检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Liang, Yingyu Long, Jiangxuan Shi, Zhenmei Song, Zhao Zhou, Yufa The University of Hong Kong Hong Kong University of Wisconsin-Madison United States South China University of Technology China The Simons Institute for the Theory of Computing The UC Berkeley United States University of Pennsylvania United States

Large Language Models (LLMs) have shown immense potential in enhancing various aspects of our daily lives, from conversational AI to search and AI assistants. However, their growing capabilities come at the cost of extremely large model sizes, making deployment on edge devices challenging due to memory and computational constraints. This paper introduces a novel approach to LLM weight pruning that directly optimizes for approximating the attention matrix, a core component of transformer architectures. Unlike existing methods that focus on linear approximations, our approach accounts for the non-linear nature of the Softmax attention mechanism. We provide theoretical guarantees for the convergence of our Gradient Descent-based optimization method to a near-optimal pruning mask solution. Our empirical results demonstrate the effectiveness of our non-linear pruning approach in maintaining model performance while significantly reducing computational costs, which is beyond the current state-of-the-art methods, i.e., SparseGPT and Wanda, by a large margin. This work establishes a new theoretical foundation for pruning algorithm design in LLMs, potentially paving the way for more efficient LLM inference on resource-constrained devices. © 2024, CC BY.

关键词： Matrix algebra

来源：评论

学校读者我要写书评

暂无评论

The ESPRIT algorithm under high noise: Optimal error scaling and noisy super-resolution

arXiv

引用

arXiv 2024年

作者： Ding, Zhiyan Epperly, Ethan N. Lin, Lin Zhang, Ruizhe Department of Mathematics University of California Berkeley United States Department of Computing and Mathematical Sciences California Institute of Technology PasadenaCA United States Applied Mathematics and Computational Research Division Lawrence Berkeley National Laboratory United States Challenge Institute for Quantum Computation University of California Berkeley United States Simons Institute for the Theory of Computing United States

Subspace-based signal processing techniques, such as the Estimation of Signal Parameters via Rotational Invariant Techniques (ESPRIT) algorithm, are popular methods for spectral estimation. These algorithms can achieve the so-called super-resolution scaling under low noise conditions, surpassing the well-known Nyquist limit. However, the performance of these algorithms under high-noise conditions is not as well understood. Existing state-of-the-art analysis indicates that ESPRIT and related algorithms can be resilient even for signals where each observation is corrupted by statistically independent, mean-zero noise of size O(1), but these analyses only show that the error ∊ decays at a slow rate ∊ = O͂(n−1/2) with respect to the cutoff frequency n (i.e., the maximum frequency of the measurements). In this work, we prove that under certain assumptions, the ESPRIT algorithm can attain a significantly improved error scaling ∊ = O͂(n−3/2), exhibiting noisy super-resolution scaling beyond the Nyquist limit ∊ = O(n−1) given by the Nyquist-–Shannon sampling theorem. We further establish a theoretical lower bound and show that this scaling is optimal. Our analysis introduces novel matrix perturbation results, which could be of independent interest. © 2024, CC BY.

关键词： Cutoff frequency

来源：评论

学校读者我要写书评

暂无评论

Generating Einstein-Podolsky-Rosen correlations for teleporting collective spin states in a two-dimensional trapped ion crystal

引用

Physical Review Research 2025年第2期7卷 L022019-L022019页

作者： Muhammad Miskeen Khan Edwin Chaparro Bhuvanesh Sundar Allison Carter John Bollinger Klaus Molmer Ana Maria Rey JILA National Institute of Standards and Technology and University of Colorado 440 UCB Boulder Colorado 80309 USA Center for Theory of Quantum Matter University of Colorado Boulder Colorado 80309 USA Present address: Rigetti Computing 775 Heinz Avenue Berkeley California 94710 USA. National Institute of Standards and Technology Boulder Colorado 80309 USA Niels Bohr Institute University of Copenhagen 2100 Copenhagen Denmark

We propose the use of phonon-mediated interactions as an entanglement resource to engineer Einstein-Podolsky-Rosen correlations and to perform teleportation of collective spin states in two-dimensional ion crystals. We emulate continuous variable quantum teleportation protocols between subsystems corresponding to different nuclear spin degrees of freedom. In each of them, a quantum state is encoded in an electronic spin degree of freedom that couples to the vibrational modes of the crystal. We show that high fidelity teleportation of spin-coherent states and their phase-displaced variant, entangled spin-squeezed states, and Dicke states, is possible for realistic experimental conditions in arrays from a few tens to a few hundred ions.

