检索结果-内蒙古大学图书馆

31st IEEE International Conference on High Performance Computing, data, and Analytics Workshops, HiPCW 2024

作者： Joshi, Sarthak Vadhiyar, Sathish Department of Computational and Data Sciences Indian Institute of Science Bengaluru India

ISBN: (纸本)9798331509118

As we have entered Exascale computing, the faults in high-performance systems are expected to increase considerably. To compensate for a higher failure rate, the standard checkpoint/restart technique would need to create checkpoints at a much higher frequency, resulting in an excessive amount of overhead, which would not be sustainable for many scientific applications. To improve application efficiency in such high-failure environments, the mechanism of replication of MPI processes was proposed. Replication allows for fast recovery from failures by simply dropping the failed processes and using their replicas to continue the regular operation of the application. We have implemented FTHP-MPI (Fault Tolerance and High-Performance MPI), a novel fault-tolerant MPI library that augments checkpoint/restart with replication to provide resilience from failures. The novelty of our work is that it is designed to provide fault tolerance in a native MPI library. This lets application developers achieve fault tolerance at high failure rates while also using efficient communication protocols in the native MPI libraries that are generally fine-tuned for specific HPC platforms. Our work employs various concepts, including MPI-agnostic checkpointing from DMTCP and communicator shrinking from ULFM, and implements them together without any code modifications on the user end. We have also implemented efficient parallel communication techniques that involve replicas. © 2024 IEEE

关键词： Failure rate

来源：评论

学校读者我要写书评

暂无评论

Prediction of Small Molecules Binding Site with RNA as a Target 15

Prediction of Small Molecules Binding Site with RNA as a Tar...

引用

15th International Conference on Computing Communication and Networking Technologies, ICCCNT 2024

作者： Majumder, Subhasree Pal, Debnath Indian Institute of Science Bangalore Dept. of Computational and Data Science Bangalore India

ISBN: (纸本)9798350370249

RNA are essential biomolecules and are potent drug targets, as many small RNA molecules are known to play significant roles in many disease pathways. Predicting drug binding sites in RNA molecules is thus an essential step towards RNA-based drug discovery. Several previous works have focussed on sequence-based and structure-based features to predict binding sites in proteins and RNA. The low availability of RNA-small structures is a bottleneck for predictive model performance. We show that using an atomic-level feature set instead of residue-level feature set, increases the dataset size and can help build a promising model using a dataset of smaller size, which is the primary concern in RNA-Small molecule complexes. We focus on a rich set of physio-chemical, structural information, and dynamic features at the atomic level. Lastly we use a voting scheme to predict the small molecule binding sites for RNA. K-Nearest Neighbour (KNN), Multilayer Perceptron (MLP), Decision Tree and Random Forest models performed well with MCC values above 0.5. Among them, Random Forest performed best with an MCC of 0.78, and an accuracy of 0.89. © 2024 IEEE.

关键词： Binding sites

来源：评论

学校读者我要写书评

暂无评论

A Preliminary Performance Analysis of LLM Inference on Edge Accelerators 31

A Preliminary Performance Analysis of LLM Inference on Edge ...

引用

31st IEEE International Conference on High Performance Computing, data, and Analytics Workshops, HiPCW 2024

作者： Arya, Mayank Simmhan, Yogesh Department of Computational and Data Sciences Indian Institute of Science Bangalore560012 India

ISBN: (纸本)9798331509118

Large Language Models (LLMs) have revolutionized machine learning by efficiently handling tasks like translation, text generation, summarization, and classification, making them ideal for applications such as virtual assistants and agentic work-flows. However, their substantial computational requirements limit their wide adoption, particularly for smaller organizations lacking the computing infrastructure. Current LLM serving is done on the cloud, which imposes cost, latency and privacy concerns. Cloud-edge collaboration also struggles with network bottlenecks and model partitioning complexities. This work explores the potential for using edge accelerators, like Nvidia’s Jetson, for LLM inference. These compact devices have 100-1000s of GPU cores and 10+GB of shared GPU memory, allowing real-time edge inference close to the data source. We analyze and evaluate the efficiency of edge devices for LLM inference for different models and power modes, and make a case for tuning power modes to reduce power usage by ≈ 4-6% while matching the throughput of the highest performance mode. © 2024 IEEE

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

Performance Trade-offs in GNN Inference: An Early Study on Hardware and Sampling Configurations 31

Performance Trade-offs in GNN Inference: An Early Study on H...

