检索结果-内蒙古大学图书馆

Semantic-Preserving Adversarial Code Comprehension

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Li, Yiyang Wu, Hongqiu Zhao, Hai Department of Computer Science and Engineering Shanghai Jiao Tong University China Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai Jiao Tong University China

Based on the tremendous success of pre-trained language models (PrLMs) for source code comprehension tasks, current literature studies either ways to further improve the performance (generalization) of PrLMs, or their robustness against adversarial attacks. However, they have to compromise on the tradeoff between the two aspects and none of them consider improving both sides in an effective and practical way. To fill this gap, we propose Semantic-Preserving Adversarial Code Embeddings (SPACE) to find the worst-case semantic-preserving attacks while forcing the model to predict the correct labels under these worst cases. Experiments and analysis demonstrate that SPACE can stay robust against state-of-the-art attacks while boosting the performance of PrLMs for code1 © 2022, CC BY.

关键词： Semantics

Sentence Representation Learning with Generative Objective rather than Contrastive Objective

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wu, Bohong Zhao, Hai Department of Computer Science and Engineering Shanghai Jiao Tong University China Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai Jiao Tong University China

Though offering amazing contextualized token-level representations, current pre-trained language models take less attention on accurately acquiring sentence-level representation during their self-supervised pre-training. However, contrastive objectives which dominate the current sentence representation learning bring little linguistic interpretability and no performance guarantee on downstream semantic tasks. We instead propose a novel generative self-supervised learning objective based on phrase reconstruction. To overcome the drawbacks of previous generative methods, we carefully model intra-sentence structure by breaking down one sentence into pieces of important phrases. Empirical studies show that our generative learning achieves powerful enough performance improvement and outperforms the current state-of-the-art contrastive methods not only on the STS benchmarks, but also on downstream semantic retrieval and reranking tasks. Our code is available at https://***/chengzhipanpan/PaSeR. Copyright © 2022, The Authors. All rights reserved.

关键词： Benchmarking

Learning Better Masking for Better Language Model Pre-training

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Yang, Dongjie Zhang, Zhuosheng Zhao, Hai Department of Computer Science and Engineering Shanghai Jiao Tong University China Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai Jiao Tong University China

Masked Language Modeling (MLM) has been widely used as the denoising objective in pretraining language models (PrLMs). Existing PrLMs commonly adopt a Random-Token Masking strategy where a fixed masking ratio is applied and different contents are masked by an equal probability throughout the entire training. However, the model may receive a complicated impact from pre-training status, which changes accordingly as training time goes on. In this paper, we show that such time-invariant MLM settings on masking ratio and masked content are unlikely to deliver an optimal outcome, which motivates us to explore the influence of time-variant MLM settings. We propose two scheduled masking approaches that adaptively tune the masking ratio and masked content in different training stages, which improves the pre-training efficiency and effectiveness verified on the downstream tasks. Our work is a pioneer study on time-variant masking strategy on ratio and content and gives a better understanding of how masking ratio and masked content influence the MLM pretraining1 Copyright © 2022, The Authors. All rights reserved.

关键词： Modeling languages

Multidisciplinary integration of frontier technologies facilitating the development of anesthesiology and perioperative medicine in aging society

学校读者我要写书评

暂无评论

Fundamental Research 2024年第4期4卷 795-796页

作者： Lize Xiong Junli Cao Hailong Dong Weihong Song Dong Ming Translational Research Institute of Brain and Brain-Like Intelligence and Department of Anesthesiology and Perioperative Medicine Shanghai Fourth People's Hospital Affiliated to Tongji University School of MedicineShanghai 200434China Jiangsu Province Key Laboratory of Anesthesiology Xuzhou Medical UniversityXuzhou 221004China Department of Anesthesiology and Perioperative Medicine Xijing Hospitalthe Fourth Military Medical UniversityXi'an 710032China The Second Affliated Hospital and Yuying Children's Hospital Institute of AgingKey Laboratory of Alzheimer's Disease of Zhejiang ProvinceWenzhou Medical UniversityWenzhou325035China State Key Laboratory of Advanced Medical Materials and Devices Medical SchoolTianjin UniversityTianjin 300072China Haihe Laboratory of Brain-Computer Interaction and Human-Machine Integration Tianjin 300380China

