portrait

Constantine Lignos

Assistant Professor of Linguistics
Department of Computer Science
Volen National Center for Complex Systems
Brandeis University

Email: lastname at brandeis dot edu
Twitter: @ConstantineLig

I direct the Broadening Linguistic Technologies Lab (website coming soon) at Brandeis University, where I am affiliated with the Computer Science Department, Computational Linguistics Program, and Linguistics Program. The overarching goal of my research is to broaden the depth and breadth of human language technology, with a focus on understudied problems in computational linguistics.

The primary thrust of my current work is eliminating the barriers to useful language technology for every living written language, especially lower-resourced and minoritized languages. I also continue to study the representation of language in the mind, including language acquisition, processing, and change.

I did my graduate work in Computer Science at The University of Pennsylvania (Ph.D. 2013), advised by Mitch Marcus and Charles Yang. I then completed a post-doctoral fellowship at The Children's Hospital of Philadelphia exploring clinical applications of statistical models of language processing. I was a researcher at BBN Technologies and USC Information Sciences Institute. In summer 2019, I joined the computational linguistics faculty at Brandeis University.


Latest news

Website update coming soon

After 11 years, the current website design has finally reached the end of its life. I’ll start posting 2021 news once I finish the new design.

Presenting at an AACL 2020 workshop

I’ll be presenting my AACL 2020 Technologies for MT of Low Resource Languages paper Effective Architectures for Low Resource Multilingual Named Entity Transliteration. This work was done with my student Molly Moran, who graduated from the Brandeis Computational Linguistics MS Program in 2020.

Presenting at an EMNLP 2020 workshop

I’ll be presenting my EMNLP 2020 Insights from Negative Results Workshop paper If You Build Your Own NER Scorer, Non-replicable Results Will Come. This work was done with my student Marjan Kamyab, who graduated from the Brandeis Computational Linguistics MS Program in 2020.

Presenting at EMNLP 2019

I’ll be presenting a poster for my EMNLP 2019 paper The Challenges of Optimizing Machine Translation for Low Resource Cross-Language Information Retrieval. This work was done at USC Information Sciences Institute in collaboration with UMass Amherst Center for Intelligent Information Retrieval.


Software

SeqScore: a Python package for evaluating named entity recognition (NER) and other chunking tasks.

MORSEL: a cognitively-motivated state-of-the-art unsupervised morphological analyzer I developed for Morpho Challenge 2010. It achieved state-of-the-art results in English and Finnish.

Codeswitchador: a system for identifying code-switching in social media data. This work enables the creation of large scale corpora of code-switching and identification of bilingual users. I developed this as a participant of the SCALE summer workshop at the Johns Hopkins Center of Excellence in Human Language Technology.

Regrettably, much of my research over the last few years is closed-source. However, many projects are publicly available on GitHub.


Teaching

My current teaching at Brandeis includes introductory and advanced courses in computational linguistics. My course materials are all posted on the Brandeis LATTE learning management system, but names of courses I've taught recently are available on my faculty guide page.

In the past I have taught researchers to write great Python code. The notes for the bootcamps I've done are available at Python Boot Camp for Researchers.

I used to maintain a list of common mistakes that programmers new to Python make: Anti-Patterns in Python Coding. It's now outdated, but still has some good content.

In Spring 2011 and 2012, I taught one of the CIS department's "mini-courses," Python Programming (CIS 192).

In the past I've also led some informal groups for learning Python. The slides from those groups can be found on my Python for Language Researchers Site.