UCL NLP Reading Group

Schedule: Mondays 12:00-13:00 [Europe/London]

2022

  • 04/04/2022: Word Order Does Matter (And Shuffled Language Models Know It)  [ PDF]. Presented by Max Bartolo.
  • 28/03/2022: Adversarial Weight Perturbation Helps Robust Generalization [ PDF]. Presented by Maximilian Mozes.
  • 14/03/2022: Rethinking the Role of Demonstrations: What Makes In-Context Learning Work? [ PDF]. Presented by Yao Lu.
  • 07/03/2022: Hindsight: Posterior-guided training of retrievers for an improved open-ended generation [ PDF]. Presented by Yuxiang Wu.
  • 21/02/2022: CoAuthor: Designing a Human-AI Collaborative Writing Dataset for Exploring Language Model Capabilities [ PDF]. Presented by Sebastian Riedel.
  • 14/02/2022: Large-Scale Representation Learning on Graphs via Bootstrapping [ PDF]. Presented by Yihong Chen.
  • 04/02/2022: The curious case of neural text degeneration [ PDF]. Presented by Luca Franceschi.
  • 28/01/2022: DREAM: Uncovering Mental Models behind Language Models [ PDF]. Presented by Patrick Lewis.
  • 21/01/2022: MAUVE: Measuring the Gap Between Neural Text and Human Text using Divergence Frontiers [ PDF]. Presented by Linqing Liu.
  • 14/01/2022: Editing a Classifier by Rewriting Its Prediction Rules [ PDF]. Presented by Maximilian Mozes.

2021

  • 26/11/2021: True Few-Shot Learning with Language Models [ PDF]. Presented by Yao Lu.
  • 29/10/2021: Symbolic Knowledge Distillation: from General Language Models to Commonsense Models [ PDF]. Presented by Yihong Chen.
  • 22/10/2021: Prompt-Learning for Fine-Grained Entity Typing [ PDF]. Presented by Yao Lu.
  • 13/10/2021: PonderNet: Learning to Ponder [ PDF]. Presented by Minqi Jiang.
  • 01/10/2021: NeuroLogic Decoding: (Un)supervised Neural Text Generation with Predicate Logic Constraints [ PDF]. Presented by Linqing Liu.
  • 24/09/2021: Editing Factual Knowledge in Language Models [ PDF]. Presented by Max Bartolo.
  • 17/09/2021: Neural Bellman-Ford Networks: A General Graph Neural Network Framework for Link Prediction [ PDF]. Presented by guest speaker Zhaocheng Zhu, Mila-Quebec AI Institute.

The UCL NLP reading group is hosted by Yao Lu. Previously, it has also been hosted by Yihong Chen.