tl4nlp.github.io domain details

Example domain paragraphs

Transfer Learning for NLP Workshop 2022 Call For Papers Program Accepted Works Proceedings Transfer Learning for NLP Workshop 2022 ♦ Workshop co-located with NeurIPS 2022 ♦ December 3, 2022 ♦ ♦ Theater C, New Orleans Convention Center ♦ ♦ Contact: [email protected] ♦ Abstract Transfer learning has become ubiquitous in natural language processing due in part to the ease of access to large pre-trained language models (PLM). Current transfer learning methods, combined with PLMs, have seen out

One particular hope for this workshop is to help to answer the question: Can we characterize the transfer behaviors between source and target tasks/domains/languages in terms of their fundamental properties?

A current weakness of transfer learning is the limited understanding of when transfer will lead to performance improvements, or predicting how positive or negative the effect of transfer will be. Although negative transfer is a known issue in transfer learning, multitask learning, continual learning, and domain adaptation, the causes remain unclear. As research efforts scale up to address an ever expanding set of domains, tasks, and languages, understanding positive and negative transfer becomes increasingl

Links to tl4nlp.github.io (1)