I work in the intersection of Machine Learning and Natural Language Processing, and specifically on Machine Comprehension and Knowledge Base Inference. My goal is to build agents that read, comprehend, and reason with factual texts.
Recently I created two multi-step reading comprehension datasets based on Wikipedia and PubMed. Together with the Allen Institute for Artificial Intelligence, I also gathered a dataset of 13K+ Questions for Science QA.
I am generally interested in understanding theoretical bridges between different Machine Learning models, and in how language and intelligence are interrelated.
I co-organise the South England NLP meetup.
Constructing Datasets for Multi-Hop Reading Comprehension Across Documents
Johannes Welbl, Pontus Stenetorp, Sebastian Riedel
Complex Embeddings for Simple Link Prediction
Théo Trouillon, Johannes Welbl, Sebastian Riedel, Éric Gaussier, Guillaume Bouchard -- ICML, 2016
Casting Random Forests as Artificial Neural Networks (and Profiting from it)
Johannes Welbl -- GCPR, 2014