Reducing Data Complexity Using Autoencoders With Class-Informed Loss Functions.
David CharteFrancisco CharteFrancisco HerreraPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2022)
Keyphrases
- loss function
- data complexity
- expressive power
- reproducing kernel hilbert space
- pairwise
- loss minimization
- query evaluation
- query answering
- conjunctive queries
- learning to rank
- support vector
- np complete
- risk minimization
- convex loss functions
- bregman divergences
- query complexity
- denoising
- tuple generating dependencies
- data sets
- dl lite
- relational databases
- databases