检索结果-内蒙古大学图书馆

2023 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2023

作者： Zhu, Wen Chen, Yiwen Nie, Shanling Yang, Hai East China University of Science and Technology Department of Computer Science and Engineering Shanghai China Shanghai Key Laboratory of Computer Software Evaluating and Testing Shanghai China National University of Singapore Center for Continuing and Lifelong Education Singapore The University of Sydney Faculty of Engineering Australia

ISBN: (纸本)9798350337488

Cancer survival prediction is pivotal in tailoring individualized treatment strategies and guiding clinician decision-making. Yet, existing methodologies grapple with efficiently harnessing the intricate distribution of medical data spanning various modalities. In response, we present SAMMS, an advanced multi-omics multimodal deep learning framework tailored for survival prediction. SAMMS leverages the robust image segmentation model, "Segment Anything"to adeptly characterize pathological images. This prowess is further enhanced by integrating multi-omics data and clinical insights, facilitating holistic modeling across a diverse modal spectrum. The framework weaves a modality-specific subnetwork with a cross-modality common subnetwork, meticulously capturing intra-modality nuances and inter-modality correlations. SAMMS eclipsed its contemporaries by delivering remarkable performance on TCGA's LGG and KIRC tumor datasets. A battery of analyses underscored SAMMS's unparalleled capability to distill multifaceted insights from multimodal datasets, yielding richer and more integrative multimodal representations. Such strides promise significant advancements in cancer survival analytics, bolstering the precision and efficacy of patient-centric treatments, disease oversight, and clinical decision processes. © 2023 IEEE.

关键词： Multi-omics Data Pathological Images Segment Anything Survival Analysis

来源：评论

学校读者我要写书评

暂无评论

Revisiting Benchmark and Assessment: An Agent-based Exploratory Dynamic Evaluation Framework for LLMs

arXiv

引用

arXiv 2024年

作者： Wang, Wanying Ma, Zeyu Liu, Pengfei Chen, Mingang Shanghai Key Laboratory of Computer Software Testing and Evaluating Shanghai Development Center of Computer Software Technology China Shanghai Normal University China Shanghai Jiao Tong University China

While various vertical domain large language models (LLMs) have been developed, automatically evaluating their performance across different domains remains a critical challenge. Current benchmark-based methods often rely on static and costly datasets, are misaligned with practical user needs, and lack flexibility across domains. To address these limitations, we revisit the evaluation process and introduce two key concepts: Benchmark+, which extends the traditional question-answer benchmark into a more flexible "strategy-criterion" format;and Assessment+, which enhances the interaction process, enabling deeper exploration and supporting analysis from broader perspectives. We propose TESTAGENT, an agent-based evaluation framework that implements these concepts using retrieval-augmented generation and reinforcement learning. TESTAGENT enables automatic dynamic benchmark generation and in-depth assessment across diverse vertical domain scenarios. Experiments on tasks ranging from constructing multiple vertical domain evaluations to converting static benchmarks into dynamic forms demonstrate the effectiveness of TESTAGENT. This work offers an interesting perspective on automatic evaluation for LLMs and highlights a pathway for dynamic and domain-adaptive assessments. Copyright © 2024, The Authors. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Data Generation, testing and Evaluation of Chinese Natural Language Processing in the Cloud 7

Data Generation, Testing and Evaluation of Chinese Natural L...

引用

7th IEEE International Conference on Smart Cloud, SmartCloud 2022

作者： Ding, Minjie Chen, Mingang Chen, Wenjie Cai, Lizhi Chai, Yuanhao Shanghai Key Laboratory of Computer Software Testing and Evaluating Shanghai Development Center Computer Software Technology Shanghai China Washington University in St. Louis McKelvey School of Engineering St. LouisMO United States

ISBN: (数字)9781665451796

ISBN: (纸本)9781665451796

With the rapid development of artificial intelligence, natural language processing, as an important branch, has also become a hot research field. A series of super large-scale pre-trained models represented by BERT and GPT have made great progress in natural language understanding and natural language generation, even some of the experimental accuracy exceed the human benchmark. However, these models will also make some mistakes and even fairness problems when they have the language ability equivalent to human beings. In order to verify whether the models can truly understand natural language, the evaluation of these models is particularly important. More methods are needed to evaluate the model. The language model-based evaluation tools often require a lot of computing resources. In this paper, we propose a method for testing and evaluation of Chinese natural language processing in cloud, generate testing data and design tests for Chinese data and test two pre-trained models. The experimental results show that our method can find defects of the model, though it has high performance on specific dataset. © 2022 IEEE.

