Industrial Data Symposium, brought to you by The Manufacturer, is the virtual version of the UK’s largest gathering of manufacturing data professionals.
Industrial Data Symposium takes place on 22nd April, bringing 100 data-minded manufacturers together for the fourth year to discuss the role of Big Data in their business. Manufacturers will learn the best practices for capturing, processing, and storing data and turning it into actions to create real business value. We caught up with Bala Amavasai – Lead AI Architect & Head of AI, Stanley, Black & Decker to preview the event and give you a flavour of what his session has in-store.
Can you give me some background and how you got to where you are now?
My background is in computer engineering, and I’ve got a PhD in machine learning from the 1990s, from the University of Sheffield, UK. So, you may look at me as old school! However, I still code and I still read all the recent papers. AI is an area which is evolving so hugely. If you do not keep up with the latest trends for one or two years, you’re no longer state-of-the-art.
I spent about 10 years in academia. I ran a research group on robotics and machine vision, so I have several PhD students who have completed under me, doing research in machine learning, robotics, image processing and machine vision.
I spent another 10 years at Procter and Gamble, where I took the leadership of the computer vision and sensors area, although I subsequently focused on my core interest which is artificial intelligence and data science. I was very pleased to be involved in the early strategy for the company. I worked on several interesting projects there which led to multiple patents. AI and data science now drive new smart product development within the company.
I subsequently came to Stanley Black and Decker three years ago, as a Lead AI Architect, i.e. the most senior subject matter expert for AI. I very quickly became the head of the AI group; hence I now wear a dual-hat – I am a technologist and I manage a group. In 2020 we launched the AI strategy for the company. My group is all about bringing AI into the products and solutions. I report to our recently appointed Chief AI Officer, Mukesh Dalal.
I have always been interested in the development of the profession. To that end, I have been active within the IEEE society. In 2021, I am honoured to serve as Chair for the IEEE Systems, Man and Cybernetics (SMC) Society for the United Kingdom and Ireland. My committee has recently launched a whole host of activities for 2021 to better serve the profession in this new normal.
What are the latest trends that you are seeing in data and AI at the moment? And where do you feel that the industry is moving?
I can give you the example of predictive analytics, which is rapidly evolving, e.g. trying to predict failure in equipment and processes. Several years ago, we were at the detection stage, we were able to detect problems, and not act on them. Then we went from detection to diagnostics, diagnosing the problem and what is going to happen. Now we are looking at predictive analytics. Essentially predicting what’s going to happen next. This is largely driven by both AI and statistics.
The next is prognostics, we are converting these data driven predictions into estimations for remaining useful equipment life and deliver the insights into something that is humans readable. When you are presenting numbers that is fine, but then bear in mind that, at the end of the day, the data has got to be read by humans and someone is going to have to act upon the information.
I am also seeing very advanced algorithms being delivered on a pay-per-use basis. Essentially AI-as-a-Service (AIaaS). Take inspection using imagery for example. Instead of companies needing to stand up a whole team to work on machine vision inspection, there are services out there that we can tap into.
We are learning a lot from other industries, like the medical industry, for example, when they talk about drug performance for patients. We use the same type of algorithms for equipment to look at when they may fail or break.
What are the challenges in adopting AI into industry?
Many of the challenges of AI in industry is that data is sat in silos and may not be connected to the cloud. Hence a lot of initial effort is to do with connectivity, that is IoT and data engineering.
Also, in manufacturing, you don’t have the whole gamut of AI algorithms accessible to you. Essentially, we don’t want to deliver AI black boxes to make predictions for machines. Algorithms need to be explainable. So, some of the deep learning architectures that you see today are off-limits.
From my perspective I find that SMEs tend to find it harder to adopt new technology, just because there’s a perceived minimum investment threshold into AI. But there are so many tools out there now that democratizes AI and are able to reduce the barrier for entry. The democratization of AI is something that I have great passion for.
Tell me more about your team and AI at Stanley Black & Decker?
My team are a talented bunch. We mainly work on AI for our products. Stanley Black & Decker consists of multiple businesses. We are best known for tool brands like DeWalt, and of course, Stanley and Black & Decker. In fact, we have over 60 brands. Beyond tools we have large businesses in security monitoring, healthcare, automotive manufacturing, oil and gas, construction, and others. My team is involved in delivering AI solutions for many of the products across our varied businesses. We are very much cloud platform agnostic when delivering our solutions.
The projects we are involved in are revenue driven. So, we are building new AI enabled products and creating new AI driven solutions. We are also exploring the services space. Within Stanley, we also have other teams using analytics and AI to optimize our internal business processes, and that’s to do with improving our bottom-line.
AI is data driven. Many businesses within Stanley Black & Decker already collect this data. And if you have data, you can do lots. I’d reference one of your speakers from last year whom I befriended, Bill Schmarzo. He said, “You can use data across an infinite number of use cases at zero marginal cost”. The other thing he said was that “data is a unique asset that never wears out and never depletes”. Once businesses understand this, it will make it easier for them to buy into AI.
Is there anything else you’d like to add?
It’s always useful for us to learn from each other. Many of the biggest industries are represented at the Industrial Data Summit. We are all on the same digital transformation journey, although we are at different stages. The main thing for me is our ability to share and make connections. Case in point, was the great connections I made at the conference last year. One of the contacts I made subsequently presented a whole seminar to Stanley Black & Decker. I would encourage people to come along and get involved.