Non-profit organization SPEC has formed a committee to oversee the development of an independent benchmark for artificial intelligence algorithms, machine learning and inference tasks.
Excited to announce the formation of #SPEC‘s new Machine Learning Committee. The ML committee will develop practical methodologies for #benchmarking #AI and #ML #performance in the context of real-world platforms and environments. Learn more at: https://t.co/5us6CDT1js pic.twitter.com/ucPWuPlHVC
— SPEC (@spec_perf) March 10, 2022
The committee will develop a practical methodology for comparative analysis of algorithm performance in real-world platforms and environments.
Initially, the committee will focus on developing benchmarks to measure the end-to-end performance of the tested hardware system when executing machine learning and inference tasks.
SPEC believes that creating an independent, vendor-agnostic benchmark test will allow developers to evaluate competing platforms and technologies and make more informed decisions about purchasing hardware and tools.
The committee has also invited AI experts to join the development and governance of the benchmark.
“I urge everyone interested in the future of MO processing to join SPEC ML’s committee and help shape these invaluable tests,” said Arthur Kang, the committee chair.
SPEC publishes a suite of benchmark tests that are widely used to assess computer system performance, particularly in the high-performance computing (HPC) sector.
Recall that in November 2021, UNESCO members adopted the first global agreement on AI ethics.
In June, the World Health Organization formulated six core principles for AI in medicine.
Follow ForkLog news on Telegram: ForkLog AI — all the news from the world of AI!