I am Varun Srivastava, a graduate student at Stanford University pursuing a Masters in Electrical Engineering. A graduate of IIT Delhi, my main interests lie in establishing the mathematical foundations of and accelerating Deep learning algorithms, and their applications such as
- Biomedical (Photoplethysmography)
- Computer Vision
- Sequence Analysis and Generation
My future research endeavours aim to provide a firmer footing to algorithms used to train neural networks by using information theoretic tools and optimization theory. In practice, this may manifest as optimizations to training and inference procedures both at an algorithmic and system implementation levels.
In the past I have had the fortune to work in highly integrated teams, which have allowed me to not only bring my projects to successful completion, but also thrive in an open work environment that fosters both personal growth, leadership abilities and teamwork. My mission is to develop sound principles for (machine) learning algorithms that have cross cutting real world impact across a range of domains from medicine to engineering.
Adobe Research (BEL), Bangaloreresearch.adobe.com
May 2018 - July 2018
Across Journey Customer Experience Measurement
- Implemented Inverse Reinforcement Algorithms to model Customer Experience in an online setting.
- Proposed a new metric incorporating aspects of Consumer Psychology and Learning to measure Experience.
- Worked on distributed database management library Apache Spark to process more than 2 TB of clickstream data.
Indian Institute of Technology, Delhiiitd.ac.in
July 2018- December 2018
Acted as teaching assistant for the course ELL205 (Signals and Systems) comprising over 150+ students from diverse backgrounds. The course establishes a strong mathematical foundation in tools for understanding and manipulating signals including various Fourier Transforms, Laplace Transforms and frequency analysis.
- Designed and invigilated quizzes (6) conducted and automated via Moodle.
- Graded and disseminated solutions to all exams (3) conducted during the duration of the course.
- Taught weekly tutorials (office hours) discussing doubts and applications for ongoing topics in the class.
Projects and Publications
Adversarial Approximate Inference for Speech to Electroglottograph Conversion
Sept 2018 - Jan 2019 Prof. Prathosh AP, Dept of EE, IIT Delhi
A distribution transformation framework to map speech to the corresponding EGG signal, robust to noise, generalizing across recording conditions, speech pathologies and voice qualities.
- Optimized the Speech to Laryngograph encoder using adversarial training for the network.
- Created a cosine based loss function for enforcing amplitude invariance between ground truth and network output.
- Used a variational inference approach for learning optimal representations for the speech signal.
- Utilized continuous wavelet transforms using Ricker wavelets for robust peak picking.
Published in IEEE/ACM Transactions on Audio, Speech, and Language Processing. Manuscript available at arxiv link
Glottal Closure Instants (GCIs) correspond to the temporal locations of significant excitation to the vocal tract occurring during the production of voiced speech. GCI detection is cast as a supervised multi-task learning problem solved using a deep convolutional neural network jointly optimizing a classification and regression cost.
- Devised state of the art deep dilated CNN to detect and locate GCIs from raw speech.
- Performed multi task learning (simultaneous regression and classification) on speech for temporal event detection.
- Created a robust Weighted Histogram based Clustering Algorithm for time series clustering.
- Beat State of the Art method with Lowest Variance in Localisation Error on noisy speech.
Published in INTERSPEECH 2019. Manuscript available at arxiv link
Mode matching in GANs through latent space learning and inversion
Prof. Prathosh AP, Dept of EE, IIT Delhi
NEMGAN addresses the problem of imposing desired modal properties on the generated space using a latent distribution, engineered in accordance with the modal properties of the true data distribution. This is achieved by training a latent space inversion network in tandem with the generative network using a divergence loss.
- Construction of a multi modal latent distribution by reparameterization of a mixture of continuous and discrete random variables employable in any GAN framework.
- Imposition of modal properties of the latent space the generated space using two-stage latent inversion.
- Learnt the mode priors of the latent distribution to follow true data distribution using sparse-supervision.
