Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2405.09886 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916249147015168 |
|---|---|
| author | Cao, Han Rajan, Sivanesan Hahn, Bianka Kocak, Ersoy Durstewitz, Daniel Schwarz, Emanuel Schneider-Lindner, Verena |
| author_facet | Cao, Han Rajan, Sivanesan Hahn, Bianka Kocak, Ersoy Durstewitz, Daniel Schwarz, Emanuel Schneider-Lindner, Verena |
| contents | Multi-task learning (MTL) is a learning paradigm that enables the simultaneous training of multiple communicating algorithms. Although MTL has been successfully applied to ether regression or classification tasks alone, incorporating mixed types of tasks into a unified MTL framework remains challenging, primarily due to variations in the magnitudes of losses associated with different tasks. This challenge, particularly evident in MTL applications with joint feature selection, often results in biased selections. To overcome this obstacle, we propose a provable loss weighting scheme that analytically determines the optimal weights for balancing regression and classification tasks. This scheme significantly mitigates the otherwise biased feature selection. Building upon this scheme, we introduce MTLComb, an MTL algorithm and software package encompassing optimization procedures, training protocols, and hyperparameter estimation procedures. MTLComb is designed for learning shared predictors among tasks of mixed types. To showcase the efficacy of MTLComb, we conduct tests on both simulated data and biomedical studies pertaining to sepsis and schizophrenia. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_09886 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | MTLComb: multi-task learning combining regression and classification tasks for joint feature selection Cao, Han Rajan, Sivanesan Hahn, Bianka Kocak, Ersoy Durstewitz, Daniel Schwarz, Emanuel Schneider-Lindner, Verena Machine Learning Artificial Intelligence Biomolecules J.3; I.2.6 Multi-task learning (MTL) is a learning paradigm that enables the simultaneous training of multiple communicating algorithms. Although MTL has been successfully applied to ether regression or classification tasks alone, incorporating mixed types of tasks into a unified MTL framework remains challenging, primarily due to variations in the magnitudes of losses associated with different tasks. This challenge, particularly evident in MTL applications with joint feature selection, often results in biased selections. To overcome this obstacle, we propose a provable loss weighting scheme that analytically determines the optimal weights for balancing regression and classification tasks. This scheme significantly mitigates the otherwise biased feature selection. Building upon this scheme, we introduce MTLComb, an MTL algorithm and software package encompassing optimization procedures, training protocols, and hyperparameter estimation procedures. MTLComb is designed for learning shared predictors among tasks of mixed types. To showcase the efficacy of MTLComb, we conduct tests on both simulated data and biomedical studies pertaining to sepsis and schizophrenia. |
| title | MTLComb: multi-task learning combining regression and classification tasks for joint feature selection |
| topic | Machine Learning Artificial Intelligence Biomolecules J.3; I.2.6 |
| url | https://arxiv.org/abs/2405.09886 |