Halvtidsseminarum av Aleksandr Karakulev: “Learning from Imperfect Data at Scale: A Bayesian Approach”

Välkommen till ett halvtidsseminarium i beräkningsvetenskap presenterat av Aleksandr Karakulev. Seminariet hålls på engelska.

Extern granskare: Ashkan Panahi (Chalmers)

Abstract: As modern machine learning models increasingly rely on large and complex data sets, dealing with data contamination — such as measurement errors, labeling mistakes, or adversarial inputs — has become a significant challenge. Manual cleaning is often impractical, and even small amounts of corrupted data can degrade performance. In this seminar, we present an adaptive method for robust learning, grounded in a Bayesian latent variable framework. The approach is broadly applicable through flexible likelihood modeling, resilient to various types of contamination, and does not require manual tuning. We demonstrate its utility across standard statistical tasks — regression, classification, and dimensionality reduction — and explore extensions to online learning, overparameterized/deep models, and federated learning, where it enables robust aggregation in the presence of heterogeneous and partially corrupted data.

Referens: För den som är intresserad av en mer detaljerad genomgång av metoden och dess tillämpningar hänvisas till den fullständiga artikeln: [2312.00585] Adaptive Robust Learning using Latent Bernoulli Variables