Projects

Speech Recognition and classification for Hindi dialects

Jan 2018 - Dec 2018

The aim of this projects is to study different Hindi dialects. We focus to develop a system, recognizing and transcribing hindi speech and it’s dialects. This project is ongoing under the supervision of Prof. Kamlesh Tiwari at AIMA Labs, BITS Pilani.

Path Finding Agent

Sep 2017 - Sep 2017

Heuristic-based online terrain search aimed at finding a near-optimal solution of path cost while minimizing the computational cost using Uninformed and Informed search.

Sarcasm Detection on Twitter dataset

Feb 2017 - April 2017

Implemented a paper on Sarcasm Detection by A. Rajadesingan. Dataset was mined using Twitter API. NLP was scripted in Python using nltk. Made some improvements and tweaks in features for the better results. Applied various classification algorithms like Ensemble, Naive Bayes classifier, SVM, Random Forest, Adaboost, Gradient Boosting, etc.
Gradient Boosting achieved an accuracy of 85.71% from a feature set of only 1\7^th the original size proposed in the paper.

Code Report

Energy Efficiency of Residential Building Design

Jan 2017 - March 2017

Developed an Android application to predict the dependence of Energy Efficiency, of a residential building in a particular area, based on its structure and design. The energy efficiency is calculated in terms of the Heating Load and Cooling Load of the building which can further help to modify the building design for achieving maximum possible energy conservation in future. Achieved an accuracy of 98% using Random Forest algorithm.

Code Report

Fan Control System using 8086 Micro-Processor

April 2017

Designed a hardware system on ISIS 2007, Proteus to simulate the speed control and time settings for a DC Motor Fan using an 8086 Micro-Processor. An assembly script was used to run the model.

Code Report

Mini Projects and Academic Activities

• Implemented SkipGram model of word2vec in tensorflow with visualization of embedding in tensorboard.
• Implemented a variant of the DBSCAN clustering algorithm written in the functional programming paradigm.
• Implemented a Path Finding Agent using Informed and Uniformed search techniques.
• Wrote a Critique and an executive summary on “Rank-indexed hashing: A compact construction of bloom filters and variants” by N. Hua, H. C. Zhao, B. Lin, and J. Xu.
• Gave a Lecture on Introduction to Machine learning in layman terms as part of a Workshop organized by BITS-ACM during the college Technical Festival APOGEE BITS-PILANI.
• Successfully conducted a Machine Learning hackathon on kaggle in collaboration with MapMyIndia during the college Technical Festival APOGEE BITS-PILANI.
• Run a technical Blog wherein members can share and learn about ongoing technology.
• Acted as Moderator and Question setter in a Python competition organized by BITS-ACM to enhance the coding culture of the college campus.