You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Large pre-trained models are capable of few-shot in-context learning (ICL), i.e., performing a new task by prepending a few demonstrations before the test input. However, the concatenated demonstrations are often excessively long and induce additional computation. Inspired by fusion-in-decoder (FiD) models which efficiently aggregate more passages and thus outperforms concatenation-based models in open-domain QA, we hypothesize that similar techniques can be applied to improve the efficiency and end-task performance of ICL. To verify this, we present a comprehensive study on applying three fusion methods—concatenation-based (early fusion), FiD (intermediate), and ensemble-based (late)—to ICL. We adopt a meta-learning setup where a model is first trained to perform ICL on a mixture of tasks using one selected fusion method, then evaluated on held-out tasks for ICL. Results on 11 held-out tasks show that FiD-ICL matches or outperforms the other two fusion methods. Additionally, we show that FiD-ICL (1) is 10x faster at inference time compared to concat-based and ensemble-based ICL, as we can easily pre-compute the representations of in-context examples and reuse them; (2) enables scaling up to meta-training 3B-sized models, which would fail for concat-based ICL.
AkihikoWatanabe
changed the title
FiD-ICL: A Fusion-in-Decoder Approach for Efficient In-Context Learning
FiD-ICL: A Fusion-in-Decoder Approach for Efficient In-Context Learning, ACL'23
Oct 22, 2023
https://virtual2023.aclweb.org/paper_P4254.html
The text was updated successfully, but these errors were encountered: