Knowledge Distillation with Auxiliary Variable

Bo Peng, Zhen Fang, Guangquan Zhang, Jie Lu
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:40185-40199, 2024.

Abstract

Knowledge distillation (KD) provides an efficient framework for transferring knowledge from a teacher model to a student model by aligning their predictive distributions. The existing KD methods adopt the same strategy as the teacher to formulate the student’s predictive distribution. However, employing the same distribution-modeling strategy typically causes sub-optimal knowledge transfer due to the discrepancy in model capacity between teacher and student models. Designing student-friendly teachers contributes to alleviating the capacity discrepancy, while it requires either complicated or student-specific training schemes. To cast off this dilemma, we propose to introduce an auxiliary variable to promote the ability of the student to model predictive distribution. The auxiliary variable is defined to be related to target variables, which will boost the model prediction. Specifically, we reformulate the predictive distribution with the auxiliary variable, deriving a novel objective function of KD. Theoretically, we provide insights to explain why the proposed objective function can outperform the existing KD methods. Experimentally, we demonstrate that the proposed objective function can considerably and consistently outperform existing KD methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-peng24a, title = {Knowledge Distillation with Auxiliary Variable}, author = {Peng, Bo and Fang, Zhen and Zhang, Guangquan and Lu, Jie}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {40185--40199}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/peng24a/peng24a.pdf}, url = {https://proceedings.mlr.press/v235/peng24a.html}, abstract = {Knowledge distillation (KD) provides an efficient framework for transferring knowledge from a teacher model to a student model by aligning their predictive distributions. The existing KD methods adopt the same strategy as the teacher to formulate the student’s predictive distribution. However, employing the same distribution-modeling strategy typically causes sub-optimal knowledge transfer due to the discrepancy in model capacity between teacher and student models. Designing student-friendly teachers contributes to alleviating the capacity discrepancy, while it requires either complicated or student-specific training schemes. To cast off this dilemma, we propose to introduce an auxiliary variable to promote the ability of the student to model predictive distribution. The auxiliary variable is defined to be related to target variables, which will boost the model prediction. Specifically, we reformulate the predictive distribution with the auxiliary variable, deriving a novel objective function of KD. Theoretically, we provide insights to explain why the proposed objective function can outperform the existing KD methods. Experimentally, we demonstrate that the proposed objective function can considerably and consistently outperform existing KD methods.} }
Endnote
%0 Conference Paper %T Knowledge Distillation with Auxiliary Variable %A Bo Peng %A Zhen Fang %A Guangquan Zhang %A Jie Lu %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-peng24a %I PMLR %P 40185--40199 %U https://proceedings.mlr.press/v235/peng24a.html %V 235 %X Knowledge distillation (KD) provides an efficient framework for transferring knowledge from a teacher model to a student model by aligning their predictive distributions. The existing KD methods adopt the same strategy as the teacher to formulate the student’s predictive distribution. However, employing the same distribution-modeling strategy typically causes sub-optimal knowledge transfer due to the discrepancy in model capacity between teacher and student models. Designing student-friendly teachers contributes to alleviating the capacity discrepancy, while it requires either complicated or student-specific training schemes. To cast off this dilemma, we propose to introduce an auxiliary variable to promote the ability of the student to model predictive distribution. The auxiliary variable is defined to be related to target variables, which will boost the model prediction. Specifically, we reformulate the predictive distribution with the auxiliary variable, deriving a novel objective function of KD. Theoretically, we provide insights to explain why the proposed objective function can outperform the existing KD methods. Experimentally, we demonstrate that the proposed objective function can considerably and consistently outperform existing KD methods.
APA
Peng, B., Fang, Z., Zhang, G. & Lu, J.. (2024). Knowledge Distillation with Auxiliary Variable. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:40185-40199 Available from https://proceedings.mlr.press/v235/peng24a.html.

Related Material