Artifacts

Websites

• qasrl.org: Hub for QA-SRL publications. More data and code here.
• nlpsurvey.net: Homepage for the NLP Community Metasurvey.
• poempick.net: Webapp for poetry memorization. Ask me if you’re confused how to use it.
• julianmichael.org: You’re looking at it.

Datasets and webapps

Below are the datasets and associated webapps that I maintain. For all datasets I helped create, see my publications.

• Human and AI debates from Debate Helps Supervise Unreliable Experts.
• Results analyzer for the NLP Community Metasurvey.
• Interactive cluster browser for the induced semantic role ontologies from Inducing Semantic Roles Without Syntax.
• QA-SRL data browser for the QA-SRL Bank 2.0 and QANom dataset.
• Question-Answer Meaning Representations, described in the QAMR paper.

Code

For more code, see my Github, and for repositories specific to papers I’ve published, see my publications.

• qasrl / qasrl-modeling / qasrl-roles: Code for handling QA-SRL data, modeling QA-SRL with neural networks, and inducing semantic roles from QA-SRL, respectively.
• jjm: Base layer of utilities for functional programming in Scala with cats.
• freelog: Monadic, tree-based, rewindable logging with progress bars in Scala.
• spacro: Easily build multi-stage MTurk crowdsourcing pipelines with custom UIs in Scala.
• pubdata: data and formatting tools for making websites like this one.
• dotfiles.

Places you can find me

• Academic: Semantic Scholar, Google Scholar, ACL Anthology, DBLP
• Social: Twitter, YouTube (talks), LessWrong
• Technical: Github, Maven Central