Scientists have found a way not only for people to help on scientific research, but also prove the human brain’s intuitive instinct still outmatches computers. By taking advantage of crowdsourcing, Danish scientists were able to test the abilities of a computer against the abilities of people by using crowdsourcing to help in quantum research.
This type of research enabled a “fun” way for people to provide solutions to research problems associated with quantum research, according to a study published on Wednesday. Even though technology has been growing exponentially since the last few years, scientists from Aarhus University determined that computers couldn’t beat humans in every area just yet.
The human’s brain capacity to solve problems intuitively has been proven to beat computers when presented with a situation where processing abilities fail compared to human’s intuition. The paper published in the journal Nature was focused on making a distinction between the abilities of both computers and people, according to the author Jacob Sherson.
Along with his research team, Sherson was able to harness people’s cerebral capacity by launching a game called Quantum Moves that helped him get the data on which the study was based upon. The game developed by Danish scientists was able to associate people’s performance with quantum computing tasks while playing on an online platform. Impressively enough for random players, computers didn’t stand a chance when presented with situations based on making quick decisions.
The ability to overlook processing new information in order to rely on previous experience is what makes the most noticeable human trait so far. According to the paper published on Wednesday, these problems are quite impossible for computers to solve due to its computational complexity. Thanks to citizen science or crowdsourcing, scientists at Aarhus University compiled sufficient data to create a so-called ‘mind atlas’ in order to see the results.
Human intelligence could soon belong to computers
After analyzing the results, the best solutions were introduced to the computer as default and started showing better results. Although it was people’s data that made the computer able to improve the solution rate, it shows promising results. This could allow for further research to provide a way to improve current artificial intelligence software and take a step forward into next generation AI.
Nevertheless, people do that on a daily basis yet it’s worth noticing that humans intuitively form simple, low-dimensional heuristic strategies, said the paper. This approach on the machine’s deficiency to solve complex problems by relying on previous experiences could lead to the union of artificial intelligence with human-like insights.
The CODER research group seeks to combine the computer’s processing capacity with the human traits that sets bot apart. Nevertheless, considering that Sherson compared the current performance of a computer with the equivalent of solving a 1,000-dimensional problem, it could take some time to do so.
Source: Nature