[computer scientist, language enthusiast, puzzle addict]

about ▽

== paxia@cs.jhu.edu | ℂƲ | tl;dr | github: pitrack ==

I am a third year PhD student at JHU's Center for Language and Speech Processing working with Benjamin Van Durme. I'm interested in entity-level information extraction (coreference), sentence representations, and noisy/web data. Specifically, I've been looking at the use and helpfulness of various kinds of context for understanding language.

I also enjoy writing and solving puzzles, biking, rock climbing, and playing (board, video, and social deduction) games.

news ▽

2/19 - I will be interning in Google Pittsburgh for summer 2019.

papers ▽

Improved Lexically Constrained Decoding for Translation and Monolingual Rewriting

J. Edward Hu, Huda Khayrallah, Ryan Culkin, Patrick Xia, Tongfei Chen, Matt Post and Benjamin Van Durme
In NAACL (2019)
= pdf code website =

Probing What Different NLP Tasks Teach Machines about Function Word Comprehension

Najoung Kim, Roma Patel, Adam Poliak, Patrick Xia, Alex Wang, Tom McCoy, Ian Tenney, Alexis Ross, Tal Linzen, Benjamin Van Durme, Samuel R. Bowman and Ellie Pavlick
In *SEM (2019)
= pdf code =

What do you learn from context? Probing for sentence structure in contextualized word representations

Ian Tenney, Patrick Xia, Berlin Chen, Alex Wang, Adam Poliak, R Thomas McCoy, Najoung Kim, Benjamin Van Durme, Sam Bowman, Dipanjan Das, Ellie Pavlick
In ICLR (2019)
= OpenReview pdf related slides poster code =

Looking for ELMo's friends: Sentence-Level Pretraining Beyond Language Modeling

Samuel R. Bowman, Ellie Pavlick, Edouard Grave, Benjamin Van Durme, Alex Wang, Jan Hula, Patrick Xia, Raghavendra Pappagari, R. Thomas McCoy, Roma Patel, Najoung Kim, Ian Tenney, Yinghui Huang, Katherin Yu, Shuning Jin, Berlin Chen
= arXiv OpenReview pdf code =

Consensus for Multi-Parallel Corpora: An English Bible Study

Patrick Xia and David Yarowsky
In IJCNLP (2017)
= pdf slides code =

CoNLL-SIGMORPHON 2017 Shared Task: Universal Morphological Reinflection in 52 Languages

Ryan Cotterell, Christo Kirov, John Sylak-Glassman, Géraldine Walther, Ekaterina Vylomova, Patrick Xia, Manaal Faruqui, Sandra Kübler, David Yarowsky, Jason Eisner, and Mans Hulden
In CONLL (2017)
= pdf website =

projects ▽

In roughly reverse chronological order. Note that some projects were hacked together within a couple of days, while others over the course of a year.


- (2018-now) [private; WIP] a discord bot for puzzles and various other information-retrieval ish tasks.


- (2016) a simple word vector implementation of a codenames game AI

this website

- (2016-now) a from-scratch attempt to construct something in CSS/HTML

Translating into Morphologically Rich Languages

(Senior Honors Thesis) - (2016) learning morphology with LSTMs + context


(15-418 final project) - (2016) monte carlo tree search solver for pathery (github)

Character Relationships in Literary Text

(Independent study) - (2015) collected annotations and tried to learn relationships between fictional characters.

puzzles ▽

I write puzzles (Galactic Puzzlehunt, formerly for PuzzlehuntCMU) and participate in MIT Mystery hunt and other hunts near me. This is an incomplete list. A complete list and personal hunting history might be found here.



50/50 Solution
Now Get Ready For... Solution
Fan Video Solution
Chromesthesia Solution

📱 ▽

I take (low quality) pictures that have contain a unique sight, tell an interesting story, or represent something I want to remember. One day I'll stop uploading them to Facebook and they may appear here instead. If you are my friend on Facebook, you should be able to see my albums Seattle XIV (2014) and Yoshi's Story (2017+). There are also 100+ photos from 2016-present include pictures from various cities around the world. They are now desktop background images. You might be able to see them when you walk past my screen!

📝 ▽

Some more serious notes I've taken on language, computer science, machine learning, and math.