Past Recording
ShareSave
Language Models are Unsupervised Multitask Learners
Thursday Mar 7 2019 23:30 GMT
Please to join the live chat.
Language Models are Unsupervised Multitask Learners
Why This Is Interesting

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

Time of Recording: Thursday Mar 7 2019 23:30 GMT