Computer Science > Computation and Language
[Submitted on 31 Dec 2020 (this version), latest version 2 Jun 2021 (v2)]
Title:A Closer Look at Few-Shot Crosslingual Transfer: Variance, Benchmarks and Baselines
View PDFAbstract:We present a focused study of few-shot crosslingual transfer, a recently proposed NLP scenario: a pretrained multilingual encoder is first finetuned on many annotations in a high resource language (typically English), and then finetuned on a few annotations (the ``few shots'') in a target language. Few-shot transfer brings large improvements over zero-shot transfer. However, we show that it inherently has large variance and it is necessary to report results on multiple sets of few shots for stable results and to guarantee fair comparison of different algorithms. To address this problem, we publish our few-shot sets. In a study of why few-shot learning outperforms zero-shot transfer, we show that large models heavily rely on lexical hints when finetuned on a few shots and then overfit quickly. We evaluate different methods that use few-shot annotations, but do not observe significant improvements over the baseline. This calls for better ways of utilizing the few-shot annotations.
Submission history
From: Mengjie Zhao [view email][v1] Thu, 31 Dec 2020 16:03:48 UTC (102 KB)
[v2] Wed, 2 Jun 2021 12:34:52 UTC (194 KB)
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.