A training method and training system for a machine learning system
A technology of machine learning and training methods, applied in the field of big data processing, can solve problems such as large influence of model noise, model training stuck, training failure, etc., to enhance the ability to resist data noise, improve service capabilities, and ensure normal output Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
no. 1 example
[0026] The first embodiment of the present application proposes a training method for a machine learning system, such as figure 1 Shown is a flow chart of the training method of the machine learning system according to the first embodiment of the present application. The machine learning system is preferably a distributed machine learning system, including a parameter server (parameter server). The parameter server may include, for example, multiple workers (workers or slavers), multiple servers (servers), and a coordinator (coordinator). like figure 1 As shown, the training method includes the following steps:
[0027] Step S101, distributing the training data to multiple working machines;
[0028] In this step, for example, each working machine can read its own training data according to its identification, and the data among the working machines do not overlap. In this step, for example, the coordinator may divide the training data into training data belonging to each wo...
no. 2 example
[0045] The second embodiment of the present application proposes a training method for a machine learning system, such as figure 2 Shown is a flow chart of the training method of the machine learning system according to the second embodiment of the present application. The machine learning system is preferably a distributed machine learning system, such as figure 2 As shown, the training method includes the following steps:
[0046] S201, distribute the training data to multiple working machines;
[0047] S202, divide the training data obtained by each working machine into multiple data slices;
[0048] S203. Obtain the local weight and local loss function value calculated by each working machine based on each data piece;
[0049] S204, summarizing the local weights and local loss function values calculated by each working machine based on each piece of data, to obtain the current weight and the current loss function value;
[0050] S205, perform model anomaly detectio...
no. 3 example
[0073] The third embodiment of the present application proposes a training method for a machine learning system, such as image 3 Shown is a flowchart of the training method of the machine learning system according to the third embodiment of the present application. The machine learning system is preferably a distributed machine learning system, such as image 3 As shown, the training method includes the following steps:
[0074] S301. Distributing the training data to multiple working machines;
[0075] S302, divide the training data obtained by each working machine into multiple data slices;
[0076] S303. Obtain the local weight and local loss function value calculated by each working machine based on each data slice;
[0077] S304, summarizing the local weights and local loss function values calculated by each working machine based on each piece of data, to obtain the current weight and the current loss function value;
[0078] S305, perform model anomaly detection b...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap