Fotografi av Amanda Olmin

Amanda Olmin

Doktorand

I min forskning utvecklar jag algoritmer som kombinerar djupinlärning med probabilistisk modellering. Målet är att ta fram metoder som kan modellera komplexa datamängder och samtidigt hantera det brus som finns i majoriteten av data.

Publikationer

2024

Amanda Olmin, Jakob Lindqvist, Lennart Svensson, Fredrik Lindsten (2024) On the connection between Noise-Contrastive Estimation and Contrastive Divergence

2023

Jakob Lindqvist, Amanda Olmin, Lennart Svensson, Fredrik Lindsten (2023) Generalised Active Learning With Annotation Quality Selection IEEE 33rd International Workshop on Machine Learning for Signal Processing (MLSP) Vidare till DOI
Amanda Olmin, Jakob Lindqvist, Lennart Svensson, Fredrik Lindsten (2023) Active Learning with Weak Supervision for Gaussian Processes Neural Information Processing 29th International Conference, ICONIP 2022, Virtual Event, November 22–26, 2022, Proceedings, Part V, s. 195-204 Vidare till DOI

2022

Amanda Olmin, Fredrik Lindsten (2022) Robustness and Reliability When Training With Noisy Labels Proceedings of the 25th International Conference on Artificial Intelligence and Statistics (AISTATS) 2022, s. 922-942
Amanda Olmin (2022) On Uncertainty Quantification in Neural Networks: Ensemble Distillation and Weak Supervision

Forskning

Organisation

Kollegor vid STIMA