引用

31st IEEE International Conference on High Performance Computing, data, and Analytics Workshops, HiPCW 2024

作者： Naman, Pranjal Simmhan, Yogesh Department of Computational and Data Sciences Indian Institute of Science Bangalore560012 India

ISBN: (纸本)9798331509118

Inferencing over Graph Neural Network (GNN) has recently gained traction in applications like recommender systems and fraud detection. However, building a high throughput, low latency GNN serving system for large graphs in real-world applications is challenging due to the irregular memory access patterns and high host-device data transfer times. In this poster, we study the effects of various hardware and sampling configurations on the latency and performance of GNN serving, which will help build scalable GNN serving systems. Our results show that sampling on CPU and inference on GPU, popular for training, does not do well for inferencing. Smaller fan-outs give reasonable accuracies but suffer from high variability while GPU-based sampling fails with high fan-outs, making Unified Virtual Addressing (UVA) based sampling a preferred alternative. © 2024 IEEE

关键词： Virtual addresses

来源：评论

学校读者我要写书评

暂无评论

Towards a Generalized SDK for a Programmable Drones-as-a-Service 31

Towards a Generalized SDK for a Programmable Drones-as-a-Ser...

引用

31st IEEE International Conference on High Performance Computing, data, and Analytics Workshops, HiPCW 2024

作者： Raj, Suman Singh, Rajdeep Astu, Kautuk Simmhan, Yogesh Department of Computational and Data Sciences Indian Institute of Science Bangalore560012 India

ISBN: (纸本)9798331509118

The proliferation of Unmanned Aerial Vehicles (UAVs) in spatio-temporal applications has given rise to the need for Drone-as-a-Service (DaaS) platforms that support diverse tasks, ranging from urban air mobility to precision agriculture. However, existing DaaS platforms are often tightly-coupled to hardware-specific frameworks that lack programmability and flexibility for custom applications. For instance, using UAVs to help a Visually Impaired Person (VIP) navigate requires a suite of applications, such as obstacle avoidance, route planning, and real-time tracking. In this poster, we propose Ocularone SDK (O-SDK) that integrates on our prior Ocularone stack for VIP applications with Aerostack2, which generalizes the hardware abstraction. O-SDK offers an analytics layer as a novel contribution to enable developers to easily integrate Deep Neural Network (DNN)-based models, and interface swarms of UAVs for complex tasks like navigating the visually impaired. Our evaluations show that the minimal overheads of using O-SDK for real-time experiments using Tello drones and Jetson Orin Nano. This opens the potential to develop a domain-neutral, programmable and customizable DaaS Software Development Kit (SDK) in future. © 2024 IEEE

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

A Tutorial on Federated Learning from Theory to Practice:Foundations,Software Frameworks,Exemplary Use Cases,and Selected Trends

引用

IEEE/CAA Journal of Automatica Sinica 2024年第4期11卷 824-850页

作者： M.Victoria Luzón Nuria Rodríguez-Barroso Alberto Argente-Garrido Daniel Jiménez-López Jose M.Moyano Javier Del Ser Weiping Ding Francisco Herrera Department of Software Engineering Andalusian Research Institute in Data Science and Computational Intelligence(DaSCI)University of GranadaGranada 18071Spain Department of Computer Science and Artificial Intelligence Andalusian Research Institute in Data Science and Computational Intelligence(DaSCI)University of GranadaGranada 18071Spain Department of Communications Engineering University of the Basque Country(UPV/EHU)and also with TECNALIABasque Research&Technology Alliance(BRTA)Spain School of Information Science and Technology Nantong UniversityNantong 226019China