关键词： Artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

Mprldroid: A Reinforcement Learning Malware Detection Model Based on Heterogeneous Information Network Path Representation

SSRN

引用

SSRN 2023年

作者： Yang, Kang Cai, Lizhi Wu, Jianhua Liu, Zhenyu Zhang, Meng Shanghai Key Laboratory of Computer Software Testing Evaluating Shanghai201112 China Shanghai Development Center of Computer Software Technology Shanghai201112 China

With the significant increase of Android malware, the application privacy data leakage incidents occur frequently, which poses a great threat to user property and information security. However, the new malware has the characteristics of high evolution rate and diverse variants, leading to the fact that the current malware detection methods still have three key problems: (1) Difficulty in acquiring Android sample structural features;(2) Weakly in representing malware behavior structure;(3) Poor robustness of the detection model. To address the above limitations, we proposes a new malware detection framework MPRLDroid with reinforcement learning. First of all, the MPRLDroid model extracts the Android application structural features and constructs the heterogeneous information network data based on the semantic call structure between APP, API and permission. Subsequently, the model uses reinforcement learning algorithm to adaptively generate meta-paths for each sample and incorporates graph attention network to effectivelyrepresentation the graph of node. Finally, the low-dimensional graph node vector data brought into the downstream detection task for classification, where the performance change of the classification result is used as a reward function for reinforcement learning. The experimental results show that the performance of the MPRLDRroid model combined with reinforcement learning outperforms the baseline models, and the robustness of the detection model is also better than other models. © 2023, The Authors. All rights reserved.

关键词： Android malware

来源：评论

学校读者我要写书评

暂无评论

Contrastive Pseudo Learning for Open-World DeepFake Attribution

Contrastive Pseudo Learning for Open-World DeepFake Attribut...

引用

International Conference on computer Vision (ICCV)

作者： Zhimin Sun Shen Chen Taiping Yao Bangjie Yin Ran Yi Shouhong Ding Lizhuang Ma Shanghai Jiao Tong University Tencent YouTu Lab Shanghai Key Laboratory of Computer Software Testing & Evaluating

The challenge in sourcing attribution for forgery faces has gained widespread attention due to the rapid development of generative techniques. While many recent works have taken essential steps on GAN-generated faces, more threatening attacks related to identity swapping or expression transferring are still overlooked. And the forgery traces hidden in unknown attacks from the open-world unlabeled faces still remain under-explored. To push the related frontier research, we introduce a new benchmark called Open-World DeepFake Attribution (OW-DFA), which aims to evaluate attribution performance against various types of fake faces under open-world scenarios. Meanwhile, we propose a novel framework named Contrastive Pseudo Learning (CPL) for the OW-DFA task through 1) introducing a Global-Local Voting module to guide the feature alignment of forged faces with different manipulated regions, 2) designing a Confidence-based Soft Pseudo-label strategy to mitigate the pseudo-noise caused by similar methods in unlabeled set. In addition, we extend the CPL framework with a multi-stage paradigm that leverages pre-train technique and iterative learning to further enhance traceability performance. Extensive experiments verify the superiority of our proposed method on the OW-DFA and also demonstrate the interpretability of deepfake attribution task and its impact on improving the security of deepfake detection area.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Integrating Extractive and Abstractive Models for Code Comment Generation

Integrating Extractive and Abstractive Models for Code Comme...

引用

IEEE International Conference on software Quality, Reliability and Security (QRS)

作者： Weisong Sun Yuling Hu Yingfei Xu Yuchen Chen Chunrong Fang State Key Laboratory for Novel Software Technology Nanjing University Nanjing China Shanghai Key Laboratory of Computer Software Evaluating and Testing Shanghai China

Code comments play an essential role in aiding developers understand and maintain source code. Current code comment generation techniques can be classified into categories: extractive methods and abstractive methods. Extractive methods use text retrieval techniques to extract important code tokens to constitute comments. Such comments contain important factual details articulated explicitly in code tokens, but are poor in naturalness. Abstractive methods usually regard code comment generation as a neural machine translation task. By leveraging powerful deep learning-based language models, abstractive methods can generate comments that resemble human writing. However, compared with natural language, programming language code is more complex. Comments generated by abstractive methods often leave out important factual details. In this paper, we propose a novel method for code comment generation by integrating extractive and abstractive models. Our extractive model is built on the Latent Semantic Analysis (LSA) model, effectively extracting important factual details in code snippets. Meanwhile, our abstractive model is built on a deep learning-based encoder-decoder model, enabling it to generate concise and human-written-like comments. We evaluate the effectiveness of our method, called ICS, by conducting extensive experiments on the CodeSearchNet dataset involving six programming languages. The results demonstrate that ICS outperforms state-of-the-art techniques in three widely used metrics: BLEU, METEOR, and ROUGE-L. Moreover, the outcomes of the human evaluation indicate that the comments generated by ICS exhibit superior naturalness and informativeness, and closely align with the provided code snippets.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Anonym-Recognizer: Relationship-preserving Face Anonymization and Recognition 3

Anonym-Recognizer: Relationship-preserving Face Anonymizatio...