The population aged 60 and over in China reached approximately 297 million in 2023 according to the National Bureau of Statistics,accounting for 21.1%of the national *** population aged 65 and over was about 217 million,accounting for 15.4%[1].It is estimated that by 2035,China will have an elderly population of over 400 million,accounting for more than 30%of the total population,thereby entering a stage of severe *** brings many challenges to the society,especially in terms of medical care for elderly *** continuous advancement of science and technology,especially the breakthrough in basic research as well as new technologies and interdisciplinary integration,has brought new opportunities for the development of various professional *** better address the challenges of anesthesia and perioperative medicine in the aging era and to accelerate the rehabilitation of patients while maintaining life and health services,it is crucial to promote interdisciplinary integration.

关键词： breakthrough frontier thereby

Energy Efficiency Optimization for Full-Duplex D2D Communications Underlaying Distributed Antenna Systems

学校读者我要写书评

暂无评论

IEEE Transactions on Green Communications and Networking

IEEE Transactions on Green Communications and Networking 2024年

作者： Liu, Zhan Liao, Zhiyuan Li, Chunquan Zhang, Zhijun Yu, Junzhi Liu, P.X. Hunan University of Humanities Science and Technology School of Information Loudi417000 China Nanchang University School of Information Engineering Nanchang330031 China Jiangxi Provincial Key Laboratory of Intelligent Systems and Human-Machine Interaction Nanchang330031 China South China University of Technology School of Automation Science and Engineering Guangzhou510640 China Peking University State Key Laboratory for Turbulence and Complex Systems Department of Advanced Manufacturing and Robotics College of Engineering Beijing100871 China Carleton University Department of Systems and Computer Engineering OttawaONK1S 5B6 Canada

In this paper, we investigate the total system energy efficiency (EE) of full-duplex (FD) device-to-device (D2D) communications underlaying distributed antenna systems (DAS), where remote access units (RAUs), D2D users (DUs), and cellular users (CUs) are all capable of FD operation. Specifically, we jointly optimize subcarrier assignment and power allocation under the quality of service (QoS) requirements of CUs and DUs and the maximum power constraints of RAUs, CUs, and DUs. In addition, we propose a novel spectrum sharing strategy that allows each subcarrier to be assigned to multiple CUs and/or multiple D2D pairs (DPs) for flexibility. To solve the formulated non-convex optimization problem, we first employ fractional programming to transform the objective function in the optimization problem from the fractional form into the equivalent subtractive form. Then, an efficient iterative resource allocation algorithm is proposed, which needs to solve an inner problem in each iteration. After relaxing the variables and introducing penalty factors, the non-convex inner problem is transformed into a convex problem through the successive convex approximation (SCA) method and solved by iterative algorithm. Simulation results demonstrate that the proposed algorithm can considerably improve the system EE compared to other benchmark schemes. Furthermore, the proposed spectrum sharing strategy is superior to the existing sharing strategies. © 2017 IEEE.

关键词： Convex optimization

Exploiting aliasing for manga restoration

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Xie, Minshan Xia, Menghan Wong, Tien-Tsin Department of Computer Science and Engineering Chinese University of Hong Kong Hong Kong Shenzhen Key Laboratory of Virtual Reality and Human Interaction Technology SIAT CAS

As a popular entertainment art form, manga enriches the line drawings details with bitonal screentones. However, manga resources over the Internet usually show screentone artifacts because of inappropriate scanning/rescaling resolution. In this paper, we propose an innovative twostage method to restore quality bitonal manga from degraded ones. Our key observation is that the aliasing induced by downsampling bitonal screentones can be utilized as informative clues to infer the original resolution and screentones. First, we predict the target resolution from the degraded manga via the Scale Estimation Network (SENet) with spatial voting scheme. Then, at the target resolution, we restore the region-wise bitonal screentones via the Manga Restoration Network (MR-Net) discriminatively, depending on the degradation degree. Specifically, the original screentones are directly restored in pattern-identifiable regions, and visually plausible screentones are synthesized in pattern-agnostic regions. Quantitative evaluation on synthetic data and visual assessment on real-world cases illustrate the effectiveness of our method. Copyright © 2021, The Authors. All rights reserved.

