检索结果-内蒙古大学图书馆

Using GPT-4 to guide causal machine learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Constantinou, Anthony C. Kitson, Neville K. Zanga, Alessio research group School of Electronic Engineering and Computer Science Queen Mary University of London LondonE1 4NS United Kingdom Department of Informatics Systems and Communication University of Milano - Bicocca Milan Italy Data Science and Advanced Analytics F. Hoffmann - La Roche Ltd Basel Switzerland

Since its introduction to the public, ChatGPT has had an unprecedented impact. While some experts praised AI advancements and highlighted their potential risks, others have been critical about the accuracy and usefulness of Large Language Models (LLMs). In this paper, we are interested in the ability of LLMs to identify causal relationships. We focus on the well-established GPT-4 (Turbo) and evaluate its performance under the most restrictive conditions, by isolating its ability to infer causal relationships based solely on the variable labels without being given any other context by humans, demonstrating the minimum level of effectiveness one can expect when it is provided with label-only information. We show that questionnaire participants judge the GPT-4 graphs as the most accurate in the evaluated categories, closely followed by knowledge graphs constructed by domain experts, with causal Machine Learning (ML) far behind. We use these results to highlight the important limitation of causal ML, which often produces causal graphs that violate common sense, affecting trust in them. However, we show that pairing GPT-4 with causal ML overcomes this limitation, resulting in graphical structures learnt from real data that align more closely with those identified by domain experts, compared to structures learnt by causal ML alone. Overall, our findings suggest that despite GPT-4 not being explicitly designed to reason causally, it can still be a valuable tool for causal representation, as it improves the causal discovery process of causal ML algorithms that are designed to do just that. Copyright © 2024, The Authors. All rights reserved.

关键词： Knowledge graph

MAXIMAL NONCOMPACTNESS OF EMBEDDINGS INTO MARCINKIEWICZ SPACES

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Malý, Jan Mihula, Zdeněk Musil, Vít Pick, Luboš Department of Mathematical Analysis Faculty of Mathematics and Physics Charles University Sokolovská 83 Praha 8186 75 Czech Republic Department of Mathematics Faculty of Electrical Engineering Czech Technical University in Prague Technická 2 Praha 6166 27 Czech Republic Masaryk University Faculty of Informatics Department of Computer Science Botanická 68a Brno602 00 Czech Republic

We develop a new functional-analytic technique for investigating the degree of noncompactness of an operator defined on a quasinormed space and taking values in a Marcinkiewicz space. The main result is a general principle from which it can be derived that such operators are almost always maximally noncompact in the sense that their ball measure of noncompactness coincides with their operator norm. We point out specifications of the universal principle to the case of the identity *** Codes 41A46, 46E35, 46E30 Copyright © 2024, The Authors. All rights reserved.

关键词：

Integrated ARM ***-Mali pipeline for high-throughput CNN inference

学校读者我要写书评

暂无评论

TechRxiv

TechRxiv 2021年

作者： Aghapour, Ehsan Pathania, Anuj Ananthanarayanan, Gayathri The Informatics Departments University of Amsterdam Netherlands The Department of Computer Science and Engineering Indian Institute of Technology Karnataka Dharwad India

Heterogeneous Multi-Processor System on Chips (HMPSoCs) combine several types of processors on a single chip. State-of-the-art embedded devices are becoming ever more powerful thanks to advancements in the computation power of HMPSoCs that enable them. Consequently, these devices increasingly perform the task of Machine Learning (ML) inferencing at the edge. State-of-the-art HMPSoCs can perform on-chip embedded inference on its CPU and GPU. Multi-component pipelining is the method of choice to provide high-throughput Convolutions Neural Network (CNN) inference on embedded platforms. In this work, we provide details for the first CPU-GPU pipeline design for CNN inference called Pipe-All. Pipe-All uses the ARM-CL library to integrate an ARM *** CPU with an ARM Mali GPU. Pipe-All is the first three-stage CNN inference pipeline design with ARM’s big CPU cluster, Little CPU cluster, and Mali GPU as its stages. Pipe-All provides on average 75.88% improvement in inference throughput (over peak single-component inference) on Amlogic A311D HMPSoC in Khadas Vim 3 embedded platform. We also provide an open-source implementation for Pipe-All under MIT License. © 2021, CC BY.

