2019

How Can We Know What Language Models Know?

Jiang, Zhengbao, Xu, Frank F., Araki, Jun et al.

Understand

Recent work has presented intriguing results examining the knowledge contained in language models (LM) by having the LM fill in the blanks of prompts such as "Obama is a _ by profession".

These prompts are usually manually created, and quite possibly sub-optimal; another prompt such as "Obama worked as a _" may result in more accurately predicting the correct profession.
Because of this, given an inappropriate prompt, we might fail to retrieve facts that the LM does know, and thus any given prompt only provides a lower bound estimate of the knowledge contained in an LM.
In this paper, we attempt to more accurately estimate the knowledge contained in LMs by automatically discovering better prompts to use in this querying process.

Built on

Assessing BERT’s syntactic abilities
Original
Yoav Goldberg. 2019 · 1901
Earlier work this paper cites.
BERT is not a knowledge base (yet): Factual knowledge vs. name-based reasoning in unsupervised QA
Original
Nina Pörner, Ulli Waltinger, and Hinrich Schütze. 2019 · 1911
Earlier work this paper cites.
Snowball : Extracting relations from large plain-text collections
Eugene Agichtein and Luis Gravano. 2000 · 2000
Earlier work this paper cites.
Learning surface text patterns for a question answering system
Deepak Ravichandran and Eduard Hovy. 2002 · 2002
Earlier work this paper cites.
Sentence-level MT evaluation without reference translations: Beyond language modeling
Michael Gamon, Anthony Aue, and Martine Smets. 2005 · 2005
Earlier work this paper cites.
Investigating a generic paraphrase-based approach for relation extraction
Lorenza Romano, Milen Kouylekov, Idan Szpektor, Ido Dagan, and Alberto Lavelli. 2006 · 2006
Earlier work this paper cites.
Open information extraction from the web
Michele Banko, Michael J. Cafarella, Stephen Soderland, Matthew Broadhead, and Oren Etzioni. 2007 · 2007
Earlier work this paper cites.
Large scale acquisition of paraphrases for learning surface patterns
Rahul Bhagat and Deepak Ravichandran. 2008 · 2008
Earlier work this paper cites.
Identifying relations for open information extraction
Anthony Fader, Stephen Soderland, and Oren Etzioni. 2011 · 2011
Earlier work this paper cites.
The Microsoft Research sentence completion challenge
Geoffrey Zweig and Christopher JC Burges. 2011 · 2011
Earlier work this paper cites.
A survey of automatic query expansion in information retrieval
Claudio Carpineto and Giovanni Romano. 2012 · 2012
Earlier work this paper cites.
Context dependent recurrent neural network language model
Tomas Mikolov and Geoffrey Zweig. 2012 · 2012
Earlier work this paper cites.
Semi-supervised sequence learning
Andrew M. Dai and Quoc V. Le. 2015 · 2015
Earlier work this paper cites.
Adam: A method for stochastic optimization
Original
Diederik P. Kingma and Jimmy Ba. 2015 · 2015
Earlier work this paper cites.
Representing text for joint embedding of text and knowledge bases
Kristina Toutanova, Danqi Chen, Patrick Pantel, Hoifung Poon, Pallavi Choudhury, and Michael Gamon. 2015 · 2015
Earlier work this paper cites.
A neural knowledge language model
Original
Sungjin Ahn, Heeyoul Choi, Tanel Pärnamaa, and Yoshua Bengio. 2016 · 2016
Earlier work this paper cites.
A diversity-promoting objective function for neural conversation models
Jiwei Li, Michel Galley, Chris Brockett, Jianfeng Gao, and Bill Dolan. 2016a · 2016
Earlier work this paper cites.
Assessing the ability of LSTMs to learn syntax-sensitive dependencies
Tal Linzen, Emmanuel Dupoux, and Yoav Goldberg. 2016 · 2016
Earlier work this paper cites.