关键词： Restoration

Adversarial Self-Attention for Language Understanding

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wu, Hongqiu Ding, Ruixue Zhao, Hai Xie, Pengjun Huang, Fei Zhang, Min Department of Computer Science and Engineering Shanghai Jiao Tong University China Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai Jiao Tong University China School of Computer Science and Technology Soochow University China Damo Academy Alibaba Group China

Deep neural models (e.g. Transformer) naturally learn spurious features, which create a "shortcut" between the labels and inputs, thus impairing the generalization and robustness. This paper advances the self-attention mechanism to its robust variant for Transformer-based pre-trained language models (e.g. BERT). We propose Adversarial Self-Attention mechanism (ASA), which adversarially biases the attentions to effectively suppress the model reliance on features (e.g. specific keywords) and encourage its exploration of broader semantics. We conduct a comprehensive evaluation across a wide range of tasks for both pre-training and fine-tuning stages. For pre-training, ASA unfolds remarkable performance gains compared to naive training for longer steps. For fine-tuning, ASA-empowered models outweigh naive models by a large margin considering both generalization and robustness. Copyright © 2022, The Authors. All rights reserved.

关键词： Semantics

Forging Multiple Training Objectives for Pre-trained Language Models via Meta-Learning

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wu, Hongqiu Ding, Ruixue Zhao, Hai Chen, Boli Xie, Pengjun Huang, Fei Zhang, Min Department of Computer Science and Engineering Shanghai Jiao Tong University China Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai Jiao Tong University China School of Computer Science and Technology Soochow University China Damo Academy Alibaba Group China

Multiple pre-training objectives fill the vacancy of the understanding capability of single-objective language modeling, which serves the ultimate purpose of pre-trained language models (PrLMs), generalizing well on a mass of scenarios. However, learning multiple training objectives in a single model is challenging due to the unknown relative significance as well as the potential contrariety between them. Empirical studies have shown that the current objective sampling in an ad-hoc manual setting makes the learned language representation barely converge to the desired optimum. Thus, we propose MOMETAS, a novel adaptive sampler based on meta-learning, which learns the latent sampling pattern on arbitrary pre-training objectives. Such a design is lightweight with negligible additional training overhead. To validate our approach, we adopt five objectives and conduct continual pre-training with BERT-base and BERT-large models, where MOMETAS demonstrates universal performance gain over other rule-based sampling strategies on 14 natural language processing tasks. Copyright © 2022, The Authors. All rights reserved.

关键词： Modeling languages

CoCo-Agent: A Comprehensive Cognitive MLLM Agent for Smartphone GUI Automation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Ma, Xinbei Zhang, Zhuosheng Zhao, Hai School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University China Department of Computer Science and Engineering Shanghai Jiao Tong University China Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai Jiao Tong University China Shanghai Key Laboratory of Trusted Data Circulation and Governance in Web3 China

Multimodal large language models (MLLMs) have shown remarkable potential as human-like autonomous language agents to interact with real-world environments, especially for graphical user interface (GUI) automation. However, those GUI agents require comprehensive cognition ability including exhaustive perception and reliable action response. We propose a Comprehensive Cognitive LLM Agent, CoCo-Agent, with two novel approaches, comprehensive environment perception (CEP) and conditional action prediction (CAP), to systematically improve the GUI automation performance. First, CEP facilitates the GUI perception through different aspects and granularity, including screenshots and complementary detailed layouts for the visual channel and historical actions for the textual channel. Second, CAP decomposes the action prediction into sub-problems: action type prediction and action target conditioned on the action type. With our technical design, our agent achieves new state-of-the-art performance on AITW and META-GUI benchmarks, showing promising abilities in realistic scenarios. Code is available at https://***/xbmxb/CoCo-Agent. Copyright © 2024, The Authors. All rights reserved.

关键词： Graphical user interfaces