Elliot Schumacher

I am a Machine Learning Researcher at Curai. In Fall 2022, I defended my Ph.D. from Center for Language and Speech Processing at Johns Hopkins University, working with Mark Dredze, Jim Mayfield, and others on Natural Language Processing and Information Extraction.

In Fall 2020, I interned with the Subjective NLP group at Amazon Search in Barcelona, Spain, as an Applied Scientist Intern. I received my masters from the Language Technologies Institute in the School of Computer Science at Carnegie Mellon University. My research focused on the DSCoVAR project with Maxine Eskenazi, Kevyn Collins-Thompson and others. Before attending CMU, I graduated from Ohio State University with a degree in both Computer Science and Linguistics.

Feel free to contact me at [my first name] AT elliotschu.com

CV.

Publications (for the latest, see Google Scholar)

• A. Sun, V. Nair, E. Schumacher, A. Kannan. CONSCENDI: A Contrastive and Scenario-Guided Distillation Approach to Guardrail Models for Virtual Assistants . NAACL 2024

• V. Nair, E. Schumacher, A. Kannan. Generating medically-accurate summaries of patient-provider dialogue: A multi-stage approach using large language models. Clinical Natural Language Processing Workshop at ACL 2023

• E. Schumacher, J. Mayfield, M. Dredze. Improving Zero-Shot Multi-Lingual Entity Linking. The 2nd Workshop on Multi-lingual Representation Learning (MRL) at EMNLP 2022

• E. Schumacher, J. Mayfield, M. Dredze. Cross-Lingual Transfer in Zero-Shot Cross-Language Entity Linking. Findings of the 2021 Association of Computational Linguistics.

• E. Schumacher, A. Mulyar, M. Dredze. Clinical Concept Linking with Contextualized Neural Representations. Proceedings of the 2020 Association of Computational Linguistics.

• A. Mulyar, E. Schumacher, M. Rouhizadeh, M. Dredze. Phenotyping of Clinical Notes with Improved Document Classification Models Using Contextualized Neural Language Models. NeurIPS Workshop on Machine Learning for Health (ML4H), Non-Archival Abstract.

• E. Schumacher, M. Dredze. Learning Unsupervised Contextual Representations for Medical Synonym Discovery JAMIA Open.

• E. Schumacher, M. Dredze. Discriminative Candidate Generation for Medical Concept Linking. Proceedings of the 2019 AKBC Conference.

• E. Schumacher, M. Eskenazi, G. Frishkoff, K. Collins-Thompson. Predicting the Relative Difficulty of Single Sentences With and Without Surrounding Context. Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing (EMNLP 2016).

Technical Reports and Press Coverage

• E. Schumacher, M. Eskenazi. A Readability Analysis of Campaign Speeches from the 2016 US Presidential Campaign. arxiv.org.
  • CMU Press Release: Most Presidential Candidates Speak at Grade 6 - 8 Level.
  • Huffington Post: Trump’s Speeches Are At A Middle-School Reading Level, Study Says.
  • Washington Post: Trump’s grammar in speeches ‘just below 6th grade level,’ study finds.
  • Pittsburgh Tribune-Review: Speeches dip below 6th-grade level, study says