Ying Feng

Ying Feng is a graduate student in electrical engineering and computer science combining theoretical rigor with practical algorithm design for high-dimensional data using randomized and geometric techniques. Existing methods for low-distortion embeddings and approximate distance computation face challenges balancing theoretical guarantees with computational efficiency. She develops fast and compact random mappings achieving best-known dimension bounds while maintaining near-linear evaluation time through structured transforms like randomized Hadamard matrices. Using MATLAB, Ying empirically validates theoretical insights by generating random operators and computing norms to test conjectures before rigorous proof. She improves algorithms for the chamfer distance, reducing computational complexity from O(n log n) to O(n log log n) by exploiting geometric structure. Ying also advances vector quantization via randomized Hadamard transforms with dithering, proving mean-squared error bounds matching dense random rotations. As a MathWorks Fellow, Ying will extend these algorithmic frameworks. Her work could establish new foundations for efficient computation in machine learning systems.