Presentation | 2015-03-05 Detection of cheating in Boltzmann machine learning : A parameter-free algorithm to sparse solution Masayuki OHZEKI, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | We generalize a mathematical model in the item response theory into that in the Boltzmann machine learning to detect "cheating students". The cheating students are hopefully expected to be rare under normal condition holding tests. (We strongly hope!) In other words, the cheating students are expected to be sparse. In the present study, we employ a greedy algorithm, the decimation algorithm, which is free from the arbitrary coefficient. We confirmed that, when the sparseness is strong, the decimation algorithm outperforms the L1 regularization and establish a basic technique to detect the cheating students. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Boltzmann machine learning / item response theory / sparseness / greedy algorithm |
Paper # | IBISML2014-85 |
Date of Issue |
Conference Information | |
Committee | IBISML |
---|---|
Conference Date | 2015/2/26(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Information-Based Induction Sciences and Machine Learning (IBISML) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Detection of cheating in Boltzmann machine learning : A parameter-free algorithm to sparse solution |
Sub Title (in English) | |
Keyword(1) | Boltzmann machine learning |
Keyword(2) | item response theory |
Keyword(3) | sparseness |
Keyword(4) | greedy algorithm |
1st Author's Name | Masayuki OHZEKI |
1st Author's Affiliation | Department of Systems Science, Kyoto University() |
Date | 2015-03-05 |
Paper # | IBISML2014-85 |
Volume (vol) | vol.114 |
Number (no) | 502 |
Page | pp.pp.- |
#Pages | 8 |
Date of Issue |