To constantly break through himself in the university, Guanzhong Yang embraces the new topics, faces the brand-new challenges, cooperates with ambitious colleagues, and pursues the life-long dreams.
In January 2026, Guanzhong Yang and his teammates completed a summary poster on Bayesian inference and modern MCMC sampling, covering both the mathematical ideas and coding implementation behind methods such as Metropolis–Hastings, Hamiltonian Monte Carlo (HMC), and NUTS. As part of the project, they also explored the theory of posterior sampling, built computational visualisations, and implemented practical examples, including a real-world case study related to Bayesian trajectory control for a robotic manipulator.
Since May 2026, Guanzhong Yang has participated in the Statistics and Data Science of MITx MicroMasters Programs, where he learned and applied graduate-level knowledge in statistical modelling and machine learning. During the programs, he completed many hand-on projects, such as implementing a neural network for handwritten digit recognition on MNIST, and designing a graph network analysis on the CAVIAR criminal network dataset using centrality measures.
In June 2026, interested in the application of linear algebra in algorithm, Guanzhong Yang improved an existing image-denoising algorithm by considering singular value decomposition and Stein’s Unbiased risk estimate. He proved that the method he proposed is optimal in theory, and validated the result through Monte Carlo and image-denoising experiments.
Each paper and conference record acts more than an academic entry. They mark moments where curiosity met rigor, and where solitary inquiry converged into shared insight.