CEE 200 Sec 1 Seminar: Newell Washburn

Speaker: Newell Washburn
Affiliation: Carnegie Mellon University

UCLA Civil & Environmental Engineering Department

 

C&EE 200 Section 1 Seminar

Structural, Geotechnical and Civil Engineering Materials

 Designing Superplasticizers with Hierarchical Machine Learning


Newell Washburn, Ph.D.


Associate Professor, Carnegie Mellon University

 

Superplasticizers are a class of chemical admixture that improve the workability of cement paste while reducing water requirements. However, new cements having lower carbon footprints and new processing methods for cement and concrete are placing greater demands on performance enhancement by polymeric dispersants. The complexity of this problem makes purely experimental or theoretical approaches challenging, but current machine learning tools have data requirements that cannot be met given the intrinsic variability of these materials. We developed hierarchical machine learning (HML) as a methodology for understanding, designing, and optimizing complex physical systems. This talk will discuss the development of the algorithm and its application to designing superplasticizers. Current studies incorporate composition and particle characteristics as explicit parameters in the model, including interactions between superplasticizers and sulfate or metakaolin. The broader goal in this research is to design cementitious materials for function instead of composition, aligning cement with the broader goals of the Materials Genome Initiative. The talk will close with perspectives on transfer learning with the HML algorithm and extending the application of HML to other material infrastructure systems.

Where: 4275 Boelter Hall

When: 4:00 – 5:00 PM on Monday, February 12, 2018

 Newell Washburn received his BS in Chemistry from University of Illinois at Urbana-Champaign and his PhD in Chemistry from University of California (Berkeley). Following post-doctoral work in Chemical Engineering at the University of Minnesota (Twin Cities), he joined the Polymers Division at NIST, first as an NRC post-doctoral associate then as leader of the Biomaterials Group. In 2004, he moved to Carnegie Mellon University where he is currently Associate Professor of Chemistry, Biomedical Engineering, and Materials Science and Engineering (by courtesy). His research interests focus on the development of materials for a diversity of technological applications primarily involving polymers or biomaterials.  With colleagues at CMU, he recently led the development of an AI algorithm called hierarchical machine learning, which is designed to leverage understanding of complex physical systems for analysis of sparse datasets. They are currently applying this to design of materials, formulations, and additive manufacturing processes.

 

Date/Time:
Date(s) - Feb 12, 2018
4:00 pm - 5:00 pm

Location:
Boelter Hall 4275
4275 Boelter Hall Los Angeles CA 90095