×
Updating only these task-specific modules then allows the model to be adapted to low-data tasks for as many steps as necessary without risking overfitting.
In particular, we develop general techniques based on Bayesian shrinkage to automatically discover and learn both task-specific and general reusable modules.
Sep 12, 2019 · We develop general techniques based on Bayesian shrinkage to automatically discover and learn both task-specific and general reusable modules.
The goal of meta-learning is to extract common knowledge from a set of training tasks in order to solve held-out tasks more efficiently and accurately. One ...
In this work, we develop techniques based on Bayesian shrinkage to meta-learn how task-independent each module is and to regularize it accordingly. We show that ...
Dec 6, 2020 · In particular, we develop general techniques based on Bayesian shrinkage to automatically discover and learn both task-specific and general ...
This work develops general techniques based on Bayesian shrinkage to automatically discover and learn both task-specific and general reusable modules and ...
Dec 6, 2020 · Neural Information Processing Systems (NeurIPS) is a multi-track machine learning and computational neuroscience conference that includes ...
Sep 9, 2024 · To address this, we introduce a hierarchical Bayesian model with per-module shrinkage parameters, which we propose to learn by maximizing an ...
People also ask
Modular Meta-Learning with Shrinkage · Yutian Chen, Abe Friesen, Feryal ... Meta Learning/Software. Back to Top. Virtual NeurIPS 2020 made with MiniConf. We ...