Unsupervised Latent Debiasing of Time-Series Models
Name
Phillips-jdp99-meng-eecs-2022-thesis.pdf
Description
Thesis PDF
Size
5.33 MB
Format
Adobe PDF
Checksum (MD5)
e1cc266562394705c81061f803c47b70
Author(s)
Phillips, Jacob
Advisor(s)
Rus, Daniela L.
Date Issued
February 2022
Publisher
Massachusetts Institute of Technology
Abstract
Traditional training regimens for time-series models have been shown to encode the biases from their training corpora into the models themselves. We aim to train unbiased time-series models using existing biased datasets. However, most debiasing techniques rely on explicit labels that encapsulate the bias, such as pairs of words along some worrying axis of bias such as race or gender for language models. We propose an unsupervised latent debiasing training regimen based on [2] that simultaneously learns the latent distribution of the dataset and a separate language task; datapoints are selected for training batches by sampling weights inverse to their commonality as determined by their placement in the latent space. We adapt [2] to time-series datasets and show algorithmic improvements to bias identification and bias reduction for models trained on toy and real datasets.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
In Copyright - Educational Use Permitted
Copyright MIT
Persistent DSpace Link
