gave traditional Artificial intelligence that has developed over the last decade is a small number — looking for patterns and providing predictive analytics based on probabilities. Enter generative AI that provides a gateway to its many capabilities. Numerical AI predictions and observations open up the possibility of highly interactive verbal queries.
Also: This year’s top 8 use cases for AI, and which tech experts need to help them.
Emerson senior VP and CTO Peter Zorneau says generative AI helps unlock the previously obscure black box of AI for many enterprise functions, and bridge the divide between operational and information technology. can also help. I recently caught up with Zornio in New York, where he explained how creative AI and numerical AI represent two ends of a continuum. Two variations are based on numerical models and language-based models.
He says that the technical basis of the two AI variants is the same, but how we work with them is different. “Number-based production models are based on datasets of numbers,” he explains. “Language models use datasets based on millions of documents, images, and other objects.”
Now, he says, these two ends of AI are converging, opening up new realms for the behind-the-scenes side of traditional AI. “We’re seeing both being used together,” says Zorneau. “In industrial settings, we’ll use language-based models as a way to interface with numerical-based models that we already have are So can you imagine an operator saying something like, ‘Hey computer, why is production slowing down on this unit? And what can I do to adjust it?
This has huge productivity and time-saving implications, he continues. “It’s a natural way to interface. That’s how you can talk to a 30-year expert in a company, right? You can ask Fred in engineering: ‘What’s going on?’ Then Fred will look at all the production trends, and eventually he’ll come back and tell you, ‘Well, when it’s happening, what’s happening is you’ve got a bad catalyst, and You need to do this you probably need to stop and recreate.
Also: GPT-4 Turbo reclaims ‘Best AI Model’ crown from Anthropic’s Cloud 3
Human talent is essential, Zorneau says, and what Fred is doing in engineering is “using his model that he built up in his head from running this place for 30 years,” Zorneau says. Generative AI advances this work, interfacing with numerically-based AI involves talking to a computer in the same way as a skilled engineer, using scientific deduction. It also has the ability to “look at operations over the past five years, try to find a scenario where exactly the same set of conditions are identical to the imprint of a very similar production. And that imprint will say, ‘Well, what do we do?’ Fred must be thinking: ‘Last time it happened, we did it.’
Finally, Zomio says, the AI will “go through all these different scenarios and find, look at the answers, and tell you: ‘Here are three steps that have been taken to solve this problem in the past. produced excellent results.'”
This end-to-end AI approach offers a great way to build a “product support system, where you take all of your manuals, all of your interactions with your support people, and put them into a system that you can then use for the product.” Can ask questions about Zornio.
Also: AI does more work doesn’t mean it replaces you. Here are 12 reasons to worry less.
Applications exist in all lines of discrete and process manufacturing, from petrochemicals to auto manufacturing. Think of the winemaking industry, which also benefits from end-to-end AI, notes Zorneau. Winemakers with well-censored fields and storage vats can ask questions such as “Why was this year’s wine so much better than last year’s?” AI can “evaluate key indicators like temperature, sugar content, grape acidity, and length of fermentation. What’s the soil condition? What’s the moisture condition? How much sun was there? How much rain did it get?”
In many ways and in many industries, AI will serve as an assistant — and “a great way to interact and query the models you have,” Zorneau pointed out. “They can be more data generated — generated from numeric types of data — but you can also see scrubbing, like an operator log book. Because whenever something happens, the operators write it down. And if you look at all of these , you can ask: ‘Where in the operator logs did this happen before?’ or ‘What was done to solve the problem?’
Also: Best AI Chatbots: ChatGPT is not the only one worth trying.
It also requires greater collaboration between two sides of the house that are divided — the operational technology and information technology teams. Data is where this collaboration begins. IT and OT teams “need to rationalize data from different manufacturers, in all different formats,” Zorneau explains. “Historically, there hasn’t been a lot of love between the two organizations. Because people have built their own systems to do all these things. And they have very different ways of implementing and using it. Something more enlightened. Ideas have tried to provide more integration, but — going forward — there will have to be more collaboration between the two.”
That’s why, emphasizes Zomio, “we need to build an architecture that enables data to move seamlessly from the OT world to the IT world and back. Especially if we use AI systems.” Talk about what could happen in the cloud. OpenAI or other language-based AI models that everyone will interface with.”