关键词： System-on-chip

Fuzzy Attack Tree: Assessing Cyberattack Risk Using Attack Tree and Fuzzy Logic

学校读者我要写书评

暂无评论

Fuzzy Attack Tree: Assessing Cyberattack Risk Using Attack T...

International Conference on ICT in Business Industry & Government (ICTBIG)

作者： Nitin Naik Paul Jenkins Paul Grace Dishita Naik Jingping Song Shaligram Prajapat Durgesh Mishra Longzhi Yang Tossapon Boongoen Natthakan Iam-On School of Informatics and Digital Engineering Aston University United Kingdom Cardiff School of Technologies Cardiff Metropolitan University United Kingdom Birmingham City University United Kingdom Software College Northeastern University China International Institute of Professional Studies Devi Ahilya University India School of Computer Science & Information Technology Symbiosis University of Applied Sciences India Department of Computer and Information Sciences Northumbria University United Kingdom Department of Computer Science Aberystwyth University United Kingdom

Organisations and users have been experiencing significant rises in cyberattacks and their severity, which means that they require a greater awareness and understanding of the anatomy of cyberattacks, to prevent and mitigate their effects. In analysing cyberattacks, there are a number of different approaches that may be used to assess their potential risks and effects. However, these are utilised in specific types of cyberattacks and their analysis, which means they cannot be applied in every situation or cyberattack. Moreover, several other factors may influence the decision to use these approaches, such as cost, complexity, skills and adaptability. As a result, continuous research to design and enhance these approaches is undertaken to produce a generic, cost-effective, easy and adaptable approach. This paper presents one such approach to assess the risk of cyberattacks utilising an attack tree and fuzzy logic. An attack tree is a systematic and illustrative method for describing an attack on a system and analysing its taxonomy and other aspects. Subsequently, the probability and risk of each leaf node in the attack tree are calculated using the proposed formulas. Finally, fuzzy logic enables decision making based on imprecise data and heuristics to obtain the overall risk of attack. This proposed approach comprises systematic steps to accomplish an assessment of any cyberattack and its associated risks in an uncomplicated and effective manner, enabling its prevention and mitigation to be determined. The paper illustrates an application of the proposed approach to assess the risk of an information theft attack on an organisation, which can then be utilised to assess the risk of other cyberattacks.

关键词：

Reshaping free-text radiology notes into structured reports with generative transformers

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Bergomi, Laura Buonocore, Tommaso M. Antonazzo, Paolo Alberghi, Lorenzo Bellazzi, Riccardo Preda, Lorenzo Bortolotto, Chandra Parimbelli, Enea Department of Electrical Computer and Biomedical Engineering University of Pavia Pavia Italy Radiology Unit - Diagnostic Imaging I Department of Diagnostic Medicine Fondazione IRCCS Policlinico San Matteo Pavia Italy Diagnostic Imaging Unit Department of Clinical Surgical Diagnostic and Pediatric Sciences University of Pavia Pavia Italy LIM-IA - Laboratory of Medical Informatics and AI IRCCS Istituti Clinici Scientifici Maugeri Pavia Italy

