Authors: Will Jones (convenor, Best Practice work stream) and Jeremy Bennett (Co Chair TechNES AI/ML group)

We have seen that there is a gap between the marketing vision for what AI/ML can achieve and the actual technology being developed in universities and corporate research departments around the world.

Invariably it is the professional engineer working in industry who is expected to bridge that gap.

The TechNES AI/ML group aims to bridge that gap. Our goal is to provide materials to help professional engineers to use and develop practical AI/ML systems. Such engineers are often highly skilled, but relatively few have practical experience of AI/ML. In developing these materials we will build on existing resources where they exist and then draw on the experience of our Universities and those in industry already working in AI/ML.

Our work focuses on three work streams: i) best practice for industrial AI/ML; ii) AI/ML in development tool flows; and iii) AI/ML system design and development. Will Jones and I are leading the first of these groups, and the rest of this blog post looks at our plans.

The Best Practices stream is focused on providing “Best Practice”

guidance for AI, machine learning, and data technology to non-expert engineers and decision makers. This stream aims to fill a gap in the market by providing guidance that is practical, actionable, and, perhaps most importantly, free.

The scope of the group not just to help users understand how to robustly and fully advantage these technologies, but to understand what these technologies can do and where it is appropriate to apply them.

Furthermore, an important part of the group’s work is recognizing that the scope of AI technology reaches beyond the immediate technical challenges of it’s realization. Issues that fall under the label of “trustworthiness”, such as the challenges of bias, privacy, transparency, must be at the forefront of ones mind when working with this technology. On a similar vein, the rise of lawmakers interest in this technology, legality and compliance are becoming increasingly relevant. With the introduction of GDPR and, particularly the upcoming EU Artificial Intelligence Act, something that everyone involved in producing this technology, from high level decision makers to engineers writing code need to be aware of.

The best practice work stream meets monthly via videoconference and quarterly in person.  We intend to produce our first set of materials in early 2023.  We welcome contribution and participation from any TechWorks member with expertise to share. Please reach out to either Jeremy Bennett, Will Jones or Steve Drew if you would like to join us.