Timezone: »

Unsupervised Cross-Task Generalization via Retrieval Augmentation
Bill Yuchen Lin · Kangmin Tan · Chris Miller · Beiwen Tian · Xiang Ren

Tue Nov 29 02:00 PM -- 04:00 PM (PST) @ Hall J #913

Humans can perform unseen tasks by recalling relevant skills acquired previously and then generalizing them to the target tasks, even if there is no supervision at all. In this paper, we aim to improve this kind of cross-task generalization ability of massive multi-task language models, such as T0 and FLAN, in an unsupervised setting. We propose a retrieval-augmentation method named ReCross that takes a few unlabelled examples as queries to retrieve a small subset of upstream data and uses them to update the multi-task model for better generalization. ReCross is a straightforward yet effective retrieval method that combines both efficient dense retrieval and effective pair-wise reranking. Our results and analysis show that it significantly outperforms both non-retrieval methods and other baseline methods.

Author Information

Bill Yuchen Lin (University of Southern California)
Kangmin Tan (University of Southern California)
Chris Miller
Beiwen Tian (Tsinghua University, Tsinghua University)
Xiang Ren (University of Southern California)

More from the Same Authors