BACKGROUND: Radiology reports are typically written in a free-text format, making clinical information difficult to extract and use. Recently the adoption of structured reporting (SR) has been recommended by various medical societies thanks to the advantages it offers, e.g. standardization, completeness and information retrieval. We propose a pipeline to extract information from free-text radiology reports, that fits with the items of the reference SR registry proposed by a national society of interventional and medical radiology, focusing on CT staging of patients with lymphoma. METHODS: Our work aims to leverage the potential of Natural Language Processing (NLP) and Transformer-based models to deal with automatic SR registry filling. With the availability of 174 radiology reports, we investigate a rule-free generative Question Answering approach based on a domain-specific version of T5 (IT5). Two strategies (batch-truncation and ex-post combination) are implemented to comply with the model’s context length limitations. Performance is evaluated in terms of strict accuracy, f1, and format accuracy, and compared with the widely used GPT-3.5 Large Language Model. A 5-point Likert scale questionnaire is used to collect human-expert feedback on the similarity between medical annotations and generated answers. RESULTS: The combination of fine-tuning and batch splitting allows IT5 to achieve notable results;it performs on par with GPT-3.5 albeit its size being a thousand times smaller in terms of parameters. Human-based assessment scores show a high correlation (Spearman’s correlation coefficients>0.88, p-values © 2024, CC BY-NC-ND.

关键词： Natural language processing systems

Disentangling The Prosody And Semantic Information With Pre-Trained Model For In-Context Learning Based Zero-Shot Voice Conversion

学校读者我要写书评

暂无评论

Disentangling The Prosody And Semantic Information With Pre-...

IEEE Spoken Language Technology Workshop

作者： Zhengyang Chen Shuai Wang Mingyang Zhang Xuechen Liu Junichi Yamagishi Yanmin Qian Department of Computer Science and Engineering Auditory Cognition and Computational Acoustics Lab MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University Shanghai China Shenzhen Research Institute of Big Data Shenzhen China Chinese University of Hong Kong Shenzhen China National Institute of Informatics Tokyo Japan

ISBN: (数字)9798350392258

ISBN: (纸本)9798350392265

Voice conversion (VC) aims to modify the speaker’s timbre while retaining speech content. Previous approaches have tokenized the outputs from self-supervised into semantic tokens, facilitating disentanglement of speech content information. Recently, in-context learning (ICL) has emerged in text-to-speech (TTS) systems for effectively modeling specific characteristics such as timbre through context conditioning. This paper proposes an ICL capability enhanced VC system (ICL-VC) employing a mask and reconstruction training strategy based on flow-matching generative models. Augmented with semantic tokens, our experiments on the LibriTTS dataset demonstrate that ICL-VC improves speaker similarity. Additionally, we find that k-means is a versatile tokenization method applicable to various pre-trained models. However, the ICL-VC system faces challenges in preserving the prosody of the source speech. To mitigate this issue, we propose incorporating prosody embeddings extracted from a pre-trained emotion recognition model into our system. Integration of prosody embeddings notably enhances the system’s capability to preserve source speech prosody, as validated on the Emotional Speech Database.

关键词： Training Emotion recognition Databases Semantics Self-supervised learning Speech enhancement Tokenization Text to speech Multilingual Timbre

Disturbance-agnostic robust performance with structured uncertainties and initial state error in classical versus quantum oscillatory systems

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Jonckheere, Edmond Schirmer, Sophie G. Langbein, Frank C. Weidner, Carrie A. O'Neil, Sean The Department of Electrical and Computer Engineering University of Southern California Los AngelesCA90089 United States The Faculty of Science & Engineering Swansea University SwanseaSA2 8PP United Kingdom The School of Computer Science and Informatics Cardiff University CardiffCF24 3AA United Kingdom The Quantum Engineering Technology Laboratories University of Bristol BristolBS8 1FD United Kingdom

A method to quantify robust performance for situations where structured parameter variations and initial state errors rather than extraneous disturbances are the main performance limiting factors is presented. The approach is based on the error dynamics, the difference between nominal and perturbed dynamics, driven by either the unperturbed or perturbed state, rather than an artificially imposed disturbance. The unperturbed versus perturbed dichotomy can be interpreted as the relative error dynamics scaled by either the unperturbed or perturbed dynamics. The error dynamics driven by unperturbed state has the unique feature of decoupling the effect of physically meaningful uncertainties from an additive disturbance. The perturbed case offers the possibility to side-step Structured Singular Value (SSV) computations. Applications to a lightly damped mechanical system and a slowly dephasing quantum system demonstrate the usefulness of the concepts across a broad range of systems. Finally, a fixed-point algorithm specifically developed for quantum systems with state transitions depending in a nonlinear fashion on uncertainties is proposed as the substitute for classical SSV. Copyright © 2023, The Authors. All rights reserved.

