Joint embedding–classifier learning for interpretable collaborative filtering
Abstract Background Interpretability is a topical question in recommender systems, especially in healthcare applications. An interpretable classifier quantifies the importance of each input feature for the predicted item-user association in a non-ambiguous fashion. Results We introduce the novel Joi...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-01-01
|
| Series: | BMC Bioinformatics |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12859-024-06026-8 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832585340152446976 |
|---|---|
| author | Clémence Réda Jill-Jênn Vie Olaf Wolkenhauer |
| author_facet | Clémence Réda Jill-Jênn Vie Olaf Wolkenhauer |
| author_sort | Clémence Réda |
| collection | DOAJ |
| description | Abstract Background Interpretability is a topical question in recommender systems, especially in healthcare applications. An interpretable classifier quantifies the importance of each input feature for the predicted item-user association in a non-ambiguous fashion. Results We introduce the novel Joint Embedding Learning-classifier for improved Interpretability (JELI). By combining the training of a structured collaborative-filtering classifier and an embedding learning task, JELI predicts new user-item associations based on jointly learned item and user embeddings while providing feature-wise importance scores. Therefore, JELI flexibly allows the introduction of priors on the connections between users, items, and features. In particular, JELI simultaneously (a) learns feature, item, and user embeddings; (b) predicts new item-user associations; (c) provides importance scores for each feature. Moreover, JELI instantiates a generic approach to training recommender systems by encoding generic graph-regularization constraints. Conclusions First, we show that the joint training approach yields a gain in the predictive power of the downstream classifier. Second, JELI can recover feature-association dependencies. Finally, JELI induces a restriction in the number of parameters compared to baselines in synthetic and drug-repurposing data sets. |
| format | Article |
| id | doaj-art-242ed8a415c442199c7427bef25431f8 |
| institution | Kabale University |
| issn | 1471-2105 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Bioinformatics |
| spelling | doaj-art-242ed8a415c442199c7427bef25431f82025-01-26T12:54:52ZengBMCBMC Bioinformatics1471-21052025-01-0126113410.1186/s12859-024-06026-8Joint embedding–classifier learning for interpretable collaborative filteringClémence Réda0Jill-Jênn Vie1Olaf Wolkenhauer2Institute of Computer Science, University of RostockSoda, Inria SaclayInstitute of Computer Science, University of RostockAbstract Background Interpretability is a topical question in recommender systems, especially in healthcare applications. An interpretable classifier quantifies the importance of each input feature for the predicted item-user association in a non-ambiguous fashion. Results We introduce the novel Joint Embedding Learning-classifier for improved Interpretability (JELI). By combining the training of a structured collaborative-filtering classifier and an embedding learning task, JELI predicts new user-item associations based on jointly learned item and user embeddings while providing feature-wise importance scores. Therefore, JELI flexibly allows the introduction of priors on the connections between users, items, and features. In particular, JELI simultaneously (a) learns feature, item, and user embeddings; (b) predicts new item-user associations; (c) provides importance scores for each feature. Moreover, JELI instantiates a generic approach to training recommender systems by encoding generic graph-regularization constraints. Conclusions First, we show that the joint training approach yields a gain in the predictive power of the downstream classifier. Second, JELI can recover feature-association dependencies. Finally, JELI induces a restriction in the number of parameters compared to baselines in synthetic and drug-repurposing data sets.https://doi.org/10.1186/s12859-024-06026-8Drug repurposingInterpretabilityGene expressionCollaborative filtering |
| spellingShingle | Clémence Réda Jill-Jênn Vie Olaf Wolkenhauer Joint embedding–classifier learning for interpretable collaborative filtering BMC Bioinformatics Drug repurposing Interpretability Gene expression Collaborative filtering |
| title | Joint embedding–classifier learning for interpretable collaborative filtering |
| title_full | Joint embedding–classifier learning for interpretable collaborative filtering |
| title_fullStr | Joint embedding–classifier learning for interpretable collaborative filtering |
| title_full_unstemmed | Joint embedding–classifier learning for interpretable collaborative filtering |
| title_short | Joint embedding–classifier learning for interpretable collaborative filtering |
| title_sort | joint embedding classifier learning for interpretable collaborative filtering |
| topic | Drug repurposing Interpretability Gene expression Collaborative filtering |
| url | https://doi.org/10.1186/s12859-024-06026-8 |
| work_keys_str_mv | AT clemencereda jointembeddingclassifierlearningforinterpretablecollaborativefiltering AT jilljennvie jointembeddingclassifierlearningforinterpretablecollaborativefiltering AT olafwolkenhauer jointembeddingclassifierlearningforinterpretablecollaborativefiltering |