检索结果-内蒙古大学图书馆

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Haihong Guo Fengxin Li Jiao Li Hongyan Liu School of Information Renmin University of China China Chinese Academy of Medical Sciences / Peking Union Medical College Institute of Medical Information / Medical Library China Key Laboratory of Data Engineering and Knowledge Engineering Ministry of Education China School of Economics and Management Tsinghua University China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Offline Reinforcement Learning (RL) optimizes policy using pre-collected data instead of direct environment interaction, offering a safe and cost-effective solution for sequential decision-making in the real world. However, it faces challenges such as distribution shift issues and vulnerability under perturbations. Researchers have developed various conservative methods to improve the robustness of offline RL. Nevertheless, model-based methods can result in transition distribution shift issues, while model-free value-based uncertainty penalty methods may not be sufficiently robust. To address these problems, we propose a new method called Robust Offline RL via Conservative Smoothing and Dynamics Controlling (RCSD). To achieve reliable value estimation of out-of-distribution (OOD) actions, RCSD uses both model-free uncertainty penalty and model-based simulation methods. It introduces a new one-step simulation method with conservative dynamics controlling to avoid value overestimation caused by transition distribution shifts. Moreover, RCSD considers both current and next states when generating OOD states to ensure cautious value estimation and efficient data utilization. RCSD uses conservative Q-smoothing and policy smoothing to strengthen the policy against sudden changes under perturbations. Experiments on D4RL benchmark demonstrate that RCSD can achieve state-of-the-art performance compared to baselines in either benchmark or adversarial attack tests.

关键词： Uncertainty Smoothing methods Perturbation methods Estimation Reinforcement learning Benchmark testing Signal processing Robustness Speech processing Faces

来源：评论

学校读者我要写书评

暂无评论

Switchable Online knowledge Distillation

arXiv

引用

arXiv 2022年

作者： Qian, Biao Wang, Yang Yin, Hongzhi Hong, Richang Wang, Meng Key Laboratory of Knowledge Engineering with Big Data Ministry of Education School of Computer Science and Information Engineering Hefei University of Technology China The University of Queensland

Online knowledge Distillation (OKD) improves the involved models by reciprocally exploiting the difference between teacher and student. Several crucial bottlenecks over the gap between them — e.g., Why and when does a large gap harm the performance, especially for student? How to quantify the gap between teacher and student? — have received limited formal study. In this paper, we propose Switchable Online knowledge Distillation (SwitOKD), to answer these questions. Instead of focusing on the accuracy gap at test phase by the existing arts, the core idea of SwitOKD is to adaptively calibrate the gap at training phase, namely distillation gap, via a switching strategy between two modes — expert mode (pause the teacher while keep the student learning) and learning mode (restart the teacher). To possess an appropriate distillation gap, we further devise an adaptive switching threshold, which provides a formal criterion as to when to switch to learning mode or expert mode, and thus improves the student’s performance. Meanwhile, the teacher benefits from our adaptive switching threshold and keeps basically on a par with other online arts. We further extend SwitOKD to multiple networks with two basis topologies. Finally, extensive experiments and analysis validate the merits of SwitOKD for classification over the state-of-the-arts. Our code is available at https://***/hfutqian/SwitOKD. Copyright © 2022, The Authors. All rights reserved.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Few-Shot Charge Prediction with Multi-grained Features and Mutual Information 20

Few-Shot Charge Prediction with Multi-grained Features and M...