When data privacy is imposed as a necessity,Federated learning(FL)emerges as a relevant artificial intelligence field for developing machine learning(ML)models in a distributed and decentralized *** allows ML models to be trained on local devices without any need for centralized data transfer,thereby reducing both the exposure of sensitive data and the possibility of data interception by malicious third *** paradigm has gained momentum in the last few years,spurred by the plethora of real-world applications that have leveraged its ability to improve the efficiency of distributed learning and to accommodate numerous participants with their data *** virtue of FL,models can be learned from all such distributed data sources while preserving data *** aim of this paper is to provide a practical tutorial on FL,including a short methodology and a systematic analysis of existing software ***,our tutorial provides exemplary cases of study from three complementary perspectives:i)Foundations of FL,describing the main components of FL,from key elements to FL categories;ii)Implementation guidelines and exemplary cases of study,by systematically examining the functionalities provided by existing software frameworks for FL deployment,devising a methodology to design a FL scenario,and providing exemplary cases of study with source code for different ML approaches;and iii)Trends,shortly reviewing a non-exhaustive list of research directions that are under active investigation in the current FL *** ultimate purpose of this work is to establish itself as a referential work for researchers,developers,and data scientists willing to explore the capabilities of FL in practical applications.

关键词： data privacy distributed machine learning federated learning software frameworks

来源：评论

学校读者我要写书评

暂无评论

Machine learning-based gene expression biomarkers to distinguish Zika and Dengue virus infections: implications for diagnosis

引用

VirusDisease 2024年第3期35卷 446-461页

作者： Zeba, Ayesha Rajalingam, Aruna Sekar, Kanagaraj Ganjiwale, Anjali Department of Life Science Bangalore University Laboratory for Structural Biology and Bio-Computing Department of Computational and Data Sciences Indian Institute of Science

Zika virus (ZIKV) and Dengue virus (DENV) infections cause severe disease in humans and are significant socio-economic burden worldwide. These flavivirus infections are difficult to diagnose serologically due to antigenic overlap. The phylogenetic analysis shows that ZIKV clusters with DENVs at a higher node of the phylogenetic tree with significant genomic and structural similarity. Our study aims to identify gene biomarkers for the classification of Dengue and Zika viral infections using machine learning algorithms and bioinformatics analysis. The gene expression count matrix for single-cell RNA sequencing dataset GSE110496 was analyzed using binary classifiers, namely Logistic regression, Support Vector Machines, Random Forest, and Decision trees. The GSE110496 dataset represents a unique study of the transcriptional and translational dynamics of DENV and ZIKV infections at 4-, 12-, 24-, and 48-h time points for human hepatoma (Huh7) cells. Out of which 24-h time point has been analyzed in this study, at the optimal threshold of viral molecules. Feature selection was performed using two different approaches Random Forest Classifier (RFC) for gene ranking and Recursive Feature Elimination (RFE). Out of which RFE, showed more accuracy and precision. The classification accuracy of 89.4% and the precision of 90% were obtained using selected 10 gene features. SCY1 Like Pseudokinase 3 (SCYL3), Chromosome 1 Open Reading Frame 112 (C1orf112), Complement factor H (CFH), Heme-binding protein 1 (HEBP1), Cadherin 1 (CDH1), Nibrin (NBN), Histone deacetylase 5 (HDAC5), nuclear receptor subfamily 0, group B, member 2 (NR0B2), Annexin A9 (ANXA9) and Alcohol dehydrogenase 6 (ADH6) are the proposed gene biomarkers in this study. The functional analysis of the reported biomarkers was performed using KEGG and GO with the WEB-based Gene SeT AnaLysis Toolkit (WebGestalt). The relationship of the selected biomarkers with DENV and ZIKV infections analyzed using a gene–gene interaction n

关键词： Biomarkers Dengue Machine learning Single cell RNA sequencing Zika

来源：评论

学校读者我要写书评

暂无评论

Co-Linear Chaining on Pangenome Graphs 23

Co-Linear Chaining on Pangenome Graphs

引用

23rd International Workshop on Algorithms in Bioinformatics, WABI 2023

作者： Rajput, Jyotshna Chandra, Ghanshyam Jain, Chirag Department of Computational and Data Sciences Indian Institute of Science Bangalore India

