+ IBF (Idea-breeding-farm) +

We are proud of this study! close

+ MY-X DVD +

BME-ETDK-1. place (incl. OTDK-access-right) with a study supported by the MY-X team!

2 BPROF-Students (1. semester): 1. and 2. places (ETDK) + OTDK-access rights!

QuILT: distance education with avatars in English!

More and more qualitative studies in English from Students of the international courses!

The Liebig-rule is available in the traffic!

Kecskemét-Conference about Innovation: presentation of the 2DM-game!

MMO 2018 conference: the best presentation of the session (award for 2DM) + the best publication of the session (award for Rosling-animations)!

Last modified: 2015.VII.19.14:08 - MIAÚ-RSS
The Journal MIAU has been working for 20+ years as a kind of public service!
The MATARKA-view

Automated error-detection in the production of semiconductors based on big data from sensor-information

Leading article: 2021. August (MIAU No. 276.)
(Previous article: MIAU No. 275.)

Keywords: chained similarity analyses, optimization, Solver, input-density, output-ratio/ranking, SECOM-database, model-centric and data-centric artificial intelligence

Abstract: The automated production of semiconductors can be observed with quasi unlimited sensors and so, the big-data can be ensured for classification projects (s. SECOM-database). The question is seemingly trivial: how it is possible to derive a classification model with the less error (like false negative and false positive). The ratio of false positive and false negative objects has however no trivial direction: the equivalences are economical question especially in unbalanced cases like in SECOM-database. The bast model can not be defined in a simple way. The anti-discriminative modelling approach makes possible to derive an aggregated goodness index where the theoretical scenario of the massive-unbalanced training sets prescribe in a quasi norm-like situation, that no fails may be estimated only passes. The focused solutions (the staircase functions) are capable to derive a constant output value for the majority class. The challenge is whether this specific estimation value becomes a pure rule, or the test cases makes it dirty. The presented models will involve chains of staircase functions with innovative features like reducing processed pixels, exclusion objects, hybridization of staircase layers and/or types. More (DOC) *** More (PDF)

Please, send Your comments per email!