LSTM-based intelligent generation method for dispensing graph

A graphics and dispensing technology, applied in neural learning methods, biological neural network models, instruments, etc., to save material costs, reduce training parameters, and shorten the test cycle.

Pending Publication Date: 2021-03-12
BEIJING RES INST OF TELEMETRY +1
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The present invention aims to solve the problem that the artificial intelligence algorithm cannot be introduced into the parameter decisio

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • LSTM-based intelligent generation method for dispensing graph
  • LSTM-based intelligent generation method for dispensing graph
  • LSTM-based intelligent generation method for dispensing graph

Examples

Experimental program
Comparison scheme
Effect test

Example Embodiment

[0038] Example 1

[0039] Such as figure 1 As shown, an LSTM-based dispensing graphic intelligent generation method includes the following steps:

[0040] S1, data preparation: collect input feature and output feature of automatic mount data form a data set, divide the data set as training set, verification set and test set;

[0041] S2, establish an input feature property. Vector: Prepare the input feature to get the input feature property;

[0042] S3, establish an output feature property vector: Preprocessing the output feature to obtain the output feature property;

[0043] S4, establish a LSTM model: Figure 2-3 The LSTM structure and super parameter are designed to establish a LSTM model;

[0044] S5, training LSTM model: Design LSTM model loss function and optimization algorithm, using input feature property vectors and output feature attributes, to correct ultra-parameters until the end of the end of the final LSTM model;

[0045] S6, generating dot clamping pattern: Call ...

Example Embodiment

[0046] Example 2

[0047] Such as figure 1 As shown, an LSTM-based dispensing graphic intelligent generation method includes the following steps:

[0048] S1, data preparation: Collect the input feature and output feature of automatic mount data to form a data set, divide the data set as training set, verification set and test set; input features include: numerical features, order characteristics, and nominal features; output The characteristics are the coordinate sequence of displaced graphics; the total number of samples of the data set is not less than 1000 groups, the proportion of training sets, verification sets, and test sets is: 70%, 20%, 10%;

[0049] S2, establish an input characteristic property: Figure 4 As shown, the input feature is pre-processed to obtain an input feature attribute vector;

[0050] S21, define a numerical feature, and normalize the numerical feature, obtain a numerical feature vector; the numerical feature is a chip size and a needle size;

[0051] ...

Example Embodiment

[0061] Example 3

[0062] Such as figure 1 As shown, for a set of existing automatic mounting process parameters and corresponding dispensing patch effects (given according to GJB548B evaluation criteria), the data set is 1000, design a method of LSTM-based dispensing graphics intelligent generation method. , Including the following steps:

[0063] S1, data preparation: Collect the input feature and output feature of automatic mount data to form a data set, divide the data set as training set, verification set and test set; input features include: numerical features, order characteristics, and nominal features; output The characteristics are the coordinate sequence of displaced graphics; the total number of samples of the data set is not less than 1000 groups, the proportion of training sets, verification sets, and test sets is: 70%, 20%, 10%;

[0064] S2, establish an input characteristic property: Figure 4 As shown, the input feature is pre-processed to obtain an input feature...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides an LSTM-based intelligent generation method for a dispensing graph, which comprises the steps of collecting input features and output features of automatic mounting data to forma data set, preprocessing the input features to obtain an input feature attribute vector, preprocessing the output features to obtain an output feature attribute vector, designing an LSTM structure and hyper-parameters to establish an LSTM model; designing a loss function and an optimization algorithm of the LSTM model, performing LSTM model training by using the input feature attribute vector and the output feature attribute vector respectively, correcting hyper-parameters until the training is finished to obtain a final LSTM model, and calling the final LSTM model to generate a dispensing graph. The invention provides the intelligent generation method for the dispensing graph in the field of electronic packaging, and the method combines a long-short-term memory unit LSTM in a recurrentneural network RNN in the field of artificial intelligence with a dispensing graph generation process, thereby providing a feasible scheme for the generation of parameters with a time sequence. Therefore, the problem of time information loss caused by the fact that dispensing pattern selection is carried out only through a process test and a feedforward neural network is used for generating a pattern is solved.

Description

technical field [0001] The invention relates to the technical field of semiconductor devices, in particular to an LSTM-based intelligent generation method for dispensing graphics. Background technique [0002] Automatic placement is a key process that determines the performance and accuracy of electronic packaging products, and the result of automatic placement is determined by multiple parameters of automatic dispensing and placement. The determination of the dispensing pattern affects the effect of automatic placement key step. In the traditional process, the engineer will select the dispensing pattern based on past experience, then conduct a large number of process tests, and continuously adjust the dispensing pattern according to the placement effect. The same chip often needs to be iterated multiple times, resulting in a lot of waste of time and material costs. Engineers The relationship between the inherent characteristics of the chip and the dispensing pattern has no...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F30/27G06N3/04G06N3/08G06F113/18
CPCG06F30/27G06N3/049G06N3/08G06F2113/18G06N3/045
Inventor 杜仲辉刘德喜井津域史磊康楠刘洋景翠
Owner BEIJING RES INST OF TELEMETRY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products