引用

20th China National Conference on Computational Linguistics, CCL 2021

作者： Zhang, Han Dou, Zhicheng Zhu, Yutao Wen, Jirong School of Information Renmin University of China Beijing China Gaoling School of Artificial Intelligence Renmin University of China Beijing China Université de Montréal Montreal Canada Beijing Key Laboratory of Big Data Management and Analysis Methods Beijing China Key Laboratory of Data Engineering and Knowledge Engineering MOE Beijing China

ISBN: (纸本)9783030841850

Charge prediction aims to predict the final charge for a case according to its fact description and plays an important role in legal assistance systems. With deep learning based methods, prediction on high-frequency charges has achieved promising results but that on few-shot charges is still challenging. In this work, we propose a framework with multi-grained features and mutual information for few-shot charge prediction. Specifically, we extract coarse- and fine-grained features to enhance the model’s capability on representation, based on which the few-shot charges can be better distinguished. Furthermore, we propose a loss function based on mutual information. This loss function leverages the prior distribution of the charges to tune their weights, so the few-shot charges can contribute more on model optimization. Experimental results on several datasets demonstrate the effectiveness and robustness of our method. Besides, our method can work well on tiny datasets and has better efficiency in the training, which provides better applicability in real scenarios. © 2021, Springer Nature Switzerland AG.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

TREC: transient redundancy elimination-based convolution 22

TREC: transient redundancy elimination-based convolution

引用

Proceedings of the 36th International Conference on Neural Information Processing Systems

作者： Jiawei Guan Feng Zhang Jiesong Liu Hsin-Hsuan Sung Ruofan Wu Xiaoyong Du Xipeng Shen Key Laboratory of Data Engineering and Knowledge Engineering (MOE) and School of Information Renmin University of China Computer Science Department North Carolina State University

ISBN: (纸本)9781713871088

The intensive computations in convolutional neural networks (CNNs) pose challenges for resource-constrained devices; eliminating redundant computations from convolution is essential. This paper gives a principled method to detect and avoid transient redundancy, a type of redundancy existing in input data or activation maps and hence changing across inferences. By introducing a new form of convolution (TREC), this new method makes transient redundancy detection and avoidance an inherent part of the CNN architecture, and the determination of the best configurations for redundancy elimination part of CNN backward propagation. We provide a rigorous proof of the robustness and convergence of TREC-equipped CNNs. TREC removes over 96% computations and achieves 3.51× average speedups on microcontrollers with minimal (about 0.7%) accuracy loss.

关键词：

来源：评论

学校读者我要写书评

暂无评论

NSPG-Miner: Mining Repetitive Negative Sequential Patterns

arXiv

引用

arXiv 2025年

作者： Li, Yan Wang, Zhulin Liu, Jing Guo, Lei Fournier-Viger, Philippe Wu, Youxi Wu, Xindong School of Economics and Management Hebei University of Technology Tianjin300401 China School of Artificial Intelligence Hebei University of Technology Tianjin300401 China State Key Laboratory of Reliability and Intelligence of Electrical Equipment Hebei University of Technology Tianjin300401 China College of Computer Science and Software Engineering Shenzhen University Shenzhen518061 China Hebei Key Laboratory of Big Data Computing 300401 China Key Laboratory of Knowledge Engineering with Big Data The Ministry of Education of China Hefei University of Technology Hefei230009 China

Sequential pattern mining (SPM) with gap constraints (or repetitive SPM or tandem repeat discovery in bioinformatics) can find frequent repetitive subsequences satisfying gap constraints, which are called positive sequential patterns with gap constraints (PSPGs). However, classical SPM with gap constraints cannot find the frequent missing items in the PSPGs. To tackle this issue, this paper explores negative sequential patterns with gap constraints (NSPGs). We propose an efficient NSPG-Miner algorithm that can mine both frequent PSPGs and NSPGs simultaneously. To effectively reduce candidate patterns, we propose a pattern join strategy with negative patterns which can generate both positive and negative candidate patterns at the same time. To calculate the support (frequency of occurrence) of a pattern in each sequence, we explore a NegPair algorithm that employs a key-value pair array structure to deal with the gap constraints and the negative items simultaneously and can avoid redundant rescanning of the original sequence, thus improving the efficiency of the algorithm. To report the performance of NSPG-Miner, 11 competitive algorithms and 11 datasets are employed. The experimental results not only validate the effectiveness of the strategies adopted by NSPG-Miner, but also verify that NSPG-Miner can discover more valuable information than the state-ofthe-art algorithms. Algorithms and datasets can be downloaded from https://***/wuc567/PatternMining/tree/master/NSPG-Miner. Copyright © 2025, The Authors. All rights reserved.

关键词： Miners

来源：评论

学校读者我要写书评

暂无评论

Semantic SLAM Based on Compensated Segmentation and Geometric Constraints in Dynamic Environments

Semantic SLAM Based on Compensated Segmentation and Geometri...

引用

Artificial Intelligence Technology (ACAIT), Asian Conference on

作者： Baofu Fang Shuai Zhou Hao Wang School of Computer Science and Information Engineering Hefei University of Technology Key Laboratory of Knowledge Engineering with Big Data(Hefei University of Technology) Hefei China

Most of the existing slam algorithms are designed based on the assumption of a static environment, this strong assumption limits the practical application of most slam systems. The main reason is that moving objects will cause feature mismatch in the pose estimation process, which in turn affects the accuracy of localization and mapping. In this paper, we propose a SLAM algorithm in a dynamic environment. First, we use the BlendMask network to detect potential moving objects to generate masks for dynamic objects. The geometrically constrained joint optical flow method is used to detect dynamic feature points. Secondly, aiming at the failure of semantic segmentation network segmentation, a missed detection compensation algorithm based on the invariance of adjacent frame speed is proposed. Finally, a keyframe selection strategy is proposed to construct a semantic octree graph containing only static objects. We evaluate our algorithm on TUM RGB-D and real scene datasets. The experimental results show that the algorithm has high accuracy and real-time performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

VRFormer: 360-Degree Video Streaming with FoV Combined Prediction and Super resolution

VRFormer: 360-Degree Video Streaming with FoV Combined Predi...

引用

IEEE International Conference on Big data and Cloud Computing (BdCloud)

作者： Zhihao Zhang Haipeng Du Shouqin Huang Weizhan Zhang Qinghua Zheng Shaanxi Provincial Key Laboratory of Big Data Knowledge Engineering Xi'an Jiaotong University School of Continuing Education Xi'an Jiaotong University

360-degree video has shown great potential to the mainstream since its immersive experience. However, 360-degree video streaming requires ultrahigh bandwidth and low latency, which limit the improvement of user quality of experience (QoE). Currently, methods combining field of view (FoV) prediction and adaptive video streaming provide an effective method for addressing the above issues. However, existing FoV prediction methods based on recurrent neural networks (RNN) cannot capture long-range dependency from input to output. Current deep reinforcement learning (DRL)-based adaptive strategies fail to estimate the future bandwidth with high accuracy and fully explore the capability of VR devices. To ameliorate these limitations, we design a DRL-based 360-degree video streaming method named VRFormer with FoV combined prediction and super resolution (SR). First, we adopt a content-aware transformer-based encoder-decoder network to make the long-term FoV prediction. It combines the user's head movement history, eye-tracking history, and user attention extracted from a convolutional neural network (CNN)-based network. Second, we introduce a DNN-based SR network running on VR devices to reconstruct high-definition video content. Finally, we apply a DRL-based network to adaptively allocate rates for future tiles and dynamically control video content reconstruction. Experiments have verified that the proposed method can effectively improve the quality of experience (QoE) of the user's viewing experience compared to the state-of-the-art methods.

关键词： Recurrent neural networks Superresolution Bandwidth Reinforcement learning Prediction methods Streaming media Transformers

来源：评论

学校读者我要写书评

暂无评论

RGB-D SLAM Method Based on Feature Association in Dynamic Environment

RGB-D SLAM Method Based on Feature Association in Dynamic En...

