BR112022017493A2 - Método e sistema de treinamento universal para modelos de inteligência artificial - Google Patents

Método e sistema de treinamento universal para modelos de inteligência artificial

Info

Publication number
BR112022017493A2
BR112022017493A2 BR112022017493A BR112022017493A BR112022017493A2 BR 112022017493 A2 BR112022017493 A2 BR 112022017493A2 BR 112022017493 A BR112022017493 A BR 112022017493A BR 112022017493 A BR112022017493 A BR 112022017493A BR 112022017493 A2 BR112022017493 A2 BR 112022017493A2
Authority
BR
Brazil
Prior art keywords
artificial intelligence
training
mirror image
model
intelligence model
Prior art date
Application number
BR112022017493A
Other languages
English (en)
Inventor
Zhou Hao
Original Assignee
Fiberhome Telecommunication Tech Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Fiberhome Telecommunication Tech Co Ltd filed Critical Fiberhome Telecommunication Tech Co Ltd
Publication of BR112022017493A2 publication Critical patent/BR112022017493A2/pt

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/334—Query execution
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Image Analysis (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

MÉTODO E SISTEMA DE TREINAMENTO UNIVERSAL PARA MODELOS DE INTELIGÊNCIA ARTIFICIAL. Um método e sistema de treinamento universal para modelos de inteligência artificial, relacionados com o campo da inteligência artificial. O método compreende: armazenar o código-fonte de um modelo de inteligência artificial a ser treinado e os dados de amostra necessários para o treinamento em um repositório de código de modelo (S1); de acordo com as características de gradiente do código-fonte, combinar, a partir de um repositório de imagens espelhadas, uma imagem espelhada de treinamento correspondente ao modelo de inteligência artificial (S2); de acordo com a imagem espelhada de treinamento combinada, combinar, a partir de um agrupamento de recursos de computação, um nó de computação correspondente à imagem espelhada de treinamento (S3); e começar a treinar o modelo de inteligência artificial no nó de computação combinado e, durante o treinamento, ajustar dinamicamente o nó de computação necessário de acordo com as características do modelo de inteligência artificial para concluir o treinamento do modelo de inteligência artificial (S4). O método descrito pode melhorar a eficiência no treinamento de modelos de inteligência artificial e reduzir efetivamente os custos de treinamento.
BR112022017493A 2020-05-09 2020-08-20 Método e sistema de treinamento universal para modelos de inteligência artificial BR112022017493A2 (pt)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010387700.6A CN113626179B (zh) 2020-05-09 2020-05-09 一种通用的人工智能模型训练方法及系统
PCT/CN2020/110175 WO2021227293A1 (zh) 2020-05-09 2020-08-20 一种通用的人工智能模型训练方法及系统

Publications (1)

Publication Number Publication Date
BR112022017493A2 true BR112022017493A2 (pt) 2022-11-29

Family

ID=78377497

Family Applications (1)

Application Number Title Priority Date Filing Date
BR112022017493A BR112022017493A2 (pt) 2020-05-09 2020-08-20 Método e sistema de treinamento universal para modelos de inteligência artificial

Country Status (3)

Country Link
CN (1) CN113626179B (pt)
BR (1) BR112022017493A2 (pt)
WO (1) WO2021227293A1 (pt)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118012468B (zh) * 2024-04-08 2024-07-09 浙江深象智能科技有限公司 模型处理方法、系统及设备

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2477141B1 (en) * 2011-01-12 2013-07-03 Fujitsu Limited Processor node, artificial neural network and method of operation of an artificial neural network
CN104463324A (zh) * 2014-11-21 2015-03-25 长沙马沙电子科技有限公司 一种基于大规模高性能集群的卷积神经网络并行处理方法
CN107885762B (zh) * 2017-09-19 2021-06-11 北京百度网讯科技有限公司 智能大数据系统、提供智能大数据服务的方法和设备
CN108647785A (zh) * 2018-05-17 2018-10-12 普强信息技术（北京）有限公司 一种神经网络自动建模方法、装置及存储介质
CN109635918A (zh) * 2018-10-30 2019-04-16 银河水滴科技（北京）有限公司 基于云平台和预设模型的神经网络自动训练方法和装置
CN109508238A (zh) * 2019-01-05 2019-03-22 咪付（广西）网络技术有限公司 一种用于深度学习的资源管理系统及方法
CN110413294B (zh) * 2019-08-06 2023-09-12 中国工商银行股份有限公司 服务发布系统、方法、装置和设备
CN111124634A (zh) * 2019-12-06 2020-05-08 广东浪潮大数据研究有限公司 一种训练方法、装置及电子设备和存储介质
CN111026436B (zh) * 2019-12-09 2021-04-02 支付宝(杭州)信息技术有限公司 模型联合训练方法及装置

Also Published As

Publication number Publication date
CN113626179A (zh) 2021-11-09
CN113626179B (zh) 2023-08-22
WO2021227293A1 (zh) 2021-11-18

Similar Documents

Publication Publication Date Title
Gan et al. Towards robustness of text-to-SQL models against synonym substitution
Wang et al. Cross-lingual BERT transformation for zero-shot dependency parsing
Song et al. Mpnet: Masked and permuted pre-training for language understanding
CN107563498B (zh) 基于视觉与语义注意力相结合策略的图像描述方法及系统
Lee et al. Emergent translation in multi-agent communication
Liu et al. Qaner: Prompting question answering models for few-shot named entity recognition
Mohammadshahi et al. SMaLL-100: Introducing shallow multilingual machine translation model for low-resource languages
Stevenson Science, computational science, and computer science: At a crossroads
CN112214608A (zh) 基于知识推理的文本生成方法、介质、装置和计算设备
BR112022017493A2 (pt) Método e sistema de treinamento universal para modelos de inteligência artificial
CN114091466A (zh) 一种基于Transformer和多任务学习的多模态情感分析方法及系统
Chang et al. Time-aware ancient Chinese text translation and inference
Li et al. Semi-supervised domain adaptation for dependency parsing via improved contextualized word representations
Li et al. Explicit connection distillation
Assem et al. QASAR: self-supervised learning framework for extractive question answering
Gao et al. UNISON: Unpaired cross-lingual image captioning
Wang et al. LLMs in the Imaginarium: tool learning through simulated trial and error
Yu et al. Boost transformer-based language models with gpu-friendly sparsity and quantization
Zhu et al. Chain-of-questions training with latent answers for robust multistep question answering
Kim et al. Embeddistill: A geometric knowledge distillation for information retrieval
CN117251555A (zh) 一种语言生成模型训练方法和装置
Dong et al. Understand what llm needs: Dual preference alignment for retrieval-augmented generation
Bowman et al. Collecting entailment data for pretraining: New protocols and negative results
Madaan et al. Improving neural model performance through natural language feedback on their explanations
CN116910217A (zh) 基于小语言模型集群的自然语言问答方法、装置及介质