Yash Mehta

Yash Mehta

Research Engineer

AutoML Lab, Freiburg

Biography

Hi! I’m currently a research engineer working on challenging neural architecture search research under the supervision of Prof Frank Hutter (ELLIS Fellow). Here, I am also working on advancing the state of the art in deep learning for EEG data prediction. Previously, I was a researcher at the Gatsby Computational Neuroscience Unit at UCL, where I was working on evaluating biologically plausible perturbation-based learning algorithms to train deep networks under the guidance of Prof Peter Latham (Gatsby) and Tim Lillicrap (DeepMind).

In the past, I did my research thesis on deep learning-based personality detection from text with Prof Erik Cambria at NTU Singapore. Before getting into research, I was working as a software developer at Amazon.

I thoroughly enjoy coding and working on hard algorithmic problems. My interests lie at the intersection of deep learning and neuroscience.

Download my resumé.

Experience

 
 
 
 
 
Machine Learning Lab
Research Engineer
Oct 2020 – Present Freiburg, Germany
Supervisor, Frank Hutter. Working on neural architecture search (NAS) applied to EEG data, development of NASLib - a modular, extensible and easy to use NAS library.
 
 
 
 
 
Gatsby Computational Neuroscience Unit, UCL
Research Staff
Jan 2019 – Feb 2020 London, UK
Supervisors, Timothy Lillicrap (DeepMind) and Peter Latham (Gatsby). Two papers under review at NeurIPS'21 on biologically plausible learning algorithms.
 
 
 
 
 
Amazon
Software Development Engineer
Amazon
Jun 2018 – Dec 2018 Bangalore, India
Full-time developer part of the Amazon Music team. Quit this job to get into research.
 
 
 
 
 
SenticTeam, Nanyang Technological University
Research Intern
Jan 2018 – Jun 2018 Singapore
Supervisor, Erik Cambria (NTU Singapore). Ongoing collaboration resulting in 4 published papers on deep learning based personality prediction.

Publications

Google Scholar

On the Limitations of Perturbation-Based Methods for Training Deep Networks
Towards Biologically Plausible Convolutional Networks
Multitask Learning for Emotion and Personality Detection
Bottom-up and top-down: Predicting personality with psycholinguistic and language model features
Personality trait detection using bagged svm over bert word embedding ensembles
Recent trends in deep learning based personality detection