The mission of MLCommons® is to accelerate machine learning innovation and increase its positive impact on society. Together with its 50+ founding Members and Affiliates, including startups, leading companies, academics, and non-profits from around the globe, MLCommons will help grow machine learning from a research field into a mature industry through benchmarks, public datasets and best practices.
Every major technological advance follows a similar trajectory towards universal adoption and impact. The arc from research to broad accessibility generally takes from 30-40 years: from early automobiles to the family car, from development of ARPANET to the mainstream World Wide Web, from the first cellular phones to an smartphone in every pocket. Each of these examples started with technological breakthroughs, but for decades was limited by expertise, access, and expense.
Machine learning is no different. ML and artificial intelligence have been around for decades, but even today the technology is fragmented, bespoke, and poorly understood. We believe that we can unlock the next stage of AI/ML adoption by creating useful measures of quality and performance, large scale open data sets, and common development practices and resources. These three pillars will help to democratize machine learning and enable its widespread incorporation into new products and services in fields such as health, safety, and communication that deliver benefits to everyone.
MLCommons operates on the basis of five key principles, outlined below:
Grow ML markets and make the world a better place
Get everyone involved
- Be global, inclusive, and fair
- Bring together academia, small companies, large companies, non-profits, etc.
- Make it easy to get involved
- Be as open with our IP as possible while sustaining the community
Act through collaborative engineering
- Keep leadership mostly technical, with an emphasis on hands-on-involvement
- Favor data-driven decisions, design simplicity, and focus on real user value
Make fast but consensus-supported decisions
- Very low barrier for “experimental” working groups with well reviewed path to full endorsement
- Favor grudging consensus over 51/49 votes, especially for big decisions
- Make technical contributions easy
- Favor rapid development and iteration
Build a community that people want to be part of
- Be welcoming, informal, and friendly
- Encourage, recognize, and reward contributions
- Celebrate with cake