关键词： Degrees of freedom (mechanics)

来源：评论

学校读者我要写书评

暂无评论

Faster Sampling Algorithms for Polytopes with Small Treewidth

Faster Sampling Algorithms for Polytopes with Small Treewidt...

引用

IEEE International Conference on Big Data

作者： Yekun Ke Xiaoyu Li Zhao Song Tianyi Zhou Department of Computer Science University of Washington Seattle WA Department of Computer Science Stevens Institute of Technology Hoboken NJ The Simons Institute for the Theory of Computing University of California Berkeley Berkeley CA Department of Computer Science University of Southern California Los Angeles CA

ISBN: (数字)9798350362480

ISBN: (纸本)9798350362497

Sampling is a fundamental problem in optimization, machine learning and theoretical computer science. A common region of interest for sampling is the polytope, which is defined by a set of linear inequalities. The algorithm that is sampling from polytopes usually requires heavy matrix algebra, including matrix multiplication, matrix inversion and matrix determinant. In this work, we show how to implement the heavy matrix algebra in the area of sampling in nearly linear time for the polytope that has small treewidth. In particular, given a polytope defined by a matrix A ∈ ℝ n×d with treewidth τ, we improve the running time of each iteration for three typical sampling algorithms for polytopes such as Dikin Walk, Soft-Threshold Dikin Walk and Vaidya Walk from O(nd 2 ) to O(nτ 2 ) by exploiting the small treewidth structures of the matrices.

关键词： Computer science Costs Machine learning algorithms Machine learning Big Data Matrices Linear matrix inequalities Computational complexity Optimization

来源：评论

学校读者我要写书评

暂无评论

APT Adversarial Defence Mechanism for Industrial IoT Enabled Cyber-Physical System

引用

IEEE Access 2023年 11卷 74000-74020页

作者： Javed, Safdar Hussain Ahmad, Maaz Bin Asif, Muhammad Akram, Waseem Mahmood, Khalid Das, Ashok Kumar Shetty, Sachin College of Computing and Information Sciences Karachi75190 Pakistan Lahore Garrison University Department of Computer Science Lahore54810 Pakistan National Yunlin University of Science and Technology Graduate School of Intelligent Data Science Douliu64002 Taiwan International Institute of Information Technology at Hyderabad Center for Security Theory and Algorithmic Research Hyderabad500032 India Old Dominion University Simulation and Visualization Engineering Virginia Modeling Analysis and Simulation Center Department of Modeling SuffolkVA23435 United States Old Dominion University Center for Cybersecurity Education and Research SuffolkVA23435 United States

The objective of Advanced Persistent Threat (APT) attacks is to exploit Cyber-Physical Systems (CPSs) in combination with the Industrial Internet of Things (I-IoT) by using fast attack methods. Machine learning (ML) techniques have shown potential in identifying APT attacks in autonomous and malware detection systems. However, detecting hidden APT attacks in the I-IoT-enabled CPS domain and achieving real-time accuracy in detection present significant challenges for these techniques. To overcome these issues, a new approach is suggested that is based on the Graph Attention Network (GAN), a multi-dimensional algorithm that captures behavioral features along with the relevant information that other methods do not deliver. This approach utilizes masked self-attentional layers to address the limitations of prior Deep Learning (DL) methods that rely on convolutions. Two datasets, the DAPT2020 malware, and Edge I-IoT datasets are used to evaluate the approach, and it attains the highest detection accuracy of 96.97% and 95.97%, with prediction time of 20.56 seconds and 21.65 seconds, respectively. The GAN approach is compared to conventional ML algorithms, and simulation results demonstrate a significant performance improvement over these algorithms in the I-IoT-enabled CPS realm. © 2013 IEEE.

