AiPEX Lab

Artificial Intelligence in Products Engineered for X (AiPEX) Lab

Our research explores the use of machine learning methods that predictively improve the outcome of product design solutions through the acquisition, fusion and mining of large-scale, publicly-available data. It has been reported that 70-80% of the costs of a product are determined during the design phase. Here, the term product is used in a general sense to refer to physical/digital systems that are guided by user needs, and that require domain knowledge to create. Towards enhancing the efficiency of the design process and creating personalized solutions, our research focuses on three thrust areas, outlined below.

Our lifelong research and education goals are tightly integrated and focus on advancing personalization, both from a design and learning perspective.

Faculty

Conrad Tucker

Conrad Tucker

Professor of Mechanical Engineering

Conrad Tucker, professor of mechanical engineering, leads the research group. He joins Carnegie Mellon University in Fall 2019 from the Pennsylvania State University, where he directed the Design Analysis Technology Advancement (D.A.T.A) Laboratory. His research focuses on the design and optimization of systems through the acquisition, integration, and mining of large scale, disparate data.

Email
conradt@andrew.cmu.edu
Google Scholar
Conrad Tucker

Research thrusts

  • Research Thrust 1: Product Feature Discovery and Quantification: research imageIt is often stated that product customization is limited by the difficulty of addressing individual needs and converting them into product specifications. This research thrust seeks to advance machine understanding of user needs through the mining of large scale, publicly-available data (e.g., social media networks such as Twitter) for scalable and efficient product customization.
  • Research Thrust 2: Conceptual Design Generation and Evaluation: This research thrust seeks to create scalable ways to generate personalized design concepts, based on individuals’ unique needs. The evaluation of generated designs is grounded in a fundamental understanding of our physical universe and the laws that govern it. We explore ways to teach machines how to learn the relationships between design concepts and the environments/constraints that they must operate in.research image
  • Research Thrust 3: Information Veracity and Security: Ascertaining the veracity and security of data in the information age is a challenge both for humans (e.g., communicating within social media networks) and machines (e.g., training data for artificial neural networks). A lack of data veracity has the potential to “fool” both machines, as well as humans into achieving unintended outcomes/output. We explore ways to ensure the veracity of data used to train machine learning models in order to avoid garbage-in, garbage-out outcomes during the design process.research image

 

Project videos

Research team

Frederica Free-Nelson

Frederica Free-Nelson

Visiting Researcher

Research interests
vehicular security, machine learning, and intrusion detection methods and techniques to promote cyber resilience and foster research on autonomous active cyber defense
Email
frederica.f.nelson.civ@mail.mil
James Cunnningham

James Cunningham

Doctorate

Research interests
Deep Learning/Reinforcement Learning
Email
jamescun@andrew.cmu.edu
Sakthi Prakash

Sakthi Prakash

Doctorate

Research interests
Machine Vision/Machine Learning
Email
sarulpra@andrew.cmu.edu
Sweta Priyadarshi

Sweta Priyadarshi

Doctorate

Research interests
Machine vision and deep learning
Email
swetap@andrew.cmu.edu
Dule Shu

Dule Shu

Doctorate

Research interests
Generative Design, Deep Learning
Email
dules@andrew.cmu.edu

Publications

Engineering and STEM

Peer Reviewed Journal Publications

Peer Reviewed Conference Publications

Health Care

Peer Reviewed Journal Publications

 

Peer Reviewed Conference Publications