Manufacturing Test Sequence Prediction for Early Failure Screening
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional machine-learning techniques, such as neural networks, struggle to predict manufacturing test failures in product testing sequences due to the lack of structure in test sequences and the temporal dependency of test results, leading to inefficiencies in resource allocation and time consumption.
Innovation Solution
A machine-learning-based model is developed to predict failure probabilities of future product tests by analyzing past test outcomes, allowing for early termination or optimization of test sequences, using a computational implementation that trains on sequences of test results to improve statistical confidence and adapt testing procedures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional machine-learning techniques are used to detect production defects, then defect detection capability is improved, but the ability to predict future test failures is insufficient
Solution Approach 1:
The patent applies preliminary action by training the machine learning model on historical test result sequences before actual testing occurs. The model learns temporal patterns and dependencies from past data, enabling it to predict future test failures before they happen. This allows the system to proactively identify units likely to fail upcoming tests, rather than merely detecting defects after they manifest.
2Reliability
If all manufacturing tests are performed to ensure quality, then product quality is improved, but time and resource consumption increase
Solution Approach 1:
The patent applies partial action by performing only a subset of manufacturing tests on each unit, determined by the ML model's predictions. Instead of executing all possible tests uniformly, the system selectively performs tests based on predicted failure probabilities. Units with low predicted failure rates may skip certain tests, reducing overall testing time while maintaining quality assurance for units at risk.
Solution Approach 2:
The system applies self-service by using historical test data to automatically train and improve its own prediction capabilities. The ML model continuously learns from accumulated test results, refining its ability to predict failures and optimize test sequences without external intervention. This self-improving mechanism enhances efficiency over time while maintaining quality standards.
3Ease of manufacture
If fixed test sequences are used for manufacturing testing, then testing procedure simplicity is improved, but adaptability to individual unit characteristics is reduced
Solution Approach 1:
The patent applies dynamics by transforming fixed test sequences into adaptive, dynamic sequences. The ML model analyzes each unit's specific characteristics and historical performance data to dynamically adjust which tests are performed and in what order. This allows the testing procedure to adapt to individual unit characteristics while maintaining operational simplicity through automated decision-making based on learned patterns.
4Loss of information
If temporal dependencies in test sequences are emphasized, then sequence order importance is improved, but model complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the test sequence into distinct temporal segments or phases. Instead of treating the entire sequence as a single complex temporal dependency, the model processes different segments separately, identifying patterns within each phase. This segmentation reduces overall model complexity while still capturing essential temporal relationships that inform failure predictions.
Data Source
AI summary
Example embodiments are disclosed of systems and methods for predicting failure probabilities of future product tests of a testing sequence based on outcomes of prior tests. Predictions are made by a machine-learning-based model (MLM) trained with a set of test-result sequence records (TRSRs) including test values and pass/fail indicators (PRIs) of completed tests. Within training epochs over the set, iterations are carried out over each TRSR. Each iteration involves sub-iterations carried out successively over test results of the TRSR. Each sub-iteration involves (i) inputting to the MLM values of a given test and those of tests earlier in the sequence while masking those later in the sequence, (ii) computing probabilities of test failures for the masked tests found later in the sequence than the given test, and (iii) applying the PFIs of test results later in the sequence than the given test as ground-truths to update parameters of the MLM.