Connectivity Patterns are Task Embeddings

Zhiheng Xi, Rui Zheng, Yuansen Zhang, Xuanjing Huang, Zhongyu Wei, Minlong Peng, Mingming Sun, Qi Zhang, Tao Gui


Abstract
Task embeddings are task-specific vectors designed to construct a semantic space of tasks, which can be used to predict the most transferable source task for a given target task via the similarity between task embeddings. However, existing methods use optimized parameters and representations as task embeddings, resulting in substantial computational complexity and storage requirements. In this work, we draw inspiration from the operating mechanism of deep neural networks (DNNs) and biological brains, where neuronal activations are sparse and task-specific, and we use the connectivity patterns of neurons as a unique identifier associated with the task. The proposed method learns to assign importance masks for sub-structures of DNNs, and accordingly indicate the task-specific connectivity patterns. In addition to the storage advantages brought by the binary masking mechanism and structured sparsity, the early-bird nature of the sparse optimization process can deliver an efficient computation advantage. Experiments show that our method consistently outperforms other baselines in predicting inter-task transferability across data regimes and transfer settings, while keeping high efficiency in computation and storage.
Anthology ID:
2023.findings-acl.759
Volume:
Findings of the Association for Computational Linguistics: ACL 2023
Month:
July
Year:
2023
Address:
Toronto, Canada
Editors:
Anna Rogers, Jordan Boyd-Graber, Naoaki Okazaki
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
11993–12013
Language:
URL:
https://aclanthology.org/2023.findings-acl.759
DOI:
10.18653/v1/2023.findings-acl.759
Bibkey:
Cite (ACL):
Zhiheng Xi, Rui Zheng, Yuansen Zhang, Xuanjing Huang, Zhongyu Wei, Minlong Peng, Mingming Sun, Qi Zhang, and Tao Gui. 2023. Connectivity Patterns are Task Embeddings. In Findings of the Association for Computational Linguistics: ACL 2023, pages 11993–12013, Toronto, Canada. Association for Computational Linguistics.
Cite (Informal):
Connectivity Patterns are Task Embeddings (Xi et al., Findings 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.findings-acl.759.pdf
Video:
 https://aclanthology.org/2023.findings-acl.759.mp4