Artificial neural networks are efficient data-driven models that allow us to capture complex nonlinear input-output relationships for a variety of tasks, from estimating the ground state of a quantum many-body system to decoding neuronal activity in the brain in response to sensory stimuli. This class of algorithms is naturally biased towards learning efficient representations with respect to input symmetries; however the mechanisms that enforce such property still remain to be elucidated.
In parallel, the problem of understanding if and how symmetries, in sensory input or their statistics, define important aspects of the brain's information processing, has gained a lot of interest recently, notably due to the surge of deep convolutional neural networks.
Artificial neural networks are efficient data-driven models that allow us to capture complex nonlinear input-output relationships for a variety of tasks, from estimating the ground state of a quantum many-body system to decoding neuronal activity in the brain in response to sensory stimuli. This class of algorithms is naturally biased towards learning efficient representations with respect to input symmetries; however the mechanisms that enforce such property still remain to be elucidated.
In parallel, the problem of understanding if and how symmetries, in sensory input or their statistics, define important aspects of the brain's information processing, has gained a lot of interest recently, notably due to the surge of deep convolutional neural networks.
Areas to be covered in this Research Topic may include, but are not limited to:
• Deep Neural Networks
Implicit network bias/regularization and its relation to invariant/equivariant representations.
Learning dynamics of invariant/equivariant representations.
New architectures that encourage or enforce symmetries (beyond translations).
Sample complexity for learning invariant/equivariant representations.
Data augmentations. Symmetries and network's decision interpretability.
• Applications in the Natural Sciences
Physics: physically informed networks; models in particle physics, quantum many body systems.
Chemistry: models of molecules, proteins that enforce physical symmetries.
• Neuroscience & Cerebral Cortex
Invariant/equivariant representations in Neuronal Populations.
Symmetry and efficient information compression of sensory stimuli.
Relevance of symmetry concepts to higher brain function: abstractions.
Plasticity rules and symmetry.
Artificial neural networks are efficient data-driven models that allow us to capture complex nonlinear input-output relationships for a variety of tasks, from estimating the ground state of a quantum many-body system to decoding neuronal activity in the brain in response to sensory stimuli. This class of algorithms is naturally biased towards learning efficient representations with respect to input symmetries; however the mechanisms that enforce such property still remain to be elucidated.
In parallel, the problem of understanding if and how symmetries, in sensory input or their statistics, define important aspects of the brain's information processing, has gained a lot of interest recently, notably due to the surge of deep convolutional neural networks.
Artificial neural networks are efficient data-driven models that allow us to capture complex nonlinear input-output relationships for a variety of tasks, from estimating the ground state of a quantum many-body system to decoding neuronal activity in the brain in response to sensory stimuli. This class of algorithms is naturally biased towards learning efficient representations with respect to input symmetries; however the mechanisms that enforce such property still remain to be elucidated.
In parallel, the problem of understanding if and how symmetries, in sensory input or their statistics, define important aspects of the brain's information processing, has gained a lot of interest recently, notably due to the surge of deep convolutional neural networks.
Areas to be covered in this Research Topic may include, but are not limited to:
• Deep Neural Networks
Implicit network bias/regularization and its relation to invariant/equivariant representations.
Learning dynamics of invariant/equivariant representations.
New architectures that encourage or enforce symmetries (beyond translations).
Sample complexity for learning invariant/equivariant representations.
Data augmentations. Symmetries and network's decision interpretability.
• Applications in the Natural Sciences
Physics: physically informed networks; models in particle physics, quantum many body systems.
Chemistry: models of molecules, proteins that enforce physical symmetries.
• Neuroscience & Cerebral Cortex
Invariant/equivariant representations in Neuronal Populations.
Symmetry and efficient information compression of sensory stimuli.
Relevance of symmetry concepts to higher brain function: abstractions.
Plasticity rules and symmetry.