ISBN: (纸本)9783959772945

Pangenome reference graphs are useful in genomics because they compactly represent the genetic diversity within a species, a capability that linear references lack. However, efficiently aligning sequences to these graphs with complex topology and cycles can be challenging. The seed-chain-extend based alignment algorithms use co-linear chaining as a standard technique to identify a good cluster of exact seed matches that can be combined to form an alignment. Recent works show how the co-linear chaining problem can be efficiently solved for acyclic pangenome graphs by exploiting their small width [Makinen et al., TALG'19] and how incorporating gap cost in the scoring function improves alignment accuracy [Chandra and Jain, RECOMB'23]. However, it remains open on how to effectively generalize these techniques for general pangenome graphs which contain cycles. Here we present the first practical formulation and an exact algorithm for co-linear chaining on cyclic pangenome graphs. We rigorously prove the correctness and computational complexity of the proposed algorithm. We evaluate the empirical performance of our algorithm by aligning simulated long reads from the human genome to a cyclic pangenome graph constructed from 95 publicly available haplotype-resolved human genome assemblies. While the existing heuristic-based algorithms are faster, the proposed algorithm provides a significant advantage in terms of accuracy. © Jyotshna Rajput, Ghanshyam Chandra, and Chirag Jain;licensed under Creative Commons License CC-BY 4.0.

关键词： Graphic methods

来源：评论

学校读者我要写书评

暂无评论

PARAMETERIZED DOUGLAS-RACHFORD DYNAMICAL SYSTEM FOR MONOTONE INCLUSION PROBLEMS

Applied Set-Valued Analysis and Optimization

引用

Applied Set-Valued Analysis and Optimization 2023年第1期5卷 19-29页

作者： Gautam, Pankaj Som, Kuntal Vetrivel, V. Department of Mathematics Indian Institute of Technology Madras Chennai India Computational Mathematics and Data Science Lab Indian Institute of Technology Madras Chennai India

Douglas-Rachford splitting method with resolvent operator is a renowned algorithm to solve monotone inclusion problem involving sum of two monotone operators. In this paper, we investigate a Douglas-Rachford-based dynamical systems designed to approach the solution sets of inclusion problems involving the sum of two maximally monotone operators. Our main aim is to use parametrized resolvent instead of classical resolvent as the Douglas-Rachford operator in the framework of preconditioning. The convergence of the orbit is demonstrated. We also add a Tikhonov regularized term (both inner and outer regularization) to obtain strong convergence of the induced orbit. ©2023 Applied Set-Valued Analysis and Optimization.

关键词： Dynamical systems

来源：评论

学校读者我要写书评

暂无评论

Information Geometric Approaches for Patient-Specific Test-Time Adaptation of Deep Learning Models for Semantic Segmentation

引用

IEEE Transactions on Medical Imaging 2025年第6期44卷 2553-2567页

作者： Ravishankar, Hariharan Paluru, Naveen Sudhakar, Prasad Yalavarthy, Phaneendra K. Indian Institute of Science Department of Computational and Data Sciences Bengaluru560012 India John F. Welch Technology Centre Wipro GE HealthCare Bengaluru560066 India

The test-time adaptation (TTA) of deep-learning-based semantic segmentation models, specific to individual patient data, was addressed in this study. The existing TTA methods in medical imaging are often unconstrained, require anatomical prior information or additional neural networks built during training phase, making them less practical, and prone to performance deterioration. In this study, a novel framework based on information geometric principles was proposed to achieve generic, off-the-shelf, regularized patient-specific adaptation of models during test-time. By considering the pre-trained model and the adapted models as part of statistical neuromanifolds, test-time adaptation was treated as constrained functional regularization using information geometric measures, leading to improved generalization and patient optimality. The efficacy of the proposed approach was shown on three challenging problems: 1) improving generalization of state-of-the-art models for segmenting COVID-19 anomalies in Computed Tomography (CT) images 2) cross-institutional brain tumor segmentation from magnetic resonance (MR) images, 3) segmentation of retinal layers in Optical Coherence Tomography (OCT) images. Further, it was demonstrated that robust patient-specific adaptation can be achieved without adding significant computational burden, making it first of its kind based on information geometric principles. © 1982-2012 IEEE.

关键词： Semantic Segmentation

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：