Snap ML: A Hierarchical Framework for Machine Learning

被引:0
|
作者
Dunner, Celestine [1 ]
Parnell, Thomas [1 ]
Sarigiannis, Dimitrios [1 ]
Ioannou, Nikolas [1 ]
Anghel, Andreea [1 ]
Ravi, Gummadi [2 ]
Kandasamy, Madhusudanan [2 ]
Pozidis, Haralampos [1 ]
机构
[1] IBM Res, Zurich, Switzerland
[2] IBM Syst, Bangalore, Karnataka, India
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We describe a new software framework for fast training of generalized linear models. The framework, named Snap Machine Learning (Snap ML), combines recent advances in machine learning systems and algorithms in a nested manner to reflect the hierarchical architecture of modern computing systems. We prove theoretically that such a hierarchical system can accelerate training in distributed environments where intra-node communication is cheaper than inter-node communication. Additionally, we provide a review of the implementation of Snap ML in terms of GPU acceleration, pipelining, communication patterns and software architecture, highlighting aspects that were critical for achieving high performance. We evaluate the performance of Snap ML in both single-node and multi-node environments, quantifying the benefit of the hierarchical scheme and the data streaming functionality, and comparing with other widely-used machine learning software frameworks. Finally, we present a logistic regression benchmark on the Criteo Terabyte Click Logs dataset and show that Snap ML achieves the same test loss an order of magnitude faster than any of the previously reported results, including those obtained using TensorFlow and scikit-learn.
引用
收藏
页数:11
相关论文
共 50 条
  • [1] ML-Plan: Automated machine learning via hierarchical planning
    Felix Mohr
    Marcel Wever
    Eyke Hüllermeier
    Machine Learning, 2018, 107 : 1495 - 1515
  • [2] ML-Plan: Automated machine learning via hierarchical planning
    Mohr, Felix
    Wever, Marcel
    Huellermeier, Eyke
    MACHINE LEARNING, 2018, 107 (8-10) : 1495 - 1515
  • [3] SNAP: A Communication Efficient Distributed Machine Learning Framework for Edge Computing
    Zhao, Yangming
    Fan, Jingyuan
    Su, Lu
    Song, Tongyu
    Wang, Sheng
    Qiao, Chunming
    2020 IEEE 40TH INTERNATIONAL CONFERENCE ON DISTRIBUTED COMPUTING SYSTEMS (ICDCS), 2020, : 584 - 594
  • [4] Introduction to Snap Machine Learning
    Parnell, Thomas
    2018 IEEE INTERNATIONAL PARALLEL AND DISTRIBUTED PROCESSING SYMPOSIUM WORKSHOPS (IPDPSW 2018), 2018, : 856 - 856
  • [5] ML-FEED: Machine Learning Framework for Efficient Exploit Detection
    Saha, Tanujay
    Al Rahat, Tamjid
    Aaraj, Najwa
    Tian, Yuan
    Jha, Niraj K.
    2022 IEEE 4TH INTERNATIONAL CONFERENCE ON TRUST, PRIVACY AND SECURITY IN INTELLIGENT SYSTEMS, AND APPLICATIONS, TPS-ISA, 2022, : 140 - 149
  • [6] A HIERARCHICAL MACHINE LEARNING FRAMEWORK FOR THE IDENTIFICATION OF AUTOMATED CONSTRUCTION OPERATIONS
    Harichandran, Aparna
    Raphael, Benny
    Mukherjee, Abhijit
    JOURNAL OF INFORMATION TECHNOLOGY IN CONSTRUCTION, 2021, 26 : 591 - 623
  • [7] POSTER: ML-Compass: A Comprehensive Assessment Framework for Machine Learning Models
    Jin, Zhibo
    Zhu, Zhiyu
    Hu, Hongsheng
    Xue, Minhui
    Chen, Huaming
    PROCEEDINGS OF THE 2023 ACM ASIA CONFERENCE ON COMPUTER AND COMMUNICATIONS SECURITY, ASIA CCS 2023, 2023, : 1031 - 1033
  • [8] QoA4ML-A Framework for Supporting Contracts in Machine Learning Services
    Truong, Hong-Linh
    Nguyen, Tri-Minh
    2021 IEEE INTERNATIONAL CONFERENCE ON WEB SERVICES, ICWS 2021, 2021, : 465 - 475
  • [9] Classification Framework for Healthy Hairs and Alopecia Areata: A Machine Learning (ML) Approach
    Shakeel, Choudhary Sobhan
    Khan, Saad Jawaid
    Chaudhry, Beenish
    Aijaz, Syeda Fatima
    Hassan, Umer
    COMPUTATIONAL AND MATHEMATICAL METHODS IN MEDICINE, 2021, 2021
  • [10] A Hierarchical Framework for Drone Identification based on Radio Frequency Machine Learning
    Zhao, Xinrui
    Wang, Longhui
    Wang, Qiexiang
    Wang, Jian
    2022 IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS WORKSHOPS (ICC WORKSHOPS), 2022, : 391 - 396