Supervised Learning of Universal Sentence Representations from Natural Language Inference Data
- Alexis Conneau, Douwe Kiela, Holger Schwenk, Loïc Barrault, Antoine Bordes
- Computer ScienceConference on Empirical Methods in Natural…
- 5 May 2017
It is shown how universal sentence representations trained using the supervised data of the Stanford Natural Language Inference datasets can consistently outperform unsupervised methods like SkipThought vectors on a wide range of transfer tasks.
Personalizing Dialogue Agents: I have a dog, do you have pets too?
- Saizheng Zhang, Emily Dinan, Jack Urbanek, Arthur D. Szlam, Douwe Kiela, J. Weston
- Computer Science, PsychologyAnnual Meeting of the Association for…
- 22 January 2018
This work collects data and train models tocondition on their given profile information; and information about the person they are talking to, resulting in improved dialogues, as measured by next utterance prediction.
Poincaré Embeddings for Learning Hierarchical Representations
- Maximilian Nickel, Douwe Kiela
- Computer ScienceNIPS
- 22 May 2017
This work introduces a new approach for learning hierarchical representations of symbolic data by embedding them into hyperbolic space -- or more precisely into an n-dimensional Poincare ball -- and introduces an efficient algorithm to learn the embeddings based on Riemannian optimization.
Adversarial NLI: A New Benchmark for Natural Language Understanding
- Yixin Nie, Adina Williams, Emily Dinan, Mohit Bansal, J. Weston, Douwe Kiela
- Computer ScienceAnnual Meeting of the Association for…
- 31 October 2019
This work introduces a new large-scale NLI benchmark dataset, collected via an iterative, adversarial human-and-model-in-the-loop procedure, and shows that non-expert annotators are successful at finding their weaknesses.
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- Patrick Lewis, Ethan Perez, Douwe Kiela
- Computer ScienceNeural Information Processing Systems
- 22 May 2020
A general-purpose fine-tuning recipe for retrieval-augmented generation (RAG) -- models which combine pre-trained parametric and non-parametric memory for language generation, and finds that RAG models generate more specific, diverse and factual language than a state-of-the-art parametric-only seq2seq baseline.
SentEval: An Evaluation Toolkit for Universal Sentence Representations
- Alexis Conneau, Douwe Kiela
- Computer ScienceInternational Conference on Language Resources…
- 1 March 2018
We introduce SentEval, a toolkit for evaluating the quality of universal sentence representations. SentEval encompasses a variety of tasks, including binary and multi-class classification, natural…
The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes
- Douwe Kiela, Hamed Firooz, Davide Testuggine
- Computer ScienceNeural Information Processing Systems
- 10 May 2020
This work proposes a new challenge set for multimodal classification, focusing on detecting hate speech in multimodal memes. It is constructed such that unimodal models struggle and only multimodal…
What makes a good conversation? How controllable attributes affect human judgments
- A. See, Stephen Roller, Douwe Kiela, J. Weston
- PsychologyNorth American Chapter of the Association for…
- 22 February 2019
This work examines two controllable neural text generation methods, conditional training and weighted decoding, in order to control four important attributes for chit-chat dialogue: repetition, specificity, response-relatedness and question-asking, and shows that by controlling combinations of these variables their models obtain clear improvements in human quality judgments.
Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometry
- Maximilian Nickel, Douwe Kiela
- Computer ScienceInternational Conference on Machine Learning
- 9 June 2018
It is shown that an embedding in hyperbolic space can reveal important aspects of a company's organizational structure as well as reveal historical relationships between language families.
The Second Conversational Intelligence Challenge (ConvAI2)
- Emily Dinan, V. Logacheva, J. Weston
- Computer ScienceThe NeurIPS '18 Competition
- 31 January 2019
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.
...
...