关键词： Real time systems

来源：评论

学校读者我要写书评

暂无评论

On the Exploitation of Control Knowledge for Enhancing Automated Planning

SSRN

引用

SSRN 2023年

作者： Lu, Xu Yu, Bin Chen, Chu Tian, Cong Duan, Zhenhua Institute of Computing Theory and Technology State Key Laboratory of Integrated Services Networks Xidian University Xi’an710071 China School of Computer Science Qufu Normal University Qufu273165 China

Heuristic search provides an efficient way to automatically explore the state space in planning, while Control Knowledge (CK) also has the potential to significantly increase the performance of planners. Currently, most of the state-of-the-art planners primarily rely on sophisticated heuristic mechanisms. However, these planners fail to scale up and to provide (high-quality) solutions in a range of problems. Appropriate CK may be capable of making hard problems easy to solve, thus leading planners to achieve much higher efficiency. In this paper, we propose a planning approach by incorporating CK with heuristic search in order to leverage the advantages of both. To achieve this, we encode CK specified by a novel variant of Linear Temporal Logic (LTL), referred to as LTL$_{\mathcal{P}}$, into standard planning models. Consequently, we can directly use existing heuristic planners to solve the augmented problem, and avoid tailoring the planners in order to deal with CK implicitly. Experiments demonstrate that our approach shows highly competitive results versus heuristic search and other techniques on many intractable benchmark problems, benefiting in improving the coverage and quality of plans. © 2023, The Authors. All rights reserved.

关键词： Temporal logic

来源：评论

学校读者我要写书评

暂无评论

Improved Quantum Query Upper Bounds Based on Classical Decision Trees

arXiv

引用

arXiv 2022年

作者： Cornelissen, Arjan Mande, Nikhil S. Patro, Subhasree Simons Institute for the Theory of Computing University of California Berkeley United States University of Liverpool United Kingdom Eindhoven University of Technology Netherlands

We consider the following question in query complexity: Given a classical query algorithm in the form of a decision tree, when does there exist a quantum query algorithm with a speedup (i.e., that makes fewer queries) over the classical one? We provide a general construction based on the structure of the underlying decision tree, and prove that this can give us an up-to-quadratic quantum speed-up in the number of queries. In particular, our results give a bounded-error quantum query algorithm of cost O(√s) to compute a Boolean function (more generally, a relation) that can be computed by a classical (even randomized) decision tree of size s. This recovers an O(√n) algorithm for the Search problem, for example. Lin and Lin [theory of computing’16] and Beigi and Taghavi [Quantum’20] showed results of a similar flavor. Their upper bounds are in terms of a quantity which we call the "guessing complexity" of a decision tree. We identify that the guessing complexity of a decision tree equals its rank, a notion introduced by Ehrenfeucht and Haussler [Information and Computation’89] in the context of learning theory. This answers a question posed by Lin and Lin, who asked whether the guessing complexity of a decision tree is related to any measure studied in classical complexity theory. We also show a polynomial separation between rank and its natural randomized analog for the complete binary AND-OR tree. Beigi and Taghavi constructed span programs and dual adversary solutions for Boolean functions given classical decision trees computing them and an assignment of non-negative weights to edges of the tree. We explore the effect of changing these weights on the resulting span program complexity and objective value of the dual adversary bound, and capture the best possible weighting scheme by an optimization program. We exhibit a solution to this program and argue its optimality from first principles. We also exhibit decision trees for which our bounds are asymptotically stro

关键词： Boolean functions

来源：评论

学校读者我要写书评

暂无评论

SLA-SCA: Secure and Lightweight Authentication and Key Agreement Mechanism for Smart City Applications

SLA-SCA: Secure and Lightweight Authentication and Key Agree...

引用

2023 IEEE International Conference on Communications, computing, Cybersecurity and Informatics, CCCI 2023

作者： Pant, Bhairvi Obaidat, Mohammad S. Singh, Shreya Pant, Mudita Wazid, Mohammad Das, Ashok Kumar Hsiao, Kuei-Fang Graphic Era Deemed to Be University Department of Computer Science and Engineering Dehradun248002 India King Abdullah Ii School of Information Technology The University of Jordan Amman11942 Jordan University of Science and Technology Beijing Beijing100083 China School of Computing Srm University TN Kattankulathur603203 India Ctr. for Security Theory and Algorithmic Research International Institute of Information Technology Hyderabad 500032 India University of Texas-Permian Basin Computer Science Department OdessaTX 79762 United States

