A linearly convergent algorithm without prior knowledge of operator norms for solving ℓ1-ℓ2 minimization.
Yaru ZhuangHaitao CheHaibin ChenPublished in: Appl. Math. Lett. (2022)
Keyphrases
- prior knowledge
- improved algorithm
- dynamic programming
- objective function
- k means
- computationally efficient
- high accuracy
- learning algorithm
- matching algorithm
- optimization algorithm
- np hard
- computational complexity
- times faster
- worst case
- convergence rate
- image segmentation
- tree structure
- experimental evaluation
- clustering method
- theoretical analysis
- preprocessing
- optimal solution
- similarity measure
- cost function
- multi objective
- energy function
- search space
- prior information
- reinforcement learning
- recognition algorithm
- algebraic equations