关键词： Robust control

Assessing the Possibility of a Country's Economic Growth Using Dynamic Bayesian Network Models 15

学校读者我要写书评

暂无评论

Assessing the Possibility of a Country's Economic Growth Usi...

15th IEEE International Scientific and Technical Conference on computer Sciences and Information Technologies, CSIT 2020

作者： Lytvynenko, Volodymyr Nikytenko, Dmytro Voronenko, Mariia Savina, Nataliia Naumov, Oleksandr Kherson National Technical University Department of Informatics and Computer Science Kherson Ukraine National University of Water and Environmental Engineering Department of Economic Theories Rivne Ukraine

ISBN: (纸本)9781728174433

This paper presents the results of studies on the development of probability-determined models based on Bayesian networks to assess the degree of economic development of Ukraine. Considering that one of the problems in the development of Bayesian networks is the exponential growth in the number of parameters in conditional probability tables (CPT), this study proposes a technique for using noisy-MAX nodes to model economic processes taking into account the time component. It is shown that if the proportion of enterprises that implement innovations is increased now by 31%, while the share of profits of these enterprises increases by only 2%, at the next time step the level of manufacturability and innovation of Ukraine will increase by 81% and will tend to the maximum. © 2020 IEEE.

关键词： Bayesian networks

GainNet: Coordinates the Odd Couple of Generative AI and 6G Networks

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chen, Ning Yang, Jie Cheng, Zhipeng Fan, Xuwei Liu, Zhang Huang, Bangzhen Zhao, Yifeng Huang, Lianfen Du, Xiaojiang Guizani, Mohsen The School of Informatics Xiamen University Xiamen361005 China The National Institute for Data Science in Health and Medicine Xiamen University Xiamen361102 China The School of Future Science and Engineering Soochow University Suzhou215006 China The Department of Electrical and Computer Engineering Stevens Institute of Technology HobokenNJ United States Mohamed Bin Zayed University of Artificial Intelligence Abu Dhabi United Arab Emirates

The rapid expansion of AI-generated content (AIGC) reflects the iteration from assistive AI towards generative AI (GAI) with creativity. Meanwhile, the 6G networks will also evolve from the Internet-of-everything to the Internet-of-intelligence with hybrid heterogeneous network architectures. In the future, the interplay between GAI and the 6G will lead to new opportunities, where GAI can learn the knowledge of personalized data from the massive connected 6G end devices, while GAI’s powerful generation ability can provide advanced network solutions for 6G network and provide 6G end devices with various AIGC services. However, they seem to be an odd couple, due to the contradiction of data and resources. To achieve a better-coordinated interplay between GAI and 6G, the GAI-native networks (GainNet), a GAI-oriented collaborative cloud-edge-end intelligence framework, is proposed in this paper. By deeply integrating GAI with 6G network design, GainNet realizes the positive closed-loop knowledge flow and sustainable-evolution GAI model optimization. On this basis, the GAI-oriented generic resource orchestration mechanism with integrated sensing, communication, and computing (GaiRom-ISCC) is proposed to guarantee the efficient operation of GainNet. Two simple case studies demonstrate the effectiveness and robustness of the proposed schemes. Finally, we envision the key challenges and future directions concerning the interplay between GAI models and 6G networks. Copyright © 2024, The Authors. All rights reserved.

关键词： Queueing networks