A
Alexander H. Miller
Researcher at Facebook
Publications - 27
Citations - 4481
Alexander H. Miller is an academic researcher from Facebook. The author has contributed to research in topics: Reinforcement learning & Question answering. The author has an hindex of 19, co-authored 27 publications receiving 2826 citations.
Papers
More filters
Posted Content
Language Models as Knowledge Bases
Fabio Petroni,Tim Rocktäschel,Patrick S. H. Lewis,Anton Bakhtin,Yuxiang Wu,Alexander H. Miller,Sebastian Riedel +6 more
TL;DR: An in-depth analysis of the relational knowledge already present (without fine-tuning) in a wide range of state-of-the-art pretrained language models finds that BERT contains relational knowledge competitive with traditional NLP methods that have some access to oracle knowledge.
Proceedings ArticleDOI
Key-Value Memory Networks for Directly Reading Documents
TL;DR: In this paper, a key-value memory network is proposed to make reading documents more viable by utilizing different encodings in the addressing and output stages of the memory read operation.
Proceedings ArticleDOI
Language Models as Knowledge Bases
Fabio Petroni,Tim Rocktäschel,Patrick S. H. Lewis,Anton Bakhtin,Yuxiang Wu,Alexander H. Miller,Sebastian Riedel +6 more
TL;DR: This article presented an in-depth analysis of the relational knowledge already present (without fine-tuning) in a wide range of state-of-the-art pretrained language models.
Book ChapterDOI
The Second Conversational Intelligence Challenge (ConvAI2)
Emily Dinan,Varvara Logacheva,Valentin Malykh,Alexander H. Miller,Kurt Shuster,Jack Urbanek,Douwe Kiela,Arthur Szlam,Iulian Vlad Serban,Ryan Lowe,Ryan Lowe,Shrimai Prabhumoye,Alan W. Black,Alexander I. Rudnicky,Jason D. Williams,Joelle Pineau,Joelle Pineau,Mikhail S. Burtsev,Jason Weston +18 more
TL;DR: To improve performance on multi-turn conversations with humans, future systems must go beyond single word metrics like perplexity to measure the performance across sequences of utterances (conversations)—in terms of repetition, consistency and balance of dialogue acts.
Posted Content
Key-Value Memory Networks for Directly Reading Documents
TL;DR: This work introduces a new method, Key-Value Memory Networks, that makes reading documents more viable by utilizing different encodings in the addressing and output stages of the memory read operation.