Published on : Dec 17, 2019
Closed loop systems combining automation with AI algorithms have been game changing for numerous industries. You name it and the potential applications are seemingly limitless: DNA synthesis, genome sequencing, and drug development being at the forefront in reaping the benefits. Researchers of all hues acknowledge the high throughput they have achieved using robotics. But can the automation share a part of our burden by getting human-like in pursuit. This implies can we have a robotic technology that not just automates but thinks more like a human. Well, the issue has kept especially materials scientists on the heels if not nonplussed.
In fact, materials scientists are those hordes of researchers who have been slow to pick the baton in the race to automation coupled with AI. Reasons abound: there are infinite ways of attempting the permutation and combination of elements on earth, and so are the types of materials. Finding a material with characteristics that can best meet a certain application is no easy feat. Researchers call it akin to solving a needle-in-the-haystack problem.
AI-driven Robotics Still Need Human Oversight
Indeed robots have helped. Investigators have used the technology to mix numerous materials in different recipes on different platforms such as wafers. And, unarguably the approach has enabled them to come out with different materials with distinct properties. But, innovation! Not so easy path to tread on. A case in point is finding the toughest 3D-printed structures for different applications. This is because, the new materials that scientists and engineers unveil still need human oversight. Certainly, AI-driven robotics has changed the rule of the game and have helped them to accelerate their efforts.
Robots and AI Helping Materials Scientists in their Pursuit
But a few recent developments have upped the ante. The applications are exciting. Researchers at the Massachusetts Institute of Technology in Cambridge in the international stage of material research recently demonstrated the potential of an autonomous system in finding better perovskite photovoltaics. This might pave way for innovation in solar industry. Another case in point is leveraging the potential of AI-driven robots in finding the recipe for better electrolytes for lithium batteries. In another example, researchers in Carnegie Mellon University, Pittsburgh, developed an AI-based automation system to find novel catalysts. Examples abound.
However, the two most concerning issues draw our attention. There is lack of standardization of the system and researchers are never on the same page on the ways to relate testable properties of the material with its structure.