The Dialogue Dodecathlon: Open-Domain Knowledge and Image Grounded Conversational Agents - ACL Anthology

The Dialogue Dodecathlon: Open-Domain Knowledge and Image Grounded Conversational Agents - ACL Anthology

Kurt Shuster, Da Ju, Stephen Roller, Emily Dinan, Y-Lan Boureau, Jason Weston. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020.

1 mentions: @sei_shinagawa
Date: 2021/02/23 09:51

Referring Tweets

@sei_shinagawa 思い出したけどV&Lにおけるseq2seqのtransformerベースのマルチタスク学習についてはThe Dialogue Dodecthlonで既に議論されてるな。マルチタスクで学習してからシングルでfine-tuningすると性能上がる。この結果から考えても最初に紹介した方の結果は不自然やなと思う t.co/JK3XLAZgCH

Related Entries

Read more [1709.07604] A Comprehensive Survey of Graph Embedding: Problems, Techniques and Applications
0 users, 1 mentions 2018/12/09 05:05
Read more [1706.00130v2] Teaching Machines to Describe Images via Natural Language Feedback
0 users, 1 mentions 2019/03/22 21:57
Read more [1802.08216] ChatPainter: Improving Text to Image Generation using Dialogue
0 users, 1 mentions 2019/04/20 05:16
Read more 2020.11.21 昔の論文読み会 発表資料 - Speaker Deck
0 users, 1 mentions 2020/11/21 09:51
Read more OpenAIのPPO言語生成論文を読む② - Seitaro Shinagawaの雑記帳
0 users, 1 mentions 2020/12/29 23:58