NeurIPS 2019 accepted papers on Large language models
22 of the 1,429 papers accepted at NeurIPS 2019 (1.5%) are on Large language models, compared with 0.9% at NeurIPS 2018. They fall into 5 topics.
See NeurIPS 2019 in the interactive map
Submitting to NeurIPS? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Language, safety and interpretability · 12 papers
- XLNet: Generalized Autoregressive Pretraining for Language Understanding (1,824 citations)
- Unified Language Model Pre-training for Natural Language Understanding and Generation (941 citations)
- Interpreting and improving natural-language processing (in machines) with natural language-processing (in the brain) (140 citations)
- Defending Against Neural Fake News (90 citations)
- Approximating Interactive Human Evaluation with Self-Play for Open-Domain Dialog Systems (76 citations)
- Anti-efficient encoding in emergent communication (68 citations)
- Assessing Social and Intersectional Biases in Contextualized Word Representations (67 citations)
- Slice-based Learning: A Programming Model for Residual Learning in Critical Data Slices (41 citations)
- Inducing brain-relevant bias in natural language processing models (37 citations)
- Abstract Reasoning with Distracting Features (30 citations)
- Can Unconditional Language Models Recover Arbitrary Sentences? (15 citations)
- Learning Dynamics of Attention: Human Prior for Interpretable Machine Reasoning (4 citations)
AI and society · 4 papers
- SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems (984 citations)
- Bridging Machine Learning and Logical Reasoning by Abductive Learning (89 citations)
- Learning Perceptual Inference by Contrasting (61 citations)
- Ask not what AI can do, but what AI should do: Towards a framework of task delegability (37 citations)
Agents, code and math · 3 papers
Retrieval and knowledge · 2 papers
- Quantum Embedding of Knowledge for Reasoning (16 citations)
- Implicitly learning to reason in first-order logic (5 citations)
