Abstract:A key driver of mammalian intelligence is the ability to represent incoming sensory information across multiple abstraction levels. For example, in the visual ventral stream, incoming signals are first represented as low-level edge filters and then transformed into high-level object representations. Similar hierarchical structures routinely emerge in artificial neural networks (ANNs) trained for object recognition tasks, suggesting that similar structures may underlie biological neural networks. However, the classical ANN training algorithm, backpropagation, is considered biologically implausible, and thus alternative biologically plausible training methods have been developed such as Equilibrium Propagation, Deep Feedback Control, Supervised Predictive Coding, and Dendritic Error Backpropagation. Several of those models propose that local errors are calculated for each neuron by comparing apical and somatic activities. Notwithstanding, from a neuroscience perspective, it is not clear how a neuron could compare compartmental signals. Here, we propose a solution to this problem in that we let the apical feedback signal change the postsynaptic firing rate and combine this with a differential Hebbian update, a rate-based version of classical spiking time-dependent plasticity (STDP). We prove that weight updates of this form minimize two alternative loss functions that we prove to be equivalent to the error-based losses used in machine learning: the inference latency and the amount of top-down feedback necessary. Moreover, we show that the use of differential Hebbian updates works similarly well in other feedback-based deep learning frameworks such as Predictive Coding or Equilibrium Propagation. Finally, our work removes a key requirement of biologically plausible models for deep learning and proposes a learning mechanism that would explain how temporal Hebbian learning rules can implement supervised hierarchical learning.

Toward a Universal Cortical Algorithm: Examining Hierarchical Temporal Memory in Light of Frontal Cortical Function

How the Brain Formulates Memory: A Spatio-Temporal Model Research Frontier.

A model of cortical cognitive function using hierarchical interactions of gating matrices in internal agents coding relational representations

A computational hierarchy in human cortex

Sequence learning, prediction, and replay in networks of spiking neurons

Derivation and analysis of basic computational operations of thalamocortical circuits

A Hierarchically Organized Memory Model With Temporal Population Coding

A model for integrating elementary neural functions into delayed-response behavior

Frontal cortex function as derived from hierarchical predictive coding

A Neural Network Model For A Hierarchical Spatio-Temporal Memory

Cortical hierarchy, dual counterstream architecture and the importance of top-down generative networks

A Computational Theory of Hippocampal Function, and Empirical Tests of the Theory.

Learning cortical hierarchies with temporal Hebbian updates

Exploring the functional hierarchy of different pyramidal cell types in temporal processing

Mechanisms of Hierarchical Reinforcement Learning in Corticostriatal Circuits 1: Computational Analysis

A Model of the Hippocampal–cortical Memory System

Exploring the Architectural Biases of the Canonical Cortical Microcircuit

Backprojections in the cerebral cortex : implications for memory

Non-volatile Hierarchical Temporal Memory: Hardware for Spatial Pooling

Could a Computer Architect Understand our Brain?

Fundamental principles of cortical computation: unsupervised learning with prediction, compression and feedback