My research focuses on developing machine learning algorithms that learn from and subsequently aid human design and creativity; a mixture of topics that I call Design Informatics. This involves using a combination of Artificial Intelligence, Machine Learning, Computational Linguistics, Ethnography, Human-Computer Interaction, Social Science, and Crowdsourcing techniques to analyze and build web-based software tools for designers on top of scalable machine learning systems.

Below are some research projects and publications that I have done in these areas. I believe in reproducible and open-source science whenever possible, and do my part by making most of my research code available on my GitHub account. You can also check out my information on Google Scholar for various citation information

Current Projects

Inferring Creativity

Measuring design creativity is crucial to evaluating the effectiveness of idea generation methods. This research project uses probabilistic models to create a family of repeatable creativity metrics that can be trained on expert data. It can be applied across domains, and provides a systematic way for researchers to explore what makes a design creative.
Highlight Related publications

Supply Chain Sustainability

In complex designs, we often have to make decisions based on imperfect information - maybe the data is unreliable or even missing. Can we leverage related design information to help us fill in the gaps? This research project on sustainable supply chain design uses a special form of uncertainty propogation to borrow information from related supply chains to update environmental, social, and economic indicators for a desired supply chain.
Highlight Related publications, Demo

Past Projects

Design Metaphors

Metaphors are a rich method for helping designer understanding problems and share a common vision. However, coming up with good metaphors can be a difficult task. This project explores how computational tools can help designers come up with better metaphors. Our tool visualizes a metaphor database, helping teams explore possible options for their design.
Meta4Explorer Tool

3D CAD Interfaces

Specifying complex 3D shapes using traditional 2D interfaces can be time consuming and unintuitive. This project explored this problem by proposing a new kind of 3D user interface that would allow a designer to quickly "sketch" in 3D, using their hands in an augmented reality environment. It allows a user to quickly prototype complex surfaces by sketching curves or point clouds in 3D and then using those points to fit parametric surfaces.
Highlight Related publications

Human Sketch Understanding

This project explored how people perceive and make sense of engineering diagrams. It used a tablet-based interface that allowed a user to slowly uncover portions of a diagram, tracking the regions of interest. By recording and analysing the tracking data generated by individuals, we can inform new models of sketch understanding.
Highlight Related publications

Papers in Preparation

Journals

Peer-Reviewed Conferences

Workshop Papers

Working Papers or Technical Reports