Theory behind Image Compression and Semantic Search.

Accepted Session
Short Form
Scheduled: Thursday, June 22, 2017 from 10:00 – 10:45am in B304


This talk will focus on describing a matrix decomposition technique called Singular Value Decomposition that conveys important geometrical and theoretical insights about linear transformations. This technique is not as famous as it should be given the range of applications from science and engineering.


Singular Value Decomposition (SVD) is a matrix decomposition technique developed during the 18th century and has been in use ever since. SVD has applications in several areas including image processing, natural language processing (NLP), genomics, and data compression. In NLP context, SVD is called latent semantic indexing (LSI) and used for concept based search and topic modeling. In this talk, we will describe the math and intuition behind eigenvalues, eigenvectors and their relation to SVD. We will also discuss specific applications of SVD in image processing and NLP with examples. Python code snippets on performing SVD using open source libraries like numpy and scikit-learn will be shared.


Linear Algebra, machine learning, artificial intelligence, Natural Language Processing, image processing

Speaking experience

I have spoken at several Society for Industrial and Applied Math (SIAM) conferences in optimization and scientific computing areas. During my corporate days, I conducted 300 + hours of hands-on training classes. Nowadays, I frequently give talks on Artificial Intelligence and Machine Learning in PDX meetups. I spoke about IBM Watson in a Marketing Technology meetup:

I recently gave this talk at a data science meetup


  • Img 0381 (3)

    Santi Adavani



    Santi co-founded RocketML, where his team is building a superfast engine for building machine learning models. Before that, Santi worked as a Product manager and software development lead at Intel’s technology and manufacturing group. Prior to Intel, he got his Ph.D. in computational sciences from the University of Pennsylvania. His areas of expertise include high-performance computing, non-linear optimization, partial differential equations, machine learning, and big data.