bibtype C - Conference Paper (international conference)
ARLID 0642813
utime 20251209092511.8
mtime 20251209235959.9
SCOPUS 105018304037
DOI 10.1007/978-3-032-05134-9_8
title (primary) (eng) From RBMs to BN2A Models: Parameter Transformation for Interpretable Educational Diagnostics
specification
page_count 14 s.
media_type P
serial
ARLID cav_un_epca*0639225
ISBN 978-3-032-05133-2
ISSN 0302-9743
title Symbolic and Quantitative Approaches to Reasoning with Uncertainty. ECSQARU 2025 Proceedings
page_num 104-117
publisher
place Cham
name Springer
year 2026
editor
name1 Sauerwald
name2 K.
editor
name1 Thimm
name2 M.
keyword Restricted Boltzmann Machines
keyword BN2A models
keyword Educational Assessment
keyword Interpretable machine learning
author (primary)
ARLID cav_un_auth*0458433
name1 Pérez Cabrera
name2 Iván
institution UTIA-B
full_dept (cz) Zpracování signálů
full_dept (eng) Department of Signal Processing
department (cz) ZS
department (eng) ZS
country MX
fullinstit Ústav teorie informace a automatizace AV ČR, v. v. i.
author
ARLID cav_un_auth*0101228
name1 Vomlel
name2 Jiří
institution UTIA-B
full_dept (cz) Matematická teorie rozhodování
full_dept Department of Decision Making Theory
department (cz) MTR
department MTR
full_dept Department of Decision Making Theory
fullinstit Ústav teorie informace a automatizace AV ČR, v. v. i.
author
ARLID cav_un_auth*0226898
name1 Martinková
name2 P.
country CZ
source
url https://library.utia.cas.cz/separaty/2025/ZS/perez-0642813.pdf
cas_special
project
project_id GA25-18070S
agency GA ČR
country CZ
ARLID cav_un_auth*0484465
abstract (eng) Restricted Boltzmann Machines (RBMs) are bipartite graphical models with binary latent and observed variables that have shown promise for representation learning. However, their lack of interpretable parameters limits their utility in domains requiring explainability, like educational assessment. Despite extensive RBM research, non-negativity constraints on weights—essential for monotonicity in educational contexts—remain largely unexplored. To address this, we propose a method to translate RBMs into a specialized class of bipartite Bayesian networks, which we term BN2A networks, characterized by strict 2-layer separation (hidden and observed variables), Noisy-AND conditional probability tables, and directly interpretable parameters for educational models. Our work establishes a mathematical transformation from RBM weights to BN2A’s interpretable parameters (leak and penalty probabilities), theoretical analysis showing BN2A’s constrained connectivity is a subset of RBM architectures, and empirical evidence that the transformation preserves model fidelity under realistic conditions. By bridging these paradigms, our method leverages RBM’s representational power while achieving BN2A’s interpretability, opening new possibilities for adaptive learning systems and diagnostic tools.
action
ARLID cav_un_auth*0498692
name European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty, ECSQARU 2025 /19./
dates 20250923
mrcbC20-s 20250926
place Hagen
country DE
RIV IN
FORD0 10000
FORD1 10200
FORD2 10201
reportyear 2026
presentation_type PR
inst_support RVO:67985556
permalink https://hdl.handle.net/11104/0372668
confidential S
arlyear 2026
mrcbU14 105018304037 SCOPUS
mrcbU24 PUBMED
mrcbU34 WOS
mrcbU63 cav_un_epca*0639225 Symbolic and Quantitative Approaches to Reasoning with Uncertainty. ECSQARU 2025 Proceedings Springer 2026 Cham 104 117 978-3-032-05133-2 Lecture Notes in Artificial Intelligence 16099 0302-9743
mrcbU67 Sauerwald K. 340
mrcbU67 Thimm M. 340