The Largest Cognitive Systems Will be Optoelectronic

Abstract

Electrons and photons offer complimentary strengths for information processing. Photons are excellent for communication, while electrons are superior for computation and memory. Cognition requires distributed computation to be communicated across the system for information integration. We present reasoning from neuroscience, network theory, and device physics supporting the conjecture that large-scale cognitive systems will benefit from electronic devices performing synaptic, dendritic, and neuronal information processing operating in conjunction with photonic communication. On the chip scale, integrated dielectric waveguides enable fan-out to thousands of connections. On the system scale, fiber and free-space optics can be employed. The largest cognitive systems will be limited by the distance light can travel during the period of a network oscillation. We calculate that optoelectronic networks the area of a large data center (10 5 m 2 ) will be capable of system-wide information integration at 1 MHz. At frequencies of cortex-wide integration in the human brain (4 Hz, theta band), optoelectronic systems could integrate information across the surface of the earth.

Fluxonic Processing of Photonic Synapse Events

“…Superconducting circuits are complimentary to photonic circuits in these regards. The proposed hardware aspires to achieve greater than one-toone-thousand fan-out in the photonic domain from each neuron to its thousands of connections [14], and subsequently, at each neuronal terminal, the hardware aspires to achieve an additional factor of roughly one-to-ten fanout in the electronic domain, providing each receiving neuron with the capability of analyzing much more information about synaptic activity. Fan-in is envisioned to occur in the electronic domain as the dendritic tree computes and feeds its signals into the neuron cell body, ultimately resulting in a binary decision of whether or not to fire.…”

“…The significance of light and superconductors for scaling was analyzed in Ref. 14. Other work has investigated photonics [15][16][17][18][19] and superconducting electronics [20][21][22][23][24][25] for neuromorphic computation, but to our knowledge none of this work has pursued dendritic processing beyond point neurons or the integration of photonics with superconducting electronics to leverage their complementary strengths for communication and computation.…”

Functional Oxides for Photoneuromorphic Engineering: Toward a Solar Brain

“…15). While the fastest timescale on which biological neurons operate is in the order of milliseconds, the photonic integrate-and-fire devices operate on picoseconds thus representing a virtually unexplored sensory and cognitive processing paradigm of the radio frequency spectrum [881] , [882] , [883] , [884] , [593] , [885] .…”

Resonant tunnelling diode nano-optoelectronic spiking nodes for neuromorphic information processing

“…Neuromorphic photonics is a nascent field, recently gaining significant traction due to increasing importance of AI algorithms and rapid advances in the field of photonic integrated circuits (PICs). Optoelectronic systems in particular are considered as highly suitable for future cognitive computing hardware, as they benefit from operation with both electrons and photons, each excelling at different key functionalities [8]. Thanks to their capability to address bandwidth and interconnect energy limits in a scalable fashion, optoelectronic systems might prove as the optimal solution to overcome these limitations [9].…”