From Brain Models to Robotic Embodied Cognition: How Does Biological Plausibility Inform Neuromorphic Systems?

Abstract

We examine the challenging "marriage" between computational efficiency and biological plausibility-A crucial node in the domain of spiking neural networks at the intersection of neuroscience, artificial intelligence, and robotics. Through a transdisciplinary review, we retrace the historical and most recent constraining influences that these parallel fields have exerted on descriptive analysis of the brain, construction of predictive brain models, and ultimately, the embodiment of neural networks in an enacted robotic agent. We study models of Spiking Neural Networks (SNN) as the central means enabling autonomous and intelligent behaviors in biological systems. We then provide a critical comparison of the available hardware and software to emulate SNNs for investigating biological entities and their application on artificial systems. Neuromorphics is identified as a promising tool to embody SNNs in real physical systems and different neuromorphic chips are compared. The concepts required for describing SNNs are dissected and contextualized in the new no man's land between cognitive neuroscience and artificial intelligence. Although there are recent reviews on the application of neuromorphic computing in various modules of the guidance, navigation, and control of robotic systems, the focus of this paper is more on closing the cognition loop in SNN-embodied robotics. We argue that biologically viable spiking neuronal models used for electroencephalogram signals are excellent candidates for furthering our knowledge of the explainability of SNNs. We complete our survey by reviewing different robotic modules that can benefit from neuromorphic hardware, e.g., perception (with a focus on vision), localization, and cognition. We conclude that the tradeoff between symbolic computational power and biological plausibility of hardware can be best addressed by neuromorphics, whose presence in neurorobotics provides an accountable empirical testbench for investigating synthetic and natural embodied cognition. We argue this is where both theoretical and empirical future work should converge in multidisciplinary efforts involving neuroscience, artificial intelligence, and robotics.

Keywords: embodied cognition; neuromorphics; neurorobotics; spiking neural networks.

Figure 1

Simplified diagram of biological neuron and its main parts of interest.

Figure 2

Compartmental neuronal model, the adjacency of compartments considers the physical structure of a neuron. Compartments may also be prescribed with appropriate subcellular morphology.

Figure 3

Pointwise neuronal models and canonical equations.

Figure 4

Incoming spikes between neurons are summed using a synaptic weight.

Figure 5

Patterns of connectivity are determined by weights between neurons.

Figure 6

Large collections of motifs form layers or populations (e.g., convolutional, recurrent). Two networks consisting of the same number of neurons and edges may differ depending on organization.

Figure 7

Organizations of neural layers and populations form architectures.

Figure 8

STDP synaptic learning rule window function: the weight between neurons is modified depending on when the pre- and post-synaptic neurons fired.

Figure 9

SNNs on general-purpose computing machines with a cache memory hierarchy that may be optimized to simulate the network efficiently, or SNNs compiled to neuromorphic chips to be run as in-memory computing.

Figure 10

Event-based camera sensors treat each pixel as a neuron sensitive to the log change in photon intensity. This form of sensor is comportable with SNN models for image processing because a spike train may be read out from the sensor as input to downstream populations.

Figure 11

The deployment of neural models for event-based sensing may offer a power-efficient way to process large streams of data in real time by only processing data as changes (to both proprioceptive and exteroceptive senses) arrive.