In this paper, we present a domain adaptation based generative framework for Zero-Shot Learning. We explicitly target the problem of …


KotlinSyft makes it easy for you to train and execute PySyft models on Android devices. This allows you to utilize training data located directly on the device itself, bypassing the need to send a user’s data to a central server. The first use case for our library is here

This work aims to derive a non-trivial breakdown point for an algorithm for training a single hiddenlayer Neural Network. In pursuit of this goal, we propose two algorithms for training a network with ReLU activations. The first approach utilizes the partitioning property of the ReLU function while the second approach utilizes the convexity of the activation function.

This is an android app that relays computer graphics like arrows on real time camera feed for navigation. The app uses data from in-mobile sensors and google maps API to identify roads on the screen.



Research Intern

University of California, Berkeley

Jun 2020 – Present Berkeley, USA

Advisor : Dawn Song

Objective : Neural symbolic hybrids for few shot recognition

  • Using program induction to sample programs for few shot image classification.

  • Training procedure involves Supervised pre-training with teacher-forcing followed by reinforcement learning using Hindsight Experience Replay.

  • We use memory augmented networks with attention to allow multiple chains of execution.

Objective : Meta learning in SQL query synthesis

  • Divided the Spider dataset into 13 meta categories.

  • We use transformers to generate embeddings for natural language question tokens.

  • A meta-training phase for decoder to learn predicting the structure of SQL query

  • A domain specific training phase for token prediction using a separate multi head attention module.


Visiting Research Scholar

MPI Brain Research Institute

Aug 2019 – Mar 2020 Frankfurt

Advisor : Prof. Moritz Helmstaedter

Objective : Myelin segmentation in 3D mSEM and connectomic analysis

  • We used 3D Unet trained on multi Scanning Electron Microscope raw data to generate 3D segmentation masks

  • Dynamically oversampling (with linearly decaying probability) myelinated voxel cubes to provide non-zero gradients countering highly skewed data (0.01% positively labeled voxels)

  • Reached over 90% precision-recall. Using the detection to analyze thalamocortical neurons with myelinated axons at the beginning of innervation.


Visiting Research Scholar

NexT++ centre

May 2018 – Jul 2018 NUS Singapore

Advisor : Prof. Tat Seng Chua

Objective: Monocular 3D object instance recognition and Pose Estimation

  • Proposed (alongside a post graduate student) a novel end-to-end architecture consisting of two modules for robust pose prediction and 3D instance recognition via extracting Marr’s 2.5 D sketches from images.

  • The learned embedding explicitly disentangles a shape vector and a pose vector, which alleviates both pose bias for 3D shape retrieval and categorical bias for pose estimation

  • One sub module learns to reconstruct 3D model, from the 2.5D sketches, in its canonical viewpoint via multi-task learning DNNs. Another NN sub module uses Faster R-CNN style anchor boxes to predict the 6 DoF poses in continuous domain.

  • The method achieves state of the art 10.3 median error for pose estimation and 0.592 top-1-accuracy for category agnostic 3D object retrieval on the Pascal3D+dataset.


Lead Software Developer

New York Office

May 2016 – Jul 2018 IIT Kanpur

Objective: Industrial grade deployment of ML backend and android application for NYO

  • Lead a team of 16 people at NYO.

  • ML systems:

    • Collaborative Filtering for Recommendation engine
    • Automated response collection from handwritten characters markings on response sheets for grading
    • NLU chatbot using RASA pipeline with NER, Relationship extraction and quantity association for extracting required information from user.
  • Android app: REST APIs, SSE notifications, app-caching, Continuous integration with Jenkins, data and property binding, Reactive Java for observables

Recent Posts

Biologically-Plausible Learning Algorithms Can Scale to Large Datasets by Will Xiao[1], Honglin Chen[2], Qianli Liao[2] and Tomaso …