Huda Khayrallah

Huda Khayrallah 

Center for Language and Speech Processing
Department of Computer Science
The Johns Hopkins University

firstname at jhu dot edu

Hackerman 226
3400 N Charles Street
Baltimore, MD 21218

About | Publications | Talks | Teaching | CV | Google Scholar

News

Simulated Multiple Reference Training Improves Low-Resource Machine Translation was nominated to for best paper at WeCNLP.
[3 min video poster] [8 min lightning talk]
Simulated Multiple Reference Training Improves Low-Resource Machine Translation was accepted to EMNLP. This is work with Brian Thompson, Matt Post, and Philipp Koehn.
SMRT Chatbots: Improving Non-Task-Oriented Dialog with Simulated Multiple Reference Training was accepted in findings of EMNLP. This is work with João Sedoc.
We had the highest scoring submission in all 5 language pairs to the Duolingo Shared Task on Simultaneous Translation and Paraphrase at WNGT at ACL 2020. [paper] [slides.ppt] [slides.pdf]
I was awarded the 2019 Jelinek Fellowship. Thanks CLSP!

About

I am a PhD student in Computer Science at The Johns Hopkins University (JHU), advised by Philipp Koehn. I am part of the Center for Language and Speech Processing (CLSP) and the machine translation group. I work on Natural Language Processing (NLP), specifically machine translation (MT). I am particularly interested in the challenges arising from low resource and domain adaptation situations.

In Summer 2019, I was a research intern at Lilt, working on translator-in-the-loop machine translation.

I graduated from UC Berkeley with a B.A. in Computer Science in May 2015. I worked with Colleen Lewis on computer science education, Andreas Stolcke on tools for speech processing, and Jerry Feldman on natural language understanding.

Below are highlights about me; you may also be interested in my full CV. You can also find my work on Google Scholar.

Education

The Johns Hopkins University | August 2015 - present

University of California, Berkeley | August 2011 - May 2015

Publications

* indicates authors contributed equally.

Unrefereed Reports

Talks

Teaching

Huda rymes with CUDA (parallel framework) and gouda (cheese) – WHO-duh – /ˈhuːdə/
Khayrallah is less scary than it looks! kay-RA-la – /keɪˈɹɑlɑ/
My pronouns are she/her.