Learning an Invertible Output Mapping Can Mitigate Simplicity Bias in Neural Networks.
Sravanti AddepalliAnshul NaseryR. Venkatesh BabuPraneeth NetrapalliPrateek JainPublished in: CoRR (2022)
Keyphrases
- neural network
- learning algorithm
- learning systems
- learning process
- online learning
- reinforcement learning
- multi agent
- fuzzy logic
- inductive inference
- cascade correlation
- decision trees
- recurrent networks
- spiking neural networks
- training process
- learning community
- neural nets
- incremental learning
- learning problems
- learning tasks
- background knowledge
- data sets
- supervised learning
- semi supervised
- pattern recognition