Improving Structured Data Ml Processes with Generative Adversarial Networks
Improving Structured Data Ml Processes with Generative Adversarial Networks

Abstract: 

Structured data is everywhere across enterprises and industries delivering significant business value by powering insights and actions derived through machine learning. Structured data comes in several forms and levels of complexity, like Flat tabular data (e.g. customer profile, dimension tables), Time series data (e.g. weekly sales, monthly promotional activity, instrument telemetry), Panel data (e.g. customer responses over time, geographic media spend) and Transactional data (e.g. administrative claims, EHR, website click stream) to name a few. Along with significant utility, there are also several challenges associated with this data that require companies to invest heavily to address, like partial data capture, missing values, rare occurrence of events of interest, limited labelled data, small data, etc. Data scientists have to come up with several custom pipeline adjustments iteratively to address these problems which is both time consuming, tedious and error prone. Generative Adversarial Networks (GANs) have had a lot of success in addressing these challenges for unstructured data but their application for structured data has been less explored. With the recent developments of its application for tabular data, GANs can act as a scalable approach to address these issues. In this session, we show how GANs can be used to augment structured data and thereby improve the structured data machine learning processes across datasets and domains. Examples include tasks like imputing missing data, over-sampling minority class for imbalanced data, creating synthetic data to improve training data and semi-supervized learning to leverage unlabeled data.

Bio: 

Srinivas leads ZS AI Research Lab with a focus on frontier innovation and development of cutting edge algorithms. Srinivas’s core expertise areas include automated machine learning, natural language processing, and marketing AI across industries. He has authored several thought leadership articles and presented at conferences. Prior to joining ZS, Srinivas spent time as a solution architect building expert systems to automate product design and manufacturing across multiple industries viz., automobile, power systems, medical devices and retail

Open Data Science

Open Data Science
Innovation Center
101 Main St
Cambridge, MA 02142
info@odsc.com

Privacy Settings
We use cookies to enhance your experience while using our website. If you are using our Services via a browser you can restrict, block or remove cookies through your web browser settings. We also use content and scripts from third parties that may use tracking technologies. You can selectively provide your consent below to allow such third party embeds. For complete information about the cookies we use, data we collect and how we process them, please check our Privacy Policy
Youtube
Consent to display content from Youtube
Vimeo
Consent to display content from Vimeo
Google Maps
Consent to display content from Google