Similar

context2vec: Learning generic context embedding with bidirectional LSTM
Oren Melamud, Jacob Goldberger, and Ido Dagan. 2016 · 2016
Cited alongside, same era.
Improving neural machine translation models with monolingual data
Rico Sennrich, Barry Haddow, and Alexandra Birch. 2016 · 2016
Cited alongside, same era.
Does string-based neural MT learn source syntax?
Xing Shi, Inkit Padhi, and Kevin Knight. 2016 · 2016
Cited alongside, same era.
What do neural machine translation models learn about morphology?
Yonatan Belinkov, Nadir Durrani, Fahim Dalvi, Hassan Sajjad, and James Glass. 2017 · 2017
Cited alongside, same era.
Towards decoding as continuous optimisation in neural machine translation
Cong Duy Vu Hoang, Gholamreza Haffari, and Trevor Cohn. 2017 · 2017
Cited alongside, same era.
Paraphrasing revisited with neural machine translation
Jonathan Mallinson, Rico Sennrich, and Mirella Lapata. 2017 · 2017
Cited alongside, same era.
What do recurrent neural network grammars learn about syntax?
Noah A. Smith, Chris Dyer, Miguel Ballesteros, Graham Neubig, Lingpeng Kong, and Adhiguna Kuncoro. 2017 · 2017
Cited alongside, same era.
Reference-aware language models
Zichao Yang, Phil Blunsom, Chris Dyer, and Wang Ling. 2017 · 2017
Cited alongside, same era.
HotFlip: White-box adversarial examples for text classification
Javid Ebrahimi, Anyi Rao, Daniel Lowd, and Dejing Dou. 2018 · 2018
Cited alongside, same era.
T-rex: A large scale alignment of natural language with knowledge base triples
Hady ElSahar, Pavlos Vougiouklis, Arslen Remaci, Christophe Gravier, Jonathon S. Hare, Frédérique Laforest, and Elena Simperl. 2018 · 2018
Cited alongside, same era.
The natural language decathlon: Multitask learning as question answering
Original
Bryan McCann, Nitish Shirish Keskar, Caiming Xiong, and Richard Socher. 2018 · 2018
Cited alongside, same era.
On the state of the art of evaluation in neural language models
Gábor Melis, Chris Dyer, and Phil Blunsom. 2018 · 2018
Cited alongside, same era.

Then

BERT: Pre-training of deep bidirectional transformers for language understanding
Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019 · 2019
Closest in time.
Mask-predict: Parallel decoding of conditional masked language models
Marjan Ghazvininejad, Omer Levy, Yinhan Liu, and Luke Zettlemoyer. 2019 · 2019
Closest in time.
A structural probe for finding syntax in word representations
John Hewitt and Christopher D. Manning. 2019 · 2019
Closest in time.
Barack’s wife Hillary: Using knowledge graphs for fact-aware language modeling
Robert L. Logan IV, Nelson F. Liu, Matthew E. Peters, Matt Gardner, and Sameer Singh. 2019 · 2019
Closest in time.
What does BERT learn about the structure of language?
Ganesh Jawahar, Benoît Sagot, and Djamé Seddah. 2019 · 2019
Closest in time.
Facebook FAIR’s WMT19 news translation task submission
Nathan Ng, Kyra Yee, Alexei Baevski, Myle Ott, Michael Auli, and Sergey Edunov. 2019 · 2019
Closest in time.
Knowledge enhanced contextual word representations
Matthew E. Peters, Mark Neumann, Robert Logan, Roy Schwartz, Vidur Joshi, Sameer Singh, and Noah A. Smith. 2019 · 2019
Closest in time.
Language models as knowledge bases?
Fabio Petroni, Tim Rocktäschel, Sebastian Riedel, Patrick Lewis, Anton Bakhtin, Yuxiang Wu, and Alexander Miller. 2019 · 2019
Closest in time.
Language models are unsupervised multitask learners
Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019 · 2019
Closest in time.
Explain yourself! Leveraging language models for commonsense reasoning
Nazneen Fatema Rajani, Bryan McCann, Caiming Xiong, and Richard Socher. 2019 · 2019
Closest in time.
Atomic: An atlas of machine commonsense for if-then reasoning
Original
Maarten Sap, Ronan Le Bras, Emily Allaway, Chandra Bhagavatula, Nicholas Lourie, Hannah Rashkin, Brendan Roof, Noah A Smith, and Yejin Choi. 2019 · 2019
Closest in time.
BERT rediscovers the classical NLP pipeline
Ian Tenney, Dipanjan Das, and Ellie Pavlick. 2019a · 2019
Closest in time.
What do you learn from context? Probing for sentence structure in contextualized word representations
Ian Tenney, Patrick Xia, Berlin Chen, Alex Wang, Adam Poliak, R. Thomas McCoy, Najoung Kim, Benjamin Van Durme, Samuel R. Bowman, Dipanjan Das, and Ellie Pavlick. 2019b · 2019
Closest in time.
Universal adversarial triggers for attacking and analyzing NLP
Eric Wallace, Shi Feng, Nikhil Kandpal, Matt Gardner, and Sameer Singh. 2019 · 2019
Closest in time.
ERNIE: Enhanced language representation with informative entities
Zhengyan Zhang, Xu Han, Zhiyuan Liu, Xin Jiang, Maosong Sun, and Qun Liu. 2019 · 2019
Closest in time.
Inducing relational knowledge from BERT
Original
Zied Bouraoui, Jose Camacho-Collados, and Steven Schockaert. 2020 · 2020
Closest in time.
Latent relation language models
Original
Hiroaki Hayashi, Zecong Hu, Chenyan Xiong, and Graham Neubig. 2020 · 2020
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

Reading the bibliography…