Challenge

Benchmarking of AI system capabilities

Developing and utilizing good benchmarks for measuring algorithmic progress and objectively comparing the performance of different algorithms.

Impact: High 

Addressed by:

Addresses:

Categories

One Reply to “Benchmarking of AI system capabilities”

  1. Note that there are potentially many sub-challenges for benchmarking specific capabilities ind specific application areas, but we will not list them explicitly. Rather, the availability of benchmarks for different application areas is document through ‘metrics’.

Comments are closed.