How can we learn (more) from challenges? A statistical approach to driving future algorithm development.
Tobias RoßPierangela BrunoAnnika ReinkeManuel WiesenfarthLisa KoeppelPeter M. FullBünyamin PekdemirPatrick GodauDarya TrofimovaFabian IsenseeSara MocciaFrancesco CalimeriBeat P. Müller-StichAnnette Kopp-SchneiderLena Maier-HeinPublished in: CoRR (2021)
Keyphrases
- times faster
- detection algorithm
- theoretical analysis
- high accuracy
- convex hull
- experimental evaluation
- clustering method
- segmentation algorithm
- computationally efficient
- linear programming
- computational cost
- long term
- recognition algorithm
- optimal solution
- np hard
- learning algorithm
- optimization algorithm
- similarity measure
- particle swarm optimization
- input data
- computational complexity
- simulated annealing
- preprocessing
- improved algorithm
- hypothesis testing
- future trends
- convergence rate
- real world
- path planning
- worst case
- dynamic programming
- cost function
- k means
- search space
- objective function
- machine learning