Manuscript available at arxiv link
Non Contact CardioPulmonary Assessment using Imaging Techniques
Prof. Prathosh AP, Dept of EE, IIT Delhi
This porject builds a single end to end model - for estimation of rPPG (and its derivatives) from videos, using a neural network for all modelling conditions (Lighting, Skin Complexion etc). It also explores efficacy of using Near Infrared spectrum (650 - 1350 nm), which is optimal for remote monitoring – as it has maximal penetration depth.
- Performing non contact estimation of human vitals (heart rate, oxygen saturation etc) using infrared cameras.
- Devising deep learning models to perform human skin segmentation from videos in near infrared spectrum.
Slides available at Google Slides
Awarded Qualcomm Research Innovation Fellowship with Rs 10,00,000 of funding for the project proposal
This project integrates contextual information from Brown Corpus and WordNet into a text generating model to produce grammatically correct, meaningful sentences in the long run.
- Designed a probabilistic discriminative model using Brown Corpus and WordNet to perform text prediction.
- Implemented a LSTM in Chainer (dynamic computational graph framework) for language modelling.
- Improved word perplexity by building a factored language model incorporating both words and parts of speech.
Slides available at Google Slides
Awarded Summer Undergraduate Research Award for successful completion
NAVI: Navigation Assistance for Visually Impaired
Prof. Balakrishnan, Dept of CS, IIT Delhi
NAVI (SmartCane 2.0) is a smart navigation device to help the visually impaired navigate using minimum external help. Equipped with several accessibility features, this device warns the user if they are heading wrong and gives tactile instructions for course correction. It also allows them to save new routes while navigating and add attributes like landmarks in real time.
- Designed a standalone device on Raspberry Pi (six person team) to assist visually impaired in navigation.
- Modeled navigation as a graph traversal in Python to determine best routes and perform course correction.
- Implemented a Kalman Filter to integrate data from a GPS and a 9 axis IMU for precise localization.
Blog post available at medium link
- Qualcomm Research Innovation Fellowship 2018
- Won as a team of two with a financial reward of Rs 10,00,000. Amongst 8 selected teams from 95 eligible teams from Institutes IITB, IITD, IITK, IITM, IIT-KGP, IISc, IIIT-H.
- Narotam Sekhsaria Scholarship 2019-2020
- Granted the prestigious NSF scholarship-loan worth Rs 15,00,000 for academic excellence and pursuing post-graduate education at Stanford University. Awarded to 0.1% of applicants
- KC Mahindra Scholarship 2019-2020
- Granted the KC Mahindra scholarship-loan worth Rs 4,00,000 for academic excellence and pursuing post graduate education at Stanford University.
- Summer Undergraduate Research Award
- Awarded Summer Undergraduate Research Award 2017 by Industrial Research and Development Unit, IITD for excellence in research
- IITD Semester Merit Award
- Awarded for being in top 7% undergraduate students for 5 consecutive semesters
- Department Rank 4
- Out of 80+ students in Electrical Engineering Department
Proficiencies and Skills
MS in Electrical Engineering
Sept 2019 - Current
Indian Institute of Technology, Delhi
B.Tech in Electrical Engineering
July 2015 - July 2019
Ryan International School
- Artificial Intelligence
- A Deeper Theory of Deep Learning: The Quest for Information-Theoretic Foundations
- Deep Learning
- Machine Learning
- Signals and Systems
- Digital Image Processing
- Communication Engineering
- Operating Systems
- Control Engineering
- Data Structures
- Probability and Stochastic Processes
- Fundamentals of Language Sciences
- Econometric Methods
A Little More About Me
I love to read fantasy during my leisure time. A short (and incomplete) list of my favoured authors includes Patrick Rothfuss, Brandon Sanderson, Brent Weeks, Neil Gaiman, N.K. Jemisin. You can also find me tinkering with the shiniest deep learning framework on the block (PyTorch at the moment) with my trusty cup of tea.