Selfee, a self-supervised learning approach, is designed to extract comprehensive and discriminative features directly from raw videos of animal behaviors which can be used for in-depth analysis.

DeepEthogram automatically classifies animal behavior videos into researcher-defined behaviors of interest, saving researcher time and enabling more detailed downstream analysis of behavior.

Emerging evidence suggests a broad role for cerebellar circuits in generating and testing predictions about movement, reward, and diverse cognitive processes.

A theoretical model combines self-supervised predictive learning with structural inductive biases to reveal how quasi-continuous attractors that perform accurate angular path integration can be learned from experience during development in the Drosophila and potentially other animal models.

A supervised methodology for mutational signatures outperforms the current standard unsupervised approach revealing new tissue-dependent mutational signatures among which some for obesity.

A supervised learning approach on a high-content genome-wide siRNA screen has identified 591 likely candidates for ciliopathies and facilitated in the discovery of KIAA0586 mutations in individuals with Joubert syndrome.

Recurrent neuronal networks learn to predict movement in a self-supervised way using biologically plausible learning rules.

A machine learning model could accurately estimate daily salt consumption in the general population and can be applied to countries lacking urine sample to compute the mean salt consumption at the population level.

A biologically plausible learning rule enables recurrent neural networks to model the way in which neural circuits use supervised learning to perform time-dependent computations.

A biologically plausible learning rule allows recurrent neural networks to learn nontrivial tasks, using only sparse, delayed rewards, and the neural dynamics of trained networks exhibit complex dynamics observed in animal frontal cortices.