A step towards quantifying when an algorithm can and cannot predict an individual's wellbeing.
Orianna DeMasiBenjamin RechtPublished in: UbiComp/ISWC Adjunct (2017)
Keyphrases
- experimental evaluation
- learning algorithm
- computational complexity
- times faster
- np hard
- linear programming
- computational cost
- cost function
- optimal solution
- k means
- search space
- preprocessing
- matching algorithm
- experimental study
- estimation algorithm
- selection algorithm
- single pass
- recognition algorithm
- segmentation method
- clustering method
- segmentation algorithm
- theoretical analysis
- computationally efficient
- graph cuts
- expectation maximization
- particle swarm optimization
- simulated annealing
- worst case
- probabilistic model
- hidden markov models
- multi objective
- objective function