Title:
機械学習モデルを改良して局所性を改善させること
Document Type and Number:
Japanese Patent JP7078758
Kind Code:
B2
Abstract:
Methods, systems, and apparatus for updating machine learning models to improve locality are described. In one aspect, a method includes receiving data of a machine learning model. The data represents operations of the machine learning model and data dependencies between the operations. Data specifying characteristics of a memory hierarchy for a machine learning processor on which the machine learning model is going to be deployed is received. The memory hierarchy includes multiple memories at multiple memory levels for storing machine learning data used by the machine learning processor when performing machine learning computations using the machine learning model. An updated machine learning model is generated by modifying the operations and control dependencies of the machine learning model to account for the characteristics of the memory hierarchy. Machine learning computations are performed using the updated machine learning model.
Inventors:
Yoon, Do Hyun
Patil, Nishant
Jopee, Norman Paul
Patil, Nishant
Jopee, Norman Paul
Application Number:
JP2020566754A
Publication Date:
May 31, 2022
Filing Date:
July 29, 2019
Export Citation:
Assignee:
Google LLC
International Classes:
G06N3/10; G06F12/0811; G06F12/0862
Foreign References:
| WO2017075346A1 | ||||
| US10019668 | ||||
| EP3376441A1 |
Other References:
米尾 謙史(外2名),「Apache Sparkにおけるデータ依存グラフに基づくメモリ内キャッシュの指示および置換」,情報処理学会研究報告,日本,[online], 情報処理学会,2018年02月20日,Vol.2018-OS-142, No.15,第1~9頁,[平成30年2月20日検索],インターネット,
Tianqi Chen, et al.,"TVM: An Automated End-to-End Optimizing Compiler for Deep Learning",arXiv:1802.04799v3,version v3,[online], arXiv (Cornell University),2018年10月05日,Pages 1-16,[令和4年4月10日検索], インターネット,.
Zhi Qiao, et al.,"MR-Graph: a Customizable GPU MapReduce",Proceedings of 2015 IEEE 2nd International Conference on Cyber Security and Cloud Computing,2015年11月05日,Pages 417-422,ISBN: 978-1-4673-9300-3,.
Tianqi Chen, et al.,"TVM: An Automated End-to-End Optimizing Compiler for Deep Learning",arXiv:1802.04799v3,version v3,[online], arXiv (Cornell University),2018年10月05日,Pages 1-16,[令和4年4月10日検索], インターネット,
Zhi Qiao, et al.,"MR-Graph: a Customizable GPU MapReduce",Proceedings of 2015 IEEE 2nd International Conference on Cyber Security and Cloud Computing,2015年11月05日,Pages 417-422,ISBN: 978-1-4673-9300-3,
Attorney, Agent or Firm:
Fukami patent office