The goal for this study was to investigate the use of Sparse Topic Modelling for recommending unseen scientific articles of interest to users, based on what the users have liked in the past, and also the content of those articles. These approaches have shown promise to replace traditional topic modelling based approaches like LDA, for similar type of tasks. In this study, we investigated this for a specific of scientific article recomendation. - Joint work with Prof. Tina Eliassi-Rad
To develop a novel state-of-the-art detection and tracking framework for rear-side approaching cars. We employ various machine learning techniques to detect and then track rear side approaching cars. The software implementation takes advantage of GPUs computational capabilities for real-time performance. The objective is to later use this information to warn the biker of incoming dangerous situation, if any, based on a 3-feet safety zone rule.
We describe a combinatorial approach for investigating properties of rational numbers. The overall approach rests on structural bijections between rational numbers and familiar combinatorial objects, namely rooted trees. We emphasize that such mappings achieve much more than enumeration of rooted trees. We discuss two related structural bijections. The first corresponds to a bijective map between integers and rooted trees. The first bijection also suggests a new algorithm for sifting primes. The second bijection extends the first one in order to map rational numbers to a family of rooted trees. The second bijection suggests a new combinatorial construction for generating reduced rational numbers, thereby producing refinements of the output of the Wilf-Calkin[1] Algorithm. Link: - arXiv:1201.1936
Implementation of probabilistic Latent Semantic Analysis(pLSA) model for movie recommendation. Study of various collaborative fil- tering recommendation systems and implementation of an improved pLSA model for movie recommendation.