检索结果-内蒙古大学图书馆

IEEE Transactions on Artificial Intelligence 2024年第7期5卷 3470-3481页

作者： Burton-Barr, Jonathan Fernando, Basura Rajan, Deepu Institute of High-Performance Computing The Centre for Frontier AI Research Agency for Science Technology and Research 138632 Singapore Nanyang Technological University Nanyang School of Computer Science and Engineering 639798 Singapore

As artificial intelligence (AI) systems become more complex and widespread, they require significant computational power, increasing energy consumption. Addressing this challenge is essential for ensuring the long-term sustainability of AI technology. AI-on-AI control refers to a system with a set of AI functions controlled by an upper-level AI model. Previous work in AI-on-AI control focuses on boosting accuracy or expanding system capability by increasing overall system cost. Alternatively, we focus on applying AI-on-AI control to decrease system cost and increase the sustainability and viability of a system with multiple AI functions. Our supervised image classification evaluative controller (SICEC) is a cost-reduction oriented AI-on-AI controller that learns when vision models within an AI system should be activated based on input features. The function controller (FC) preprocesses an input and activates relevant functions, functions being distinct units of AI functionality within the system. Some functions have a set of same functional models (SFMs). These models take the same input and produce the same output but have architectural differences. We introduce a same functional controller to select a SFM using the FC's decision confidence. Results are promising, with a decrease of up to 48.9% in inference time, 67.8% in floating point operations (FLOPs), and 66.4% in energy usage. With SICEC showing significant reductions in inference time and energy cost, our work contributes to limited resource computing and sustainable AI technology. © 2020 IEEE.

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

Situational Scene Graph for Structured Human-Centric Situation Understanding

Situational Scene Graph for Structured Human-Centric Situati...

引用

2025 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2025

作者： Sugandhika, Chinthani Li, Chen Rajan, Deepu Fernando, Basura College of Computing and Data Science Nanyang Technological University Singapore Centre for Frontier Ai Research Agency for Science Technology and Research Singapore Institute of High-Performance Computing Agency for Science Technology and Research Singapore

ISBN: (纸本)9798331510831

Graph based representation has been widely used in modelling spatio-temporal relationships in video understanding. Although effective, existing graph-based approaches focus on capturing the human-object relationships while ignoring fine-grained semantic properties of the action components. These semantic properties are crucial for understanding the current situation, such as where does the action takes place, what tools are used and functional properties of the objects. In this work, we propose a graph-based representation called Situational Scene Graph (SSG) to encode both human-object relationships and the corresponding semantic properties. The semantic details are represented as predefined roles and values inspired by situation frame, which is originally designed to represent a single action. Based on our proposed representation, we introduce the task of situational scene graph generation and propose a multi-stage pipeline Interactive and Complementary Network (InComNet) to address the task. Given that the existing datasets are not applicable to the task, we further introduce a SSG dataset whose annotations consist of semantic role-value frames for human, objects and verb predicates of human-object relations. Finally, we demonstrate the effectiveness of our proposed SSG representation by testing on different downstream tasks. Experimental results show that the unified representation can not only benefit predicate classification and semantic role-value classification, but also benefit reasoning tasks on human-centric situation understanding. Code and data are available at https://***/LUNAProject22/SSG. © 2025 IEEE.

关键词： Human engineering

来源：评论

学校读者我要写书评

暂无评论

Can LLMs Perform Structured Graph Reasoning Tasks? 27th

Can LLMs Perform Structured Graph Reasoning Tasks?

引用

27th International Conference on Pattern Recognition, ICPR 2024

作者： Agrawal, Palaash Vasania, Shavak Tan, Cheston Center for Frontier AI Research Agency for Science Technology and Research Singapore Singapore Institute of High Performance Computing Agency for Science Technology and Research Singapore Singapore UWCSEA Dover High School Singapore Singapore

ISBN: (纸本)9783031781711

Pretrained Large Language Models (LLMs) have demonstrated various reasoning capabilities through language-based prompts alone, particularly in unstructured task settings (i.e. tasks purely based on language semantics). However, LLMs often struggle with structured tasks, because of the inherent incompatibility of input representation. Reducing structured tasks to uni-dimensional language semantics often renders the problem trivial. Keeping the trade-off between LLM compatibility and structure complexity in mind, we design various graph reasoning tasks as a proxy to semi-structured tasks in this paper, in order to test LLMs’ ability to use representations beyond plain text. In particularly, we design 10 distinct problems of graph traversal, each representing increasing levels of complexity, and benchmark 5 different instruct-finetuned LLMs (GPT-4, GPT-3.5, Claude-2, Llama-2 and Palm-2) on the aforementioned tasks. Further, we analyse the performance of models across various settings such as varying sizes of graphs, as well as different forms of k-shot prompting. We highlight various limitations, biases and properties of LLMs through this benchmarking process, such as an inverse relation to the average degrees of freedom of traversal per node in graphs, the overall negative impact of k-shot prompting on graph reasoning tasks, and a positive response bias which prevents LLMs from identifying the absence of a valid solution. Finally, we introduce a new prompting technique specially designed for graph traversal tasks (PathCompare), which demonstrates a notable increase in the performance of LLMs in comparison to standard prompting techniques such as Chain-of-Thought (CoT) (The code for reproducing the results of this paper can be found at https://***/PalaashAgrawal/LLMGraphTraversal). © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Simulation of the Quantum Key Distribution Algorithm Using the Intel Quantum SDK 13

Simulation of the Quantum Key Distribution Algorithm Using t...

