAS)先进交叉学科学院）；Hello Group Inc.（Hello集团有限公司）；Sun Yat-sen University（中山大学）
Two-Stage Hybrid Supervision Framework for Fast, Low-resource, and Accurate Organ and Pan-cancer Segmentation in Abdomen CT
Beijing University of Posts and Telecommunications（北京邮电大学）；Dalian University of Technology（大连理工大学）；Guilin University of Electronic Technology（桂林电子科技大学）
