Robustly Poly-Skilled

On 25 June 2024, Lee Ween Jiann successfully defended his dissertation, titled “Incorporating Intrinsic Structures into Entity Matching and Representation Learning”. The work came in two parts.

The first part looked into algorithms for facilitating entity matching, when an entity from one side could match multiple entities on the other side (poly-matching), but the “one” and the “many” could come from either side (bipoly-matching).

Ween outlining the structure of his thesis at the beginning of the defense

The second part studied representation learning, particularly for geospatial entities, whereby the relative distance and directions were taken into account in feature learning.

Ween illustrating the geospatial representation learning as applied to MNIST superpixel data

Ween first joined PreferredAI in 2017 as a research engineer, out of an undergraduate business school. In 2019, he transitioned into a PhD candidate.

Dinner at The Rice Table in July 2017 while Ween was a research engineer (Ween was at the centre-left of the picture, at the end of the long table)

He had an almost CrossFit-like approach to PhD training. While pursuing research, he had the the time, the skills, and the willingness to undertake various activities, such as teaching, mentoring younger students, organizing events, maintaining our compute system, and pursuing entrepreneurial ventures. As a result, as he graduated with credible papers under his belt (ICDM-21, TKDE-23, TSAS-24) as well as SMU Presidential Doctoral Fellowship, he also brought with him a diverse array of skillsets that would stand him in good stead in the future.

Group celebration at Pemenco CHIJMES after the defense (Ween in white shirt, second from right)

We wish him all the best for his next endeavour!