Publications

The Surprising Performance of Simple Baselines for Misinformation Detection

We examine the performance of a broad set of modern transformer-based language models and show that with basic fine-tuning, these models are competitive with and can even significantly outperform recently proposed state-of-the-art methods

Pelrine, Kellin, Jacob Danovitch, and Reihaneh Rabbany. The Surprising Performance of Simple Baselines for Misinformation Detection. arXiv preprint arXiv:2104.06952 (2021). https://arxiv.org/abs/2104.06952

ComplexDataLab at W-NUT 2020 Task 2: Detecting Informative COVID-19 Tweets by Attending over Linked Documents

We present Gapformer, which effectively classifies content as informative or not. It reformulates the problem as graph classification, drawing on not only the tweet but connected webpages and entities.

Pelrine, Kellin, et al. ComplexDataLab at W-NUT 2020 Task 2: Detecting Informative COVID-19 Tweets by Attending over Linked Documents.Proceedings of the Sixth Workshop on Noisy User-generated Text (W-NUT 2020). 2020. https://www.aclweb.org/anthology/2020.wnut-1.63/

Towards a Computational Approach to Conceptual Metaphor

Exploring a computational model of Lakoff’s conceptual metaphor theory.

Recommended citation: Lynch, B., Danovitch, J., & Davies, J. (2018). Towards a Computational Approach to Conceptual Metaphor. Poster session at CogSci 2019, Montreal, CA.