I’ve been following artificial intelligence (AI) closely since Sydney (a code name for Microsft’s Bing generative AI chatbot) told a tech reporter with The New York Times to leave his wife to be with it instead. I’ve been trying to find ways to use Chat-GPT, DALL-E, and other generative AI to make my life easier with little real success. But I’ve also read about people doing some really interesting things.
Bottom line is, like many people, I am hot and cold on AI.
I am cold on Large Language Models—chatbots like ChatGPT, Copilot, Gemini and LLaMa. While I do acknowledge that some jobs will move from writing to interfacing with an LLM, I don’t see any major lights flashing in the manufacturing industry yet. You could get the LLM to write or solve a Python problem to get an A on your homework, but I wouldn’t deploy anything written by an LLM in a manufacturing environment. I particularly recommend not doing this without so putting a lot of work into validating it first. This would require a significant amount of effort—so much that you might as well have written the code yourself.
After all, I couldn’t even get GPT-4 to write an interesting blog post for Automation World (or even a passable one that would pass a Turing test smell check).
My opinion of LLMs should be taken as being dismissive of AI. Though I’m cold on LLMs, I’m hot convolutional neural networks and Monte Carlo tree searches.
If you have a chance, I recommending reading a recent research paper on using convolutional neural networks (CNNs) with Raman spectroscopy. This paper looks at the use of CNNs to create an accurate, generic model of a process to predict attributes beyond training data. Conventional methods, such as partial least squares, limit the use of Raman spectroscopy to the quality aspects it is based on. Similarly, Monte Carlo tree searches—the same form of AI used to beat top ranked Go players—has been used to discover a new class of compounds to kill drug-resistant bacteria.
While these forms of AI provide important advances for humanity, the reason I am excited about them for automation is that they are far more results driven than the content driven and occasionally hallucinating LLMs. At their most basic, results from convolutional neural networks and Monte Carlo tree searches are testable and verifiable.
Beyond that I also see clear applications within process automation, particularly with tuning PID loops. Findings in the research paper noted about show strong promise to accelerate process development times by accelerating the development of calibration curves for process parameters.
Neural network-based controls have also shown promise in dealing with highly non-linear systems that are highly flexible and less noise prone. The application of these forms of AI may still be a way off in our industry, but I believe they could have a far greater impact than LLMs.
William TomHon is a sales director for the Process Automation division at Catalyx North America (specialists in the science of operational processes), an integrator member of the Control System Integrators Association (CSIA).