引用

13th International Conference on Advanced Computer Information Technologies, ACIT 2023

作者： Gambo, Yaknan John Shinde, Tejas Rasch, Kevin Liebelt, Helena Li, Rui Deggendorf Institute of Technology High Performance Computing and Quantum Computing Deggendorf Germany Quantum Computing Research Intel Inc HillsboroOR United States Deggendorf Institute of Technology Department of Computer Science Deggendorf Germany

Information security is an important fabric of our world today. Technologies used in the development of information security have constantly been evolving in response to new threats. Cryptographic algorithms like the RSA, DSA, AES etc. are currently being used in securing information even over insecure channels. The coming of quantum computers has now posed a great threat to classical cryptosystems which can be broken in polynomial time thanks to Shor's algorithm. A good hacker with good quantum computing power can break into encrypted messages easily. The Quantum Key Distribution (QKD) algorithm based on the BB84 protocol provides a solution to this problem. It makes it possible for communicating parties to exchange secret keys securely before a conversation can ensue. Verifying if these keys have been tampered with is a good first step to start before any other. QKD technology has made it possible to exchange these keys via an insecure channel and lets communicating parties know if an eavesdropper has tempered with the message. In this work, we discuss the QKD algorithm and simulated it on the newly released Intel quantum SDK. © 2023 IEEE.

关键词： Quantum cryptography

来源：评论

学校读者我要写书评

暂无评论

Road Network Representation Learning with the Third Law of Geography 38

Road Network Representation Learning with the Third Law of G...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Zhou, Haicang Huang, Weiming Chen, Yile He, Tiantian Cong, Gao Ong, Yew-Soon College of Computing and Data Science Nanyang Technological University Singapore Institute of High-Performance Computing Agency for Science Technology and Research Singapore Centre for Frontier Al Research Agency for Science Technology and Research Singapore

Road network representation learning aims to learn compressed and effective vectorized representations for road segments that are applicable to numerous tasks. In this paper, we identify the limitations of existing methods, particularly their overemphasis on the distance effect as outlined in the First Law of Geography. In response, we propose to endow road network representation with the principles of the recent Third Law of Geography. To this end, we propose a novel graph contrastive learning framework that employs geographic configuration-aware graph augmentation and spectral negative sampling, ensuring that road segments with similar geographic configurations yield similar representations, and vice versa, aligning with the principles stated in the Third Law. The framework further fuses the Third Law with the First Law through a dual contrastive learning objective to effectively balance the implications of both laws. We evaluate our framework on two real-world datasets across three downstream tasks. The results show that the integration of the Third Law significantly improves the performance of road segment representations in downstream tasks. Our code is available at https://***/Haicang/Garner. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Improving the spectral purity of single photons by a single-interferometer-coupled microring

引用

Chinese Physics B 2022年第3期31卷 311-315页

作者： Yang Wang Pingyu Zhu Shichuan Xue Yingwen Liu Junjie Wu Xuejun Yang Ping Xu Institute for Quantum Information & State Key Laboratory of High Performance Computing College of Computer Science and TechnologyNational University of Defense TechnologyChangsha 410013China

We experimentally engineer a high-spectral-purity single-photon source using a single-interferometer-coupled silicon microring. By the reconfiguration of the interferometer, different coupling conditions can be obtained, corresponding to different quality factors for the pump and signal/idler. The ratio between the quality factor of the pump and signal/idler ranges from 0.29 to 2.57. By constructing the signal–idler joint spectral intensity, we intuitively demonstrate the spectral correlation of the signal and idler. As the ratio between the quality factor of the pump and signal/idler increases, the spectral correlation of the signal and idler decreases, i.e., the spectral purity of the signal/idler photons increases. Furthermore,time-integrated second-order correlation of the signal photons is measured, giving a value up to 94.95 ± 3.46%. Such high-spectral-purity photons will improve the visibility of quantum interference and facilitate the development of on-chip quantum information processing.

关键词： spectral-purity microring asymmetric Mach–Zehnder interferometer

来源：评论

学校读者我要写书评

暂无评论

DoFIT: Domain-aware Federated Instruction Tuning with Alleviated Catastrophic Forgetting 38

DoFIT: Domain-aware Federated Instruction Tuning with Allevi...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Xu, Binqian Shu, Xiangbo Mei, Haiyang Bai, Zechen Fernando, Basura Shou, Mike Zheng Tang, Jinhui Nanjing University of Science and Technology China Show Lab National University of Singapore Singapore Institute of High-Performance Computing A*STAR Singapore