ISBN: (纸本)9798350302561

The notion of a smart city is a modern concept characterized by the widespread integration of the Internet of Things (IoT) across various facets of urban life, including healthcare, traffic management, agriculture, and more. This establishes real-time communication and data exchange between all users, servers and devices through the Internet. As we begin to discuss communication between parties, one must also recognize the imperative need for this exchange to be secure and private, as such exchanges are extremely vulnerable to attacks such as man-in-the-middle and replay attacks and even situations like compromising of anonymity during exchanges. This paper focuses on finding an innovative solution to the same challenge. It proposes a mechanism to establish secure communication between the smart IoT devices and the connected cloud servers through a secure mutual authentication and key establishment process with a short name 'SLA-SCA'. The security analysis of the proposed SLA-SCA is done, which proves its resilience against the various potential attacks of such a communication. The proposed SLA-SCA outperformed the other existing schemes regarding important security and functionality features. © 2023 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Propositional Projection Temporal Logic Specification Mining 14th

Propositional Projection Temporal Logic Specification Mining

引用

14th International Conference on Combinatorial Optimization and Applications, COCOA 2020

作者： Zhang, Nan Yuan, Xiaoshuai Duan, Zhenhua Institute of Computing Theory and Technology and ISN Laboratory Xidian University Xi’an710071 China

ISBN: (纸本)9783030648428

This paper proposes a dynamic approach of specification mining for Propositional Projection Temporal Logic (PPTL). To this end, a pattern library is built to collect some common temporal relation among events. Further, several algorithms of specification mining for PPTL are designed. With our approach, PPTL specifications are mined from a trace set of a target program by using patterns in the library. In addition, a specification mining tool PPTLMiner supporting this approach is developed. In practice, given a trace set and user selected patterns, PPTLMiner can capture PPTL specifications of target programs. © 2020, Springer Nature Switzerland AG.

关键词： Temporal logic

来源：评论

学校读者我要写书评

暂无评论

Inclusive Hadronic Decay Rate of the τ Lepton from Lattice QCD: The u¯s Flavor Channel and the Cabibbo Angle

引用

Physical Review Letters 2024年第26期132卷 261901-261901页

作者： Constantia Alexandrou Simone Bacchio Alessandro De Santis Antonio Evangelista Jacob Finkenrath Roberto Frezzotti Giuseppe Gagliardi Marco Garofalo Bartosz Kostrzewa Vittorio Lubicz Simone Romiti Francesco Sanfilippo Silvano Simula Nazario Tantalo Carsten Urbach Urs Wenger Department of Physics Computation-based Science and Technology Research Center Dipartimento di Fisica and Gaustrae 20 42119 Wuppertal Germany Istituto Nazionale di Fisica Nucleare HISKP (Theory) High Performance Computing and Analytics Lab Dipartimento di Matematica e Fisica Institute for Theoretical Physics Albert Einstein Center for Fundamental Physics

We present a lattice determination of the inclusive decay rate of the process τ↦Xusντ in which the τ lepton decays into a generic hadronic state Xus with u¯s flavor quantum numbers. Our results have been obtained in nf=2+1+1 isosymmetric QCD with full nonperturbative accuracy, without any operator product expansion approximation and, except for the presently missing long-distance isospin-breaking corrections, include a solid estimate of all sources of theoretical uncertainties. This has been possible by using the Hansen-Lupo-Tantalo method [M. Hansen et al., Phys. Rev. D 99, 094508 (2019)] that we have already successfully applied [A. Evangelista et al., Phys. Rev. D 108, 074513 (2023)] to compute the inclusive decay rate of the process τ↦Xudντ in the u¯d flavor channel. By combining our first-principles theoretical results with the presently available experimental data, we extract the Cabibbo-Kobayashi-Maskawa matrix element |Vus|, the Cabibbo angle, with a 0.9% accuracy, dominated by the experimental error.

关键词： Cabibbo–Kobayashi–Maskawa matrix Hadronic decays Lattice field theory Lattice gauge theory Lepton-hadron interactions Non-Abelian gauge theories Nonperturbative effects in field theory Phenomenology Quantum chromodynamics Strong interaction Total cross sections

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：