AI swaggers into world of designer drugs — Analysis

Artificial intelligence has been used by scientists to predict formulas for designer drugs. This is to help improve regulation. AI produced formulas that could be used to create nearly 9 million possible new drugs.

University of British Columbia (UBC), used a deep neural network to help it create the chemical structures of possible new drugs. Their study released this week shows that the computer intelligence performed better than scientists expected.

The research team used a database of known designer drugs – synthetic psychoactive substances – to train the AI on their structures. Designer drugs are always in flux as manufacturers constantly modify their formulations to avoid restrictions or create new products. “legal”Researchers said that substances can be cracked in a matter of months, but it takes time for law enforcement agencies to crack their structures.

Learn more

Drug overdose deaths in US hit all-time record

The vast majority of these drug designs have not been approved for human use and they are unregulated. They pose major health risks to the emergency services in all parts of the world and are an immediate concern.One of the researchers was Dr. Michael Skinnider, a UBC student physician.

The AI generated 8.9 million possible designer drugs after its training. After the training, the researchers created a database of the 196 drugs discovered in real-life. The data showed that the machine had predicted more than 90%.

“We can accurately predict the fate of designer drugs before they appear on the marketplace. This is similar to the 2002 Sci-Fi movie Minority Report. There, foreknowledge regarding criminal activities helped reduce the crime rate in a future universe.”According to Dr. David Wishart (a senior author and professor of computing science at University of Alberta), the article was published.

The AI is still unable to identify unknown compounds, according to the research team. However, they believe it could help in this task as the AI was able to predict the likelihood of certain formulas being created for designer drugs hitting the market. This model “It was 72 percent of the times that I ranked the correct chemical composition of an unidentified drug designer among the top 10 candidates”However, spectrometry analysis which can be easily obtained, increased the accuracy by around 86%.

We were shocked that our model was able to perform this well. Typically, elucidating complete chemical structures using only a precise mass measurement is considered an impossible problem.Skinnider was mentioned.

You think your friends might be interested in this story? Tell your friends!



Related Articles

Back to top button