Multiple Descents in Unsupervised Learning: The Role of Noise, Domain Shift and Anomalies.

Kobi Rahimi Tom Tirer Ofir Lindenbaum

Published in: CoRR (2024)

Keyphrases

unsupervised learning
domain specific
missing data
random noise
computer vision
supervised learning
object recognition
anomaly detection
multiple domains
database
data sets
hidden markov models
dimensionality reduction
signal to noise ratio
gaussian noise
noise model