Chemical synthesis is a scientific procedure that uses simple chemical compounds to construct complex ones. It is the exact process used to create most of the substances we use in our daily life – from drugs to those components inside our electronic devices. However, the steps involved are often time intensive and extremely complicated that it may take years for chemists to master or achieve breakthroughs in certain procedures. Nonetheless, artificial intelligence has come to the rescue.
To make it simple for everyone we can say, artificial intelligence is the new lab technician for chemical synthesis. A new approach, that brings in lots of hope to chemist enthusiasts. It’s also amazing that the tech derives its functionality from an already existing software – AlphaGo.
Adding AI to Chemical Synthesis Equations
Some of us can remember how a computer beat the best human chess player Garry Kasparov, right – the software in action was no other but AlphaGo. The same that now scientists are testing to take chemical synthesis to the next level. Ideally, Go’s success comes from a combination of two techs – machine learning which uses deep neural networks, and the Monte Carlo Tree Search principle or concept.
That is, after a successful research, scientists from Muenster University –Germany confirm that combining the above two tech approaches can be perfectly used to plan chemical systems, also named by experts as retrosynthesis, “with remarkable efficiency,” said Marwin Segler, a scientist and lead author of the new study, which appears in Nature’s recent publication.
How Exactly Does Retrosynthesis Work?
Basically, retrosynthesis is termed by scientists as the ultimate branch of organic chemistry. It entails designing and producing important chemical compounds, which at times come with years of learning and consistent research – like how chess players have to invest time in the game to master it.
In reverse, so we may grasp more, the idea is that the formed compound gets broken down to smaller components and the process persists until the point where the most basic component, which cannot be broken further, is obtained. That’s how it happens in the actual “forward procedure.”
Other than mastering the steps experts must possess a natural intuition and creativity to achieve any success in this field. Which is why — it was thought hard for computers to handle the task without a heavy input of programming commands.
Again, referring to how AlphaGo manages to master chess, the machine intelligence deployed in this case works with isolating the best promising possibilities or moves. In other words, every move in chess avails a variety of routes the player can take; however, there is that “best winning-route” which of cause doesn’t have a specific formula but relies on probabilities.
That’s Where the New Artificial Intelligence Technique Comes to Play
In short, the algorithm follows through each particular probability to master the best possible steps, for better chemical synthesis procedures. The deep neural network system deployed is also expected to autonomously improve by learning from millions of published chemical reactions, making it a superior method to humans, for inventing and predicting new processes.
The experts can also test the viability of the predicted moves by running them through the Monte Carlo Tree Search, to ascertain whether an equation will lead to the formation of the particular target molecule.
Well, employing computer power to plan synthesis may not be a new idea but it has never experienced unquestionable breakthroughs, as it now seems with artificial intelligence being the main driver.