Vignesh Gokul

Hi! I am a Ph.D. candidate in Computer Science at UC San Diego advised by Prof. Shlomo Dubnov. My research interests includes Creative Interaction/Improvisation, Video/Audio Understanding and Federated Learning.

I am on the job market for research roles in the industry (exp. graduation by Jan 2024)

Email  /  Resume  /  Research Statement  /  Teaching Statement  /  Diversity Statement  /  Google Scholar  /  Github

• Co-creativity/Improvisation: My research goal is to enable creative interactions between humans and visual assistants. These interactions can be either visual (dance) or audio (music improvisation) .

• Distributed Co-creativity: I also focus on building federated/distributed systems that can learn such interactions in the absence of large datasets.

• October 2022: Joined Samsung NEON as Research Scientist Intern.

• October 2022: Our work on Switching Machine Improvisation Models by Latent Transfer Entropy Criteria has been accepted to Physical Sciences Forum.

• July 2022: Our work on FedLTN: Federated Learning for Sparse and Personalized Lottery Ticket Networks has been accepted to ECCV 2022.

• June 2021: Joined Adobe as Machine Learning Research Intern.

• June 2021: Released preprint of our work Bias-Free FedGAN: A Federated Approach to Generate Bias-Free Datasets

• June 2021: DPD-InfoGAN: Differentially Private Distributed InfoGAN has been accepted to EuroMLSys 2021

• June 2020: Joined Unity as Machine Learning Research Intern.

• March 2019: Semantic Interaction with Human Motion Using Query-Based Recombinant Video Synthesis has been accepted to IEEE MIPR

We introduce Symmetric Transfer Entropy (SymTE) as a quantitative metric to switch between generative models based on a control signal.

In this paper we survey the methods for control and cre-ative interaction with pre-trained generative models for audio and music.

In this paper, we propose FedLTN, a novel approach motivated by the well-known Lottery Ticket Hypothesis to learn sparse and personalized lottery ticket networks (LTNs) for communication-efficient and person- alized FL under non-identically and independently distributed (non-IID) data settings.

Federated GANs propagate data biases onto the global model. We propose Bias-Free FedGAN, an approach to generate bias-free synthetic datasets using FedGAN.

We propose a differentially private version of InfoGAN (DP-InfoGAN). We also extend our framework to a distributed setting (DPD-InfoGAN) to allow clients to learn different attributes present in other clients' datasets in a privacy-preserving manner.

In this paper we describe a gestural motif extraction system that combines deep feature learning with structural similarity analysis to allow such query based human-computer motion interaction.

• Low-cost responsive video cached system for effective interaction with digital humans, U.S. Provisional Patent Application Serial No. 63/453,999

• Cached Interactive System for Cost-Effective dialog interaction with digital assistants, U.S. Provisional Patent Application Serial No. 63/453,825

• Early Research Scholars Program (ERSP): Research Mentor for 42 groups (150 students), as a part of a NSF funded program to expose students to research opportunities. Students work in groups of 4 for three academic quarters under the guidance of a faculty member in the CS department on their own research project. This involves submitting a research proposal at the end of the first quarter, implementing and evaluating it in the next two quarters, and presenting a poster at the end of the academic year. (Reviews 2018) (Reviews 2019)(Reviews 2020)

• CSE 231: Advanced Compiler Design (Reviews)

• CSE 194: Race, Gender and Computing (Reviews)

• CSE 258: Recommender Systems and Web Mining (Reviews)

• CSE 291: Machine Learning for 3D Data (Reviews)