Carding 4 Carders
Professional
- Messages
- 2,724
- Reaction score
- 1,586
- Points
- 113
Researchers have created a neural network that explains their predictions.
A team of scientists from New York University has developed a neural network that can explain how it makes its predictions. This work sheds light on the principles of functioning of neural networks, which are the basis of artificial intelligence and machine learning.
The main area of research is related to the specific application of neural networks, which has been actively used in recent years-solving complex biological problems. The study was based on the process of RNA splicing, which plays a key role in the transfer of information from DNA to functional RNA and protein products.
"Many neural networks remain black boxes because they can't explain how they work, which raises concerns about their reliability," says Oded Regev, a professor of computer science at New York University's Kurono Institute for Mathematical Sciences. He adds that with a new method that improves the volume and quality of data for machine learning, an interpreted neural network has been developed that can accurately predict complex outcomes and explain how it arrives at its predictions.
To create their model, Regev and his colleagues used existing data on RNA splicing. Their model, in some ways the equivalent of a powerful microscope, allows scientists to track and quantify the process of RNA splicing.
Regev emphasizes: "Based on the 'designed for interpretation'approach we have created a neural network model that provides insight into RNA splicing." Researchers have found that a small structure in RNA that resembles a hair clip can reduce splicing.
These conclusions were confirmed by a series of experiments: when the RNA molecule took the form of a hairpin, the splicing process stopped, and vice versa, when this structure was violated, the splicing resumed.
A team of scientists from New York University has developed a neural network that can explain how it makes its predictions. This work sheds light on the principles of functioning of neural networks, which are the basis of artificial intelligence and machine learning.
The main area of research is related to the specific application of neural networks, which has been actively used in recent years-solving complex biological problems. The study was based on the process of RNA splicing, which plays a key role in the transfer of information from DNA to functional RNA and protein products.
"Many neural networks remain black boxes because they can't explain how they work, which raises concerns about their reliability," says Oded Regev, a professor of computer science at New York University's Kurono Institute for Mathematical Sciences. He adds that with a new method that improves the volume and quality of data for machine learning, an interpreted neural network has been developed that can accurately predict complex outcomes and explain how it arrives at its predictions.
To create their model, Regev and his colleagues used existing data on RNA splicing. Their model, in some ways the equivalent of a powerful microscope, allows scientists to track and quantify the process of RNA splicing.
Regev emphasizes: "Based on the 'designed for interpretation'approach we have created a neural network model that provides insight into RNA splicing." Researchers have found that a small structure in RNA that resembles a hair clip can reduce splicing.
These conclusions were confirmed by a series of experiments: when the RNA molecule took the form of a hairpin, the splicing process stopped, and vice versa, when this structure was violated, the splicing resumed.
