University of Hong Kong
Department of Architecture
KB709
kmmt@hku.hk

Prof. Kam-Ming Mark Tam | 譚錦明

director

Assistant Professor Of Architectural Structures

Kam-Ming Mark Tam is Assistant Professor of Architectural Structures at the University of Hong Kong. He develops computational design strategies for translating design goals into solutions through the advancement and integration of computational, structural design, and modelling methods, as well as digital manufacturing techniques. His specific expertise encompasses computational design modelling, architectural and structural geometry, numerical and geometrical form-finding algorithms, optimisation methods, additive manufacturing (AM), and Machine Learning (ML) techniques.  This includes specialisations in Geometric Deep Learning (GDL) and Deep Reinforcement Learning (DRL) for learning and modelling discrete systems and decision processes.

Tam recently completed his doctoral research at ETH Zürich, where he conducted research at the Block Research Group co-led by Prof. Philippe Block and Dr. Tom Van Mele. His thesis, titled ‘Learning Discrete Equilibrium: Trans-topological Inverse Pattern and Force Design Using Machine Learning & Automatic Differentiation,’ was nominated for the ETH medal by its committee of examiners, which includes experts from ML and/or structural engineering. These include Prof. Andreas Krause (Chair of the ETH AI Center), Robert Otani (Chief Technology Officer at Thornton Tomasetti), and Prof. Nathan Brown (Professor of Architectural Engineering from PennState).

Specifically, the doctoral research contributed novel GDL methodologies for versatile learning and modelling of hierarchical discrete data structures like meshes for representing reticulated systems.  The novel framework provided the basis for surrogate models to infer performance attributes from designs, inverse models for engineering designs according to target goals, and DRL-based decision models for the effective deployment of bespoke design processes aligned with diverse performance goals. His doctoral research, partly funded by scholarships from ETH’s Institute of Technology in Architecture and the Swiss National Science Foundation, included an exchange to Imperial College London (UK), where he received guidance from the renowned GDL luminary, Prof. Michael Bronstein. Additionally, he served as a research scientist at the L3S Center (Germany), working with Dr. Daniel Kudenko, an established DRL expert, and supervised doctoral research in ML.

Tam has taught in various capacities at Singapore University of Technology and Design, Pratt Institute, Tongji University, ETH, Massachusetts Institute of Technology (MIT), and University of Waterloo (UW). Prior to his doctorate, Tam earned a Master of Engineering (MEng) from MIT, and a Master of Architecture (MArch) with Structural Engineering Certification and an Honours Bachelor of Architectural Studies (HBAS) with Economics Minors from UW. Tam’s work has been presented at a number of conferences, including the International Association for Shell and Spatial Structures (IASS), the Association for Computer Aided Design in Architecture, ROB|ARCH, and Design Computing and Cognition.

At MIT, he developed robotic-enabled AM techniques and fabrication-aware structural optimisation methods for designing AM-produced structures under the tutelage of Prof. Caitlin Mueller. A key contribution was his work on Stress Line Additive Manufacturing, which earned the Tsuboi prize for the best paper presented at the 2015 IASS conference. Professionally, Tam played a key role in collaborations with AM startup Branch Technology during his tenure as Integration Engineer at Thornton Tomasetti’s CORE Studio. His efforts were crucial to the realisation of OneC1TY located in Nashville, Tennessee, one of the world’s longest-spanning AM-produced lattice pavilions.