Hidden Progress in Deep Learning: SGD Learns Parities Near the Computational Limit.
Boaz BarakBenjamin L. EdelmanSurbhi GoelSham M. KakadeEran MalachCyril ZhangPublished in: NeurIPS (2022)
Keyphrases
- deep learning
- noise tolerant
- uniform distribution
- unsupervised learning
- boolean functions
- machine learning
- mental models
- weakly supervised
- agnostic learning
- membership queries
- decision lists
- pac learning
- higher order
- markov random field
- text classification
- graph cuts
- multi class
- training set
- object recognition
- data sets