Hi, I’m Matt Sárdi, machine learning researcher.
Currently Senior Python Developer in T-Systems Artificial Intelligence & Machine Learning Chapter at Deutsche Telekom.
I'm interested in making voice assistants (Siri, Alexa, Google Assistant) more useful and less frustrating to use, and my guess is one way to make progress is via commonsense reasoning. My research interests overlap most with those of Yejin Choi. I'm also interested in deep learning theory.
I'm a machine learning researcher with 2 years of experience in computer vision, building visual emotion recognition models (photo of a face → it's a smile) for Realeyes, 2 years of experience in natural language processing at Mozaik Education (working with transformer language models), and some more experience being an ordinary software engineer. Currently, I'm interested in commonsense reasoning with transformers and deep learning theory.
Before, I worked at MTA SZTAKI (Hungarian Academy of Sciences' Institute for Computer Science and Control, misleadingly cool name), and did my Bachelor's and Master's in computer science at Eötvös Loránd University in Hungary (The alma mater of Paul Erdős and John von Neumann, which also sounds kinda cool. (In case you actually looked that up and are confused, the university was historically renamed)).
My legal name is Máté Sárdi, but I prefer Matt. My last name is pronounced more or less like "shardy".
You can reach me at [email protected] (note, the name order starts with "sardi").
EENLP: Cross-lingual Eastern European NLP Index
Alexey Tikhonov, Alex Malkhasov, Andrey Manoshin, George Dima, Réka Cserháti, Md.Sadek Hossain Asif, Matt Sárdi
LREC 2022
Using 3D Face Reconstruction Model for Emotion Recognition
Matt Sárdi
Master’s Thesis, 2018
Research Proposal: Does Grokking Happen in the Global Valley?
2021
Source Code Generation with Language Models, a Review of 3 Papers