Federated Instruction Tuning (FIT) advances collaborative training on decentralized data, crucially enhancing model's capability and safeguarding data privacy. However, existing FIT methods are dedicated to handling data heterogeneity across different clients (i.e., client-aware data heterogeneity), while ignoring the variation between data from different domains (i.e., domain-aware data heterogeneity). When scarce data needs supplementation from related fields, these methods lack the ability to handle domain heterogeneity in cross-domain training. This leads to domain-information catastrophic forgetting in collaborative training and therefore makes model perform sub-optimally on the individual domain. To address this issue, we introduce DoFIT, a new Domain-aware FIT framework that alleviates catastrophic forgetting through two new designs. First, to reduce interference information from the other domain, DoFIT finely aggregates overlapping weights across domains on the inter-domain server side. Second, to retain more domain information, DoFIT initializes intra-domain weights by incorporating inter-domain information into a less-conflicted parameter space. Experimental results on diverse datasets consistently demonstrate that DoFIT excels in cross-domain collaborative training and exhibits significant advantages over conventional FIT methods in alleviating catastrophic forgetting. Code is available at https://***/1xbq1/DoFIT. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Solving 2 by 2 Grid Sudoku Problem using Grover's Algorithm with Intel Quantum SDK 13

Solving 2 by 2 Grid Sudoku Problem using Grover's Algorithm ...

引用

13th International Conference on Advanced Computer Information Technologies, ACIT 2023

作者： Shinde, Tejas Gambo, Yaknan John Rasch, Kevin Liebelt, Helena Li, Rui Deggendorf Institute of Technology High Performance Computing and Quantum Computing Deggendorf Germany Quantum Computing Research Intel Inc HillsboroOR United States Deggendorf Institute of Technology Department of Applied Computer Science Deggendorf Germany

Quantum technologies are moving towards enabling real world uses and as these technologies develop. The most important thing that differs Quantum from classical computers is that Quantum Computers can do 'multiple' tasks at the same time as opposed to the classical computers doing a single task at a time. This phenomenon has led to the discovery of a few algorithms that are faster than regular computers. One such algorithm is the Grover's Algorithm[l]. This algorithm has shown a few real-world applications such as Quantum Search Algorithm, Travelling salesman, etc[2]. In this paper, Grover's algorithm is introduced, and a 2 x 2 Sudoku example is solved using Intel Quantum Software Development Kit[3]. The Intel Quantum SDK is quite new and in this paper we show how a simple Grover's Problem can be implemented in the SDK. Sudoku is a popular number-based puzzle game played on a grid consisting of nine squares subdivided into smaller 3x3 grids. The objective is to fill the grid with numbers from 1 to 9, ensuring that each row, column, and 3x3 grid contains all numbers exactly once. © 2023 IEEE.

关键词： Qubits

来源：评论

学校读者我要写书评

暂无评论

Predicting the Next Action by Modeling the Abstract Goal 27th

Predicting the Next Action by Modeling the Abstract Goal

引用

27th International Conference on Pattern Recognition, ICPR 2024

作者： Roy, Debaditya Fernando, Basura Institute of High-Performance Computing Agency for Science Technology and Research Singapore Singapore Centre for Frontier AI Research Agency for Science Technology and Research Singapore Singapore

ISBN: (纸本)9783031783531

The problem of predicting human actions from observed videos is an inherently uncertain one. We present an action anticipation model that leverages latent goal information to reduce the uncertainty in future predictions. We develop a latent variable representing goal information called abstract goal which is conditioned on observed sequences of visual features for action anticipation. We design the abstract goal as a distribution whose parameters are estimated using a variational recurrent model. We sample multiple candidates for the next action and use goal consistency criterion to determine the best candidate that follows from the abstract goal. Our method obtains impressive results on the very challenging Epic-Kitchens55 (EK55) and good results in Epic-Kitchens100 (EK100) datasets. Code is available at https://***/LAHAproject/Abstract_Goal. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Thermal conductivity of GeTe crystals based on machine learning potentials

引用

Chinese Physics B 2024年第4期33卷 104-107页

作者：张健张昊春李伟峰张刚 School of Energy Science and Engineering Harbin Institute of TechnologyHarbin 150001China Institute of High Performance Computing Agency for ScienceTechnology and ResearchSingapore 138632Singapore School of Physics&State Key Laboratory of Crystal Materials Shandong UniversityJinan 250100China

GeTe has attracted extensive research interest for thermoelectric *** this paper,we first train a neuroevolution potential(NEP)based on a dataset constructed by ab initio molecular dynamics,with the Gaussian approximation potential(GAP)as a *** phonon density of states is then calculated by two machine learning potentials and compared with density functional theory results,with the GAP potential having higher ***,the thermal conductivity of a GeTe crystal at 300 K is calculated by the equilibrium molecular dynamics method using both machine learning potentials,and both of them are in good agreement with the experimental results;however,the calculation speed when using the NEP potential is about 500 times faster than when using the GAP ***,the lattice thermal conductivity in the range of 300 K-600 K is calculated using the NEP *** lattice thermal conductivity decreases as the temperature increases due to the phonon anharmonic *** study provides a theoretical tool for the study of the thermal conductivity of GeTe.

关键词： machine learning potentials thermal conductivity molecular dynamics

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

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

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：