[OpenAI GPT-2] Language Models are Unsupervised Multitask Learners

Time: Thursday 7-Mar-3919 22:30 (This is a past event.)

Motivation
Natural language processing tasks, such as question answering, machine translation, reading comprehension, and summarization, are typically approached with supervised learning on task-specific datasets. We demonstrate that language models begin to learn these tasks without any explicit supervision when trained on a new dataset of millions of webpages called WebText. When conditioned on a document plus questions, the answers generated by the language model reach 55
Subjects:
Stream Categories:
 Trending Paper