Mahsa Yarmohammadi

Assistant research scientist

Primary Appointment: Center for Language and Speech Processing

Research Interests

  • Natural Language Processing
  • Syntactic and Semantic Parsing
  • Information Extraction
  • Deep Learning
  • Automatic Speech Recognition

Mahsa Yarmohammadi is an assistant research scientist in the Center for Language and Speech Processing (CLSP) who leads state-of-the-art research in cross-lingual language and speech applications and algorithms.

A primary focus of Yarmohammadi’s research is using deep learning techniques to transfer existing resources into other languages and to learn representations of language from multilingual data. As a prime researcher in several multi-site programs in the multilingual research area, she conducts novel research and development on the core technologies, the integration of the components into an end-to-end system, refinement and tight interconnection of methods across components, and iterative end-to-end testing and analysis.

She also works in automatic speech recognition (ASR) and speech translation. Yarmohammadi develops novel finite-state algorithms for the decoder component of ASR. She explores ASR and machine translation (MT) interface issues such as normalizing ASR output to better match MT systems, or modeling ASR output lattices as neural lattice embeddings to be consumed by neural MT systems.

As a post-doctoral fellow at CLSP in 2021, Yarmohammadi was mentored by Benjamin Van Durme, associate professor in the Department of Computer Science, and Sanjeev Khudanpur, associate professor in the Department of Electrical and Computer Engineering, as well as former CLSP member, Daniel Povey.

From 2015 to 2017, Yarmohammadi worked as a software architect at Intel Corporation, and was a  Java/J2EE developer for the GAM Electronics Company from 2004 to 2008. She has also served as a reviewer or program committee member at a variety of conferences, including the 2016 Pacific Northwest Regional NLP Workshop (co-chair), 2021 European Chapter of the Association for Computational Linguistics (review committee) and 2020 Association for the Advancement of Artificial Intelligence Conference (program committee), among others.

Yarmohammadi received her PhD in computer science and engineering from Oregon Health & Science University (2016), and her master’s in computer (software) engineering (2007) at Shahid Beheshti University in Iran. She earned her undergraduate degree in computer (software) engineering (2004) at Amirkabir University of Technology (Tehran Polytechnic).

Johns Hopkins University

Johns Hopkins University, Whiting School of Engineering

Center for Language and Speech Processing
Hackerman 226
3400 North Charles Street, Baltimore, MD 21218-2680

Center for Language and Speech Processing