I have previously announced in Q-AHA: Quantum Applications, Hardware and Algorithms:
The 2021 Tayur Prize for Numerical Performance of GAMA-inspired methods for Important/Imaginative Applications.
Here are the winners (thanks to IIT-Madras faculty Prabha Mandayam and Anil Prabhakar, who mentored the undergraduate students in the course ID5840 that mirrors the (joint Tepper-ECE, in collaboration with USRA/NASA and Amazon Braket) CMU course Quantum Integer Programming and Quantum Machine Learning, and judges David Bernal, Maximiliano Stock, Davide Venturelli, Elias Towe and Peter McMahon):
First Prize: Binary Classification (Apurva Padhye, Sai Sakunthala).
Examining X-ray chest data of lungs to detect pneumonia early is crucial in increasing the chances of survival (and reduce recovery time). Rather than use Convolutional Neural Network (CNN), which is the most common method for image recognition – and requires large computational effort and may over-fit the data – they modeled instead via the use of Support Vector Machine (SVM) that is not only simpler but also more directly convertible to Quadratic Unconstrained Binary Optimization (QUBO) formulation that is. Their implementation – on real pneumonia data – showed that Graver Augmented Multi-used Algorithm (GAMA) classifies with very high accuracy in a about 40 seconds (compared to thirty minutes using commercial state-of-the-art software for the same accuracy, and speeding up classical methods via decomposition provides a 15x speedup but at an unacceptable drop in accuracy).
Runner-up: Quadratic Knapsack Problem (Neelkanth Rawat, Shivaprasad Hulyal).
QKP is a fundamental model that is useful in Operations Research and Statistics, generalizing many graph theoretic formulations and capturing the essence of the computational difficulty in several important applications. Their implementation demonstrates that GAMA can match the commercially best solvers in accuracy, and importantly, without the rapid increase in time as the density of graph increases.
Coincidentally, this past week Tinglong and I also had our invited paper accepted for publication, where we discuss what it would take for AI-based solutions (such as the First Prize winner above) to become ubiquitous and what our OM research community can contribute towards that goal:
What about the 2022 Tayur Prize? Moving the focus from Ising model (see 2020 Tayur Prize) to Gate (circuit) model of Quantum Computing (you may enjoy the Knuth-Bendix approach to compiling that I described in Language of the Gods in the World of Men):
For innovative accommodation of noise in compiling or error correction towards the development of robust quantum algorithms.