引用

Artificial Intelligence Technology (ACAIT), Asian Conference on

作者： Baofu Fang Hao Wang School of Computer Science and Information Engineering Hefei University of Technology Key Laboratory of Knowledge Engineering with Big Data (Hefei University of Technology) Hefei China

Simultaneous localization and mapping (SLAM) is one of the current research hotspots. However, in visual SLAM for dynamic environments, inaccurate detection of object motion states and incomplete dynamic region culling will lead to large localization errors. To address these issues, this paper proposes an RGB-D SLAM method based on feature association. The method has strongly correlated features in time and space according to the input image sequence. Using the moving probability of the feature points in the previous frame, the movement of the feature points in the current frame is calculated in combination with the dynamic corner points screened in the current frame. Then, the motion state of the object is determined according to the proportion of different feature points. Then combined with semantic information and object depth information, the fast search method is used to obtain accurate dynamic regions. Finally, the selected effective feature points are used to estimate the camera pose and establish a static map of the environment. This paper evaluates the robustness and accuracy of our method on the TUM dataset and real environment, and the results show that our method can significantly improve the system tracking effect and reduce the system tracking error compared with other SLAM methods in dynamic environments.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Prototypical Replay with Old-class Focusing knowledge Distillation for Incremental Named Entity Recognition 39

Prototypical Replay with Old-class Focusing Knowledge Distil...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Liu, Zesheng Zhu, Qiannan Li, Cuiping Chen, Hong School of Information Renmin University of China Beijing China Key Laboratory of Data Engineering and Knowledge Engineering MOE China Engineering Research Center of Database and Business Intelligence MOE China School of Artificial Intelligence Beijing Normal University Beijing China Engineering Research Center of Intelligent Technology and Educational Application MOE China

ISBN: (纸本)157735897X

Catastrophic forgetting is a key challenge in incremental named entity recognition (INER). Existing methods often address this issue through distillation-based approaches, which involve transferring previously learned knowledge from the old model to the new one. However, these methods may not fully equip the new model with an adequate understanding of the characteristics about old entity types, leading to confusion when classifying tokens associated with these entity types. To address this challenge, we propose a novel method called Prototypical Replay with Old-class Focusing knowledge Distillation (POF) for INER. Our approach focuses on preserving the main characteristics of each previous entity type by storing compact prototypes and replaying them with appropriate frequency. This replay strategy makes the new model review the knowledge of old entity types while minimizing storage needs. Additionally, we introduce an old-class focusing knowledge distillation (OFKD) loss, which distills features only in old-class regions to maintain the quality of old-class prototypes and prevent ineffective prototypical replay while preserving sufficient plasticity for learning new entity types. We conduct experiments on three benchmark datasets (i.e., Few-NERD, I2B2 and OntoNotes5), and the results demonstrate that our method outperforms all previous state-of-the-art methods. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

KAN v.s. MLP for Offline Reinforcement Learning

KAN v.s. MLP for Offline Reinforcement Learning

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Haihong Guo Fengxin Li Jiao Li Hongyan Liu School of Information Renmin University of China China Institute of Medical Information / Medical Library Chinese Academy of Medical Sciences / Peking Union Medical College China Key Laboratory of Data Engineering and Knowledge Engineering Ministry of Education China School of Economics and Management Tsinghua University China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Kolmogorov-Arnold Networks (KAN) is an emerging neural network architecture in machine learning. It has greatly interested the research community about whether KAN can be a promising alternative to the commonly used Multi-Layer Perceptions (MLP). Experiments in various fields demonstrated that KAN-based machine learning can achieve comparable if not better performance than MLP-based methods, but with much smaller parameter scales and are more explainable. In this paper, we explore the incorporation of KAN into the actor and critic networks for offline reinforcement learning (RL). We evaluated the performance, parameter scales, and training efficiency of various KAN and MLP-based conservative Q-learning (CQL) on the classical D4RL benchmark for offline RL. Our study demonstrates that KAN can achieve performance close to the commonly used MLP with significantly fewer parameters. This allows us to choose the base networks according to the offline RL task requirements.

关键词： Training Performance evaluation Hands Q-learning Neural networks Memory management Signal processing Benchmark testing Splines (mathematics) Speech processing

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：