引用

3rd International Workshop on Human-Centric Multimedia Analysis, HCMA 2022, held in conjunction with the ACM Multimedia 2022

作者： Peng, Chunlei Wan, Shuang Miao, Zimin Liu, Decheng Zheng, Yu Wang, Nannan School of Cyber Engineering Xidian University Xi'an China Shanghai Key Laboratory of Computer Software Evaluating and Testing Shanghai China State Key Laboratory of Integrated Services Networks Xidian University Xi'an China

ISBN: (纸本)9781450394925

With the widespread application of big data technology, we are exposed to more and more video monitoring. To prevent serious social problems caused by face data leakage, face anonymization has become an important kind of method to protect face privacy. The face anonymization mentioned in this paper refers to the anonymization generation of the visual appearance in face images. Existing face anonymization methods mainly focus on removing identity information. However, in the scenario of face recognition technology that needs to protect privacy, existing face anonymization technology makes anonymized faces that can no longer be used for face recognition, limiting the application scope of face anonymization. Therefore, when using face anonymization, it is equally important to ensure that the anonymized face images can still be used for downstream tasks such as face recognition. To this end, we propose Anonym-Recognizer, a relationship-preserving face anonymization and recognition method. Our method uses relationship cyphertext which can be any binary identity number representing the identity of the image owner and designs a generative adversarial network to perform face anonymization and relationship cyphertexts embedding. In our framework, we first use Visual Anonymizer to manipulate the visual appearance of the input image, then use Cyphertext Embedder to get the anonymized image with the identity information embedded. With the help of Anonym Recognizer, the face recognition system can extract the relationship cyphertexts from the anonymized image as the credentials to match the identity information. The proposed Anonym-Recognizer provides a new perspective for the recognition and application of anonymized face images. Experiments on the Megaface dataset show that our method can encourage a 100% recognition accuracy on anonymized faces while finishing the task of face anonymization with high qualitative and quantitative quality. © 2022 ACM.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

RGA-Reid: Neural Architecture Search Based on Relation-Aware Global Attention for Person Re-Identification

RGA-Reid: Neural Architecture Search Based on Relation-Aware...

引用

IEEE Joint International Information Technology and Artificial Intelligence Conference (ITAIC)

作者： Shijie Chen Shengbo Chen Zhou Lei School of Computer Engineering and Science Shanghai University Shanghai China Shanghai Key Laboratory of Computer Software Testing and Evaluating Shanghai China

With the continuous development of deep learning, reid has been continually improved. However, the network backbones of reid are mostly manually designed. This paper proposes RGA-Reid based on auto-reid. NASBR-Reid obtains the network backbone through neural architecture search. It replaces the part-base module in the auto-reid search algorithm with the RGA module, which effectively integrates global relationships to enhance attention training between different parts. Experimental results demonstrate that RGA-Reid can retrieve superior network backbones.

关键词：

来源：评论

学校读者我要写书评

暂无评论

FAS-reid: Fair Architecture Search for Person Re-IDentification

FAS-reid: Fair Architecture Search for Person Re-IDentificat...

引用

Information Technology,Big Data and Artificial Intelligence (ICIBA), International Conference on IEEE International Conference on

作者： Shengbo Chen Shijie Chen Zhou Lei School of Computer Engineering and Science Shanghai University Shanghai China Shanghai Key Laboratory of Computer Software Testing and Evaluating Shanghai China

The development of deep learning has driven the development of ReID, and more and more excellent methods have been proposed, but most of these are artificially designed network backbones. Automation is a trend in the development of deep learning. This paper propose FAS-ReID(Fair Architecture Search for Person Re-IDentification), which can automatically design and generate a neural network backbone for ReID. This paper compare operation selection to competition, and inject noise to make the competition fairer, while making the architecture of automated search More adaptable to ReID We use TriHard loss to improve the feature extraction ability. Our experiments show that the backbone searched by FAS-ReID is better and reduces the search time. Meanwhile, as for the problem of performance collapse caused by skip-connection enrichment in the search process, FAS-Reid does not use the early stop strategy to avoid performance loss like other solutions, which is also the reason why our method is more reliable and robust.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Evaluation of Hallucination and Robustness for Large Language Models

Evaluation of Hallucination and Robustness for Large Languag...

引用

IEEE International Conference on software Quality, Reliability and Security Companion (QRS-C)

作者： Rui Hu Junhao Zhong Minjie Ding Zeyu Ma Mingang Chen Shanghai Key Laboratory of Computer Software Testing and Evaluating Shanghai Development Center of Computer Software Technology Shanghai China Shanghai AI Industry Association Shanghai China

As large language models (LLMs) rapidly advance, rigorous testing and evaluation of these models grows increasingly crucial. To address this need, we have developed three types of questions: Chinese contextual, English contextual, and language context-independent. testing in both Chinese and English probes the LLMs' hallucination tendencies. We investigate the impact of language on hallucinations from two perspectives: the type of language used in the input prompt and the cultural context underlying the prompt's content. Additionally, 52 multi-domain single-choice questions from C-EVAL are presented in original and randomized order to assess robustness to perturbations. Among the five LLMs, the tests demonstrate GPT -4 has the strongest anti-hallucination and robustness capabilities, answering with greater accuracy, consistency, and reliability. ChatGLM ranks second and outperforms GPT -4 on Chinese context-dependent questions. Emergent testing phenomena are analyzed from the user's perspective. Hallucinated responses are categorized and potential causal factors leading to hallucination and fragility are examined. Based on these findings, viable avenues for improvement are proposed.

关键词：

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：