About

I am Ritesh Goru, currently working at DevRev. I completed my undergraduate degree in Electrical Engineering at IIT Bombay. I’m interested in Stochastic Approximation, MCMC techniques for Optimisation and Bayesian Learning Theory.

Ritesh Goru

Research Experience

I did my undergrad project on distributed SGD with straggler mitigation with Prof. Vivek Borkar and Prof. Nikhil Karamchandani. I also worked with Prof. Subhasis Chaudhuri on Active learning for Distance Metric Learning.

I was a visiting research student at Purdue University under Prof. Juan Wachs in ISAT lab, where I worked on Imitation Learning methods for Robotic Surgery.

Publications

Batch Decorrelation for Active Metric Learning

We present an active learning strategy for training parametric models of distance metrics, given triplet-based similarity assessments: object $x_{i}$ is more similar to object $x_{j}$ than to $x_{k}$ . In contrast to prior work on class-based learning, where the fundamental goal is classification and any implicit or explicit metric is binary, we focus on perceptual metrics that express the degree of (dis)similarity between objects.

Accepted to IJCAI-PRICAI 2020. Read the paper

Technical Projects

I spent 3 years working at Innovation Cell IIT Bombay, where I worked on various technical projects. I was the vision lead for SeDriCa Team. I was one of the 7 members of the winning team of ASME-SDC world finals held at Tampa Florida.

Notes & Blog

Inspired from lil’log and Feynman’s technique, I log things I learn here.

Check out my blog posts for paper summaries, technical notes, and non-technical writings.