Jason Chuang

Timeline

1. 2012

   Termite for topic models

   With Christopher D. Manning and Jeffrey Heer, he presents Termite at the Advanced Visual Interfaces conference in 2012. The visual analysis tool for assessing topic model quality uses a tabular layout that compares terms within and across latent topics.

2. 2013

   A Stanford doctorate

   He completes a PhD in computer science at Stanford in 2013, working with Jeffrey Heer and Christopher D. Manning. His dissertation, Designing Visual Text Analysis Methods to Support Sensemaking and Modeling, examines how interactive visualizations and statistical analysis algorithms can be designed together. He is also a postdoctoral researcher at the University of Washington and a visiting research scientist at the Allen Institute for Artificial Intelligence.

3. 2013

   Sentiment Treebank at EMNLP

   He is one of seven authors, with Richard Socher, Alex Perelygin, Jean Wu, Christopher D. Manning, Andrew Ng and Christopher Potts, of "Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank" at EMNLP 2013. He applies visual analytics to study how the deep learning algorithm captures the effects of sentiment negations, and introduces a framework at ICML 2013 for large-scale evaluation of statistical topic models.

4. 2014

   Interactive language learning workshop at ACL

   In 2014, he organizes the Workshop on Interactive Language Learning, Visualization, and Interfaces at ACL 2014.

5. 2015

   Topic model stability at NAACL

   With Margaret E. Roberts, Brandon M. Stewart, Rebecca Weiss, Dustin Tingley, Justin Grimmer and Jeffrey Heer, he co-authors "Interactive Alignment for Assessing Topic Model Stability" at the 2015 Conference of the North American Chapter of the ACL (NAACL).

6. 2020

   Open letter on ad transparency

   In January 2020, as a researcher at Mozilla, he reports to BuzzFeed News that Facebook's Ad Library API often fails from programming errors and returns incomplete data. In October 2020, he signs a TechCrunch open letter in which 15 researchers propose a standard for universal digital ad disclosure; his projects bring greater transparency to digital advertising and recommendation algorithms, and his research looks at how people work with and trust intelligent computing systems.

7. 2021

   A universal ad transparency standard

   In March 2021, he and Laura Edelson give the opening remarks at a Mozilla Festival session on making online advertising more transparent. In December 2021, the Knight First Amendment Institute publishes "A Standard for Universal Digital Ad Transparency", which he co-authors with Edelson, Erika Franklin Fowler, Michael M. Franz and Travis Ridout; the proposal sets criteria that trigger a transparency requirement, with ad data collected by a government agency. The Institute describes him as a researcher in data science and analytics at Mozilla.

8. 2023

   A Test-of-Time Award

   In 2023, the ACL Test-of-Time Paper Awards include the 2013 EMNLP paper "Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank", which he co-authored with Richard Socher and five others.