Projects

Implementation of Papers:

  1. Bernardino Romera-Paredes, Philip H. S. Torr. “An embarrassingly simple approach to zero-shot learning”, (ICML 2015). [Code]
  2. Edgar Schönfeld, Sayna Ebrahimi, Samarth Sinha, Trevor Darrell, Zeynep Akata. “Generalized Zero- and Few-Shot Learning via Aligned Variational Autoencoders”, (CVPR 2019). [Code]

Reinforcement-based robotic navigation in real-world

We create a partially observable environment for the agent(QuadPod) to perform its action in a real-world scenario.  We accomplished this by setting up a camera whose field of view consisted of the current position of the robot, its initial and destination points. Instead of consuming the entire frame(which will be dimensionally expensive) as the observation from the environment, we used only the information of the current position of the agent wrt to its initial and destination coordinates. We used the ϵ-greedy Q-learning with delayed reward to test the exploration of the agent, the agent was rewarded as it moved forward closer to the destination and punished when it moved backward away from the destination. 

[Code] [Video]

Customer Churn Prediction in Mobile networks

A simple classification model was built to identify the churning of customers in mobile networks. This project was completed under the guidance of Dr. Sharath Kumar from Reliance Jio. This resulted in a publication in the ICGCIoT, IEEE conference on green computing and internet of things 2018.

[Paper] [Code]

Interictal Spike Detection in EEG Signals

Interictal spikes present in EEG signals are the indicators to epilepsy. Given an EEG signal, we classified the spikes present in the signal using a simple SVM classifier. Prior to classification, thresholding was done using None linear energy operator and with the help of morphological characteristics of Spikes. This resulted in a publication in ICEECCOT, IEEE Conference on Electronics, Communication, Computer Technologie, and Optimization Techniques.

[Paper] [Code]