Congratulations to Dr. Maolin WANG of the Department of Data Science on receiving one prestigious honour: being named Runner-Up for the 2025 International Neural Network Society (INNS) Doctoral Dissertation Award for his outstanding doctoral dissertation Tensor Criterion-Based Neural Networks.
Dr. WANG's research pioneers a new paradigm for neural architecture design, Tensor Criterion-Based Neural Networks (TC-NNs), which systematically derives neural architectures from tensor mathematical principles rather than empirical engineering. While modern deep learning has achieved remarkable empirical success, many state-of-the-art models remain black boxes whose designs rely on heuristic trial-and-error, limiting interpretability, efficiency, and theoretical understanding. By establishing tensor criteria that encode desired properties such as multilinearity preservation, structural decomposition, and symmetry constraints, this thesis demonstrates how neural networks can serve as principled computational realizations of tensor operations. The resulting framework yields architectures that are simultaneously more expressive, more efficient, and more theoretically grounded, offering new perspectives for designing reliable and sustainable AI systems.The INNS award will be presented at the IEEE World Congress on Computational Intelligence (WCCI) 2026 in Maastricht, Netherlands, on 25 June 2026.
Tensor Criterion-Based Neural Networks (TC-NNs): research overview
Congratulations to Dr. WANG on international recognition and college honor and on his remarkable achievements in advancing the mathematical foundations of neural network design!