The introduction of AI meta-learning into neuroscience, the medical effect is expected to improve accurately

Recently, a technical result of the cooperation between the National University of Singapore, ByteDance and other institutions was published in the top journal "Nature Neuroscience" in neurobiology.This research is the first to introduce meta-learning methods in the field of artificial intelligence into the field of neuroscience and medicine, which can train reliable AI models on limited medical data and improve the effect of precision medicine based on brain imaging.

Brain imaging technology is an important field in the development of neuroscience. It can directly observe the neurochemical changes of the brain during information processing and response to stimuli, thus providing an important reference for the diagnosis and treatment of diseases.In theory, brain-imaging-based AI models could be applied to predict some representational properties of individuals.

For example, the clinical effect of a drug or a treatment, etc., so as to promote precision medicine for individuals and improve the level of medical care and nursing in society.

Although large-scale human neuroscience datasets such as the UK Biobank (UK Biobank) are now available, small-scale data samples of dozens to hundreds of people are still the norm when studying clinical populations or solving key neuroscience problems..Given the limited amount of accurately labeled medical data, how to train a reliable AI model is becoming a focus in the fields of neuroscience and computer science.

In the latest research results released by Nature Neuroscience, researchers from National University of Singapore, ByteDance, McGill University and other institutions proposed for the first time to use meta learning in the field of machine learning to solve the above problems.Meta-learning is one of the most popular learning methods in the past few years, and its goal is to allow models to quickly learn new tasks based on the acquired knowledge.

Analyzing small samples of data from the past, researchers have found a meaningful correlation between individual cognitive, mental health, demographic and other health characteristics and brain imaging data.Based on this correlation between small sample data and large data sets, researchers propose a method called meta-matching.This approach can transfer machine learning models trained on large datasets to small datasets, thereby training more reliable models to more accurately guess new phenotypes.

At present, this new method has been evaluated on the datasets of the UK Biobank and the Human Connectome Project, and the evaluation results show that the new method reflects a higher accuracy rate than the traditional method.Experiments show that this new training structure is very sensitive and can be combined with any machine learning algorithm, and can effectively train AI prediction models with good generalization performance on small-scale data sets.