Sharad V. Oberoi, PhD

Visiting Scholar

The Robotics Institute

Carnegie Mellon University

Carnegie Mellon University (2007-present)

As a Visiting Scholar at The Robotics Institute, I am working on the following project:

ADEPT: Assessing Design Engineering Project Classes with Multi-Disciplinary Teams

During my PhD in Civil & Environmental Engineering at Carnegie Mellon, I have worked on a couple of related projects:

Group Cognition: Learning in engineering project teams

● I was the lead PhD student on this NSF-funded project to address the problem of information management in engineering design projects by using computational linguistics.

● I developed a research prototype to aggregate the documents and student discussion threads in project-based courses to extract an interactive concept-map based graphical representation (called DesignWebs) for navigating the project corpus.

● The research has implications for design learning by allowing researchers to understand how the underlying structure of the artifact changes as the project progresses.

DesignWebs: Navigable webs for design document navigation

● My contribution in this project relates to using noun phrases as a surrogate measure for assessing design team dynamics.

● This research shows that monitoring the design vocabulary over time as it is shared across team boundaries can reveal insights about the extent of successful collaboration among the team members.

● The research has implications for enhancing our understanding of how students create design knowledge and providing additional assessment metrics to instructors.

The University of Chicago (2010-11)

As an MA Candidate in the Social Sciences Division at UChicago, I was able to leverage my background in machine learning to address one of the hot research questions in international security studies:

“Has the population-centric counterinsurgency strategy been effective in stabilizing Iraq, and does it hold promise for Afghanistan?”

This formed the core of my MA thesis:

Population-centric counterinsurgency as seen through the Iraq War Logs

● This research builds an assessment model for measuring the success of population-centric counterinsurgency (COIN) in Iraq by using machine learning approaches on the Iraq War Logs.

● The analysis demonstrates that the COIN campaign did not succeed in achieving its goals and the US led coalition forces should focus on population-related activities.

● The research has implications for the current US strategy in Afghanistan that relies on the Iraq War as a successful counterinsurgency model.

Last updated on April 16, 2012.

If you are having problems viewing this site, visit the text-only version here.