Testing-based Black-box Extraction of Simple Models from RNNs and Transformers

Edi Muškardin, Martin Tappler, Bernhard K. Aichernig
Proceedings of 16th edition of the International Conference on Grammatical Inference, PMLR 217:291-294, 2023.

Abstract

In this technical report, we outline the testing-based black-box method used to extract simple and interpretable models from RNNs and transformers. Our work was done in the scope of the TAYSIR competition, in which it won the first place.

Cite this Paper

BibTeX
@InProceedings{pmlr-v217-muskardin23a, title = {Testing-based Black-box Extraction of Simple Models from RNNs and Transformers}, author = {Mu\v{s}kardin, Edi and Tappler, Martin and K. Aichernig, Bernhard}, booktitle = {Proceedings of 16th edition of the International Conference on Grammatical Inference}, pages = {291--294}, year = {2023}, editor = {Coste, François and Ouardi, Faissal and Rabusseau, Guillaume}, volume = {217}, series = {Proceedings of Machine Learning Research}, month = {10--13 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v217/muskardin23a/muskardin23a.pdf}, url = {https://proceedings.mlr.press/v217/muskardin23a.html}, abstract = {In this technical report, we outline the testing-based black-box method used to extract simple and interpretable models from RNNs and transformers. Our work was done in the scope of the TAYSIR competition, in which it won the first place.} }
Endnote
%0 Conference Paper %T Testing-based Black-box Extraction of Simple Models from RNNs and Transformers %A Edi Muškardin %A Martin Tappler %A Bernhard K. Aichernig %B Proceedings of 16th edition of the International Conference on Grammatical Inference %C Proceedings of Machine Learning Research %D 2023 %E François Coste %E Faissal Ouardi %E Guillaume Rabusseau %F pmlr-v217-muskardin23a %I PMLR %P 291--294 %U https://proceedings.mlr.press/v217/muskardin23a.html %V 217 %X In this technical report, we outline the testing-based black-box method used to extract simple and interpretable models from RNNs and transformers. Our work was done in the scope of the TAYSIR competition, in which it won the first place.
APA
Muškardin, E., Tappler, M. & K. Aichernig, B.. (2023). Testing-based Black-box Extraction of Simple Models from RNNs and Transformers. Proceedings of 16th edition of the International Conference on Grammatical Inference, in Proceedings of Machine Learning Research 217:291-294 Available from https://proceedings.mlr.press/v217/muskardin23a.html.

Related Material