In today's world of digitization, personalization and globalization, each organization collects vast amounts of information about its own operations. This data is often collected in many different repositories, and stored in various forms--from structured information stored in databases to unstructured textual, imagery and video information stored in files). This data contains a lot of hidden information and insights about enterprise's operations. By extracting useful and actionable knowledge from these repositories and suggesting concrete transformation opportunities, one can dramatically improve an enterprise' operational efficiency.
The Machine Learning group @ SRL focuses on developing and applying various data and text mining techniques to derive actionable insights and knowledge from enterprise information repositories. Our experience with industrial-scale problems has indicated that while the literature contains diverse set of techniques for text and data mining, most of them work well only under certain circumstances (or in controlled environments) and only when parameterized appropriately. Unfortunately, the parameterization is generally specific to each problem type as well as the dataset. Today, customization of these techniques for each problem type and data set is a very manual exercise.
Thus, our mission is to develop robust and automated solutions for deriving actionable insights from enterprise repositories. We are developing custom solutions for several problem verticals commonly found in the context of IT services industry. For each problem vertical, our goal is to develop a solution that is robust to different data sets. Our approach involves two key features: (1) combine text mining and data mining effectively, and (2) perform cross-silo (or across data repositories) analysis to derive hidden and actionable insights. Currently, we are developing solutions for the following two problem verticals:

  1. Driving continuous improvement for operations support activities; and
  2. Optimizing workforce planning and management for the services industry.