Google trained artificial intelligence to discern smells by analyzing molecular structures
- An AI trained in a flavor database was able to identify the molecular structure
- The bot can evaluate molecular structure and spit descriptors like & # 39; earthy & # 39;
- Experts say method can help AI smell reach parity with sight and hearing
A team of Google researchers says they taught artificial intelligence to smell.
In an article published in ArxivGoogle's Brain team researchers describe how they used machine learning to instruct a robot to accurately categorize different odors by assessing their molecular structure.
The bot was trained using a database of 5,000 molecules that were identified and labeled by perfume manufacturers with different descriptors as "poignant". or & # 39; grassy & # 39 ;.
Google's AI was able to examine the stenosis of the molecules and relay the corresponding scent (Image)
The researchers fed two-thirds of the database into the bot's neural network and then used the remaining unknown scents to test AI on those they had not analyzed – they passed.
Researchers have long tried to program a sense of smell using artificial intelligence, but several obstacles have proved difficult to overcome.
Difficulties include the subjectivity involved in describing smells. What may be recorded as earthy for one person may be woody for another.
Some molecules also look similar at the molecular level, but when arranged in mirror images, they have different aromas – for example, caraway and mint. This also causes problems for AI.
The technology could help promote systems designed to detect gas leaks or create perfumes (Stock Photo).
Regardless, researchers say their system can help advance artificial olfaction and keep pace with other senses that have been enhanced or imitated by AI.
"Based on these early results with graphical neural networks for molecular properties, we hope that machine learning can eventually do by olfaction what has already been done for vision and hearing," they write.
Other companies, such as IBM, have also tried to advance the field of odor-enabled bots by employing technology that can detect gas leaks and even generate perfumes.
WHAT IS DEEP LEARNING?
Deep learning is a form of machine learning related to algorithms that have a wide range of applications.
It is a field that was inspired by the human brain and focuses on building artificial neural networks.
It was originally formed based on brain simulations and to enable learning algorithms to become better and easier to use.
Processing large amounts of complex data becomes much easier and allows researchers to rely on algorithms to draw accurate conclusions based on the parameters that researchers have set.
Existing task-specific algorithms are better for specific tasks and objectives, but deep learning allows for a broader scope of data collection.
Share or comment on this article:
. (tagsToTranslate) dailymail (t) sciencetech (t) Google