Performance Prediction via Bayesian Matrix Factorisation for Multilingual Natural Language Processing Tasks
Monday, May 22, 2023
Performance prediction for Natural Language Processing (NLP) seeks to reduce the experimental burden resulting from the myriad of different evaluation scenarios, e.g., the combination of languages used in multilingual transfer. In this work, we explore the framework of Bayesian matrix factorisation for performance prediction, as many experimental settings in NLP can be naturally represented in matrix format. Our approach outperforms the state-of-the-art in several NLP benchmarks, including machine translation and cross-lingual entity linking. Furthermore, it also avoids hyperparameter tuning and is able to provide uncertainty estimates over predictions.
Post Talk Link: Click Here
Passcode: D%jGQ2z0
Speaker/s
Viktoria Schram is currently working towards her PhD at the School of Computing and Information Systems at the University of Melbourne, Australia. Prior to this, she has worked for 2 years on a research project studying systems design for THz communications. She has received a Bachelor of Science in Engineering and Management and a Master of Science in Electrical Engineering from FAU Erlangen-Nuremberg, Germany in 2015 and 2018, respectively. Her current research interests include statistical machine learning and probabilistic machine learning applied to performance prediction in NLP.
Related
Polygenic Score Modeling to Investigate Genotype-Phenotype Associations
Read More
Trustworthy Machine Learning: Transparency, Collaboration, and Evaluation
Read More