Cognitive Technical Support     


Nishtha Madaan photo Vikas Agarwal photo photo photo photo photo

Cognitive Technical Support - overview


The Cognitive Technical Support department is focused on reimagining the world of IT services entrenched by AI, deep learning and cognitive technologies.  Our goal is to infuse cognitive technologies powered by Watson into the domain of IT services that are delivered via a mix of private, public and hybrid cloud based technologies.  We develop AI solutions and services intended for customer helpdesk operations, troubleshoot and triaging as well as preventive maintenance.

What comprises IT Support?

IT support is needed across the spectrum of service management beginning with customer help-desk where the problem is first reported, to the technical support agents who troubleshoot and triage to keep the infrastructure/applications up and running, all the way to the autonomous self-monitoring IT systems that can detect problems and self-heal to prevent future problems.  Traditionally this industry has been extremely human intensive mainly due to the complexity of tasks involved in maintaining applications and IT devices (servers, storage, network etc.).  However, the complexity of IT systems, their sheer scale and the large combinations of operating systems, middleware and applications that comprise any given managed system therein make automated service management a very hard research problem.

How can Cognitive Technologies Help?

A cognitive IT support system is defined as one that: (a) can diagnose problems related to  the IT entities  when they occur or can predict them before they will occur, (b) discover root case analysis of the problem and can resolve them automatically and (c) can learn and improve its understanding and resolution over time. Our aim is to build such a support system that has the following components:

  1. Mines  terabytes of operational day that are generated by IT entities every day to build machine learning models that detect patterns of failures
  2. Mine human conversations to correlate the learnt patterns to the operational data , the diagnosis of the root cause and finally to resolution actions
  3. Codify and store all the knowledge, patterns and insights in the form of IT knowledge graphs
  4. Create cognitive assists that use natural language understanding/AI and the codified knowledge graphs to aid the tasks of all human touch points in the system.

Team Updates

  • Congrats!Citicafe: Interactive Platform for Citizen Engagement” by Shubham, Pooja, Prateeti, Amol, Anwesh, and Gargi has been accepted at ACM IUI 2018.
  • Congrats! “Anomaly Detection Using Program Control Flow Graph Mining from Execution Logs" selected for the IRL Distinguished Paper Award-2016. The paper was presented at KDD 2016 and includes Atri Mandal, Shubham Atreja and Gargi B. Dasgupta as authors.
  • Congrats! Anupama Ray has been selected to attend the Global Young Scientists Summit (GYSS) happening in Singapore from 21-26 January 2018.
  • Hima Patel chaired the AI Track of GHC India (2017) held on 16-17 November, 2017 in Bangalore.
  • Gargi B. Dasgupta participated in a Panel discussion on “Importance of Domain Knowledge in Building Intelligent Applications” as part of GHC India (2017) held on 16-17 November, 2017 in Bangalore.

Latest Blogs

Blog on internship experience
Shubham Atreja, Nov 11, 2017

Getting your github code CI-ready in a few easy steps with Travis
Atri Mandal, Aug 24, 2017

Creative AI: The Future - Part 2
Chandresh Maurya, June 5, 2017