The World is waking up here is another very recent research paper funded by the EU Horizon 2020 project where the researchers are aware of Brainchip and AKIDA along with Intel and IBM but guess which company has the only commercially available neuromorphic chip and IP. This paper fits nicely with the out of the blue approach that CEO Sean Hehir mentioned from a customer in the communications space. I have extracted some of the relevant parts and also provided the link to the full paper. Gets a bit heavy but if you avoid the big words and the maths as I did it should make you very excited about the possibilities as there would be literally 10's of billions of IP licence royalties in play across every 5G then 6G network in the world:
My opinion only DYOR
FF
AKIDA BALLISTA
https://arxiv.org/abs/2206.06047
13 Jun 2022 -
Neuromorphic Wireless Cognition: Event-Driven Semantic Communications for Remote Inference
Jiechen Chen, Nicolas Skatchkovsky, Osvaldo Simeone, Fellow, IEEE
Abstract:
Neuromorphic computing is an emerging computing paradigm that moves away from bat arXiv:2206.06047v1 [cs.IT] 13 Jun 2022 1 Neuromorphic Wireless Cognition: Event-Driven Semantic Communications for Remote Inference Jiechen Chen, Nicolas Skatchkovsky, Osvaldo Simeone, Fellow, IEEE Abstract Neuromorphic computing is an emerging computing paradigm that moves away from batched processing towards the online, event-driven, processing of streaming data.
This paper proposes an end-to end design for a neuromorphic wireless Internet-of-Things system that integrates spike-based sensing, processing, and communication. In the proposed NeuroComm system, each sensing device is equipped with a neuromorphic sensor, a spiking neural network (SNN), and an impulse radio transmitter with multiple antennas. Transmission takes place over a shared fading channel to a receiver equipped with a multi-antenna impulse radio receiver and with an SNN. In order to enable adaptation of the receiver to the fading channel conditions, we introduce a hypernetwork to control the weights of the decoding SNN using pilots. Pilots, encoding SNNs, decoding SNN, and hypernetwork are jointly trained across multiple channel realizations.
The proposed system is shown to significantly improve over conventional frame-based digital solutions, as well as over alternative non-adaptive training methods, in terms of time-to-accuracy and energy consumption metrics.
Index Terms Neuromorphic computing, spiking neural networks, semantic communications.
This work of Osvaldo Simeone and Nicolas Skatchkovsky was supported by the European Research Council (ERC) under the European Union’s Horizon 2020 Research and Innovation Programme (grant agreement No. 725731), and the work by Jiechen Chen was funded by the China Scholarship Council and King’s College London for their Joint Full-Scholarship (K-CSC) under Grant CSC202108440223.
The authors are with the King’s Communications, Learning and Information Processing (KCLIP) lab, King’s College London, London, WC2R 2LS, UK. (email:{jiechen.chen, nicolas.skatchkovsky, osvaldo.simeone}@kcl.ac.uk).
2 I. INTRODUCTION
A. Context and Motivation
The recent rollout of 5G around the world has marked the start of a switch of telecom systems from network-centric carriers of bits to user-centric distributed processors of intelligence.
A key element of this switch will be the integration of wireless systems with sensing and cognition, a technology trend we will refer to as wireless cognition. In this context, this paper is motivated by the two following paradigm shifts that are widely envisioned as central to 6G:
• From universality to goal-driven specialization: With the emergence of wireless cognition, there is a need to develop semantics-aware solutions that integrate sensing, communication, and computing, tailoring resource consumption to the goals of the task at hand [1]-[5].
• From hardware agnostic to hardware-constrained design: By assuming universal computing architectures, the conventional design of communication systems is agnostic to the specific hardware systems deployed at the communicating nodes. This approach fails to acknowledge the critical importance of the computing architecture in the effective and efficient implementation of semantic tasks (see, e.g., [6]). Neuromorphic sensing and computing are emerging as alternative, brain-inspired, paradigms for efficient data collection and semantic signal processing. The main features of the technology are energy efficiency, native event-driven processing of time-varying semantic sources, spikebased computing, and always-on on-hardware adaptation [7, 8].
• Neuromorphic sensors encode information in the timing of spikes, and include neuromorphic cameras, silicon cochleas, and brain-computer interfaces. As a general principle of operation, spikes are produced only when relevant changes occur in the signals being sensed [9–12].
• Neuromorphic processors, also known as spiking neural networks (SNNs), are networks of dynamic spiking neurons that mimic the operation of biological neurons [13]. Spiking neurons communicate and process with the timings of spikes [14]. When implemented on specialized – digital or mixed analog-digital – hardware or on tailored FPGA configurations, SNNs have minimal idle and operating energy cost, and consume as little as a few picojoules per spike [15].
Current commercial use cases of neuromorphic technologies range from drone monitoring via Dynamic Vision Sensor (DVS) cameras [16, 17] through the development of brain-computer 3 interfaces1 to the development of fast and accurate COVID-19 antibody testing2 .
Neuromorphic computing platforms include Intel’s Loihi SNN chip, IBM’s TrueNorth, and Brainchip’s Akida [18, 19].
This work views the emergence of neuromorphic technologies as a unique opportunity for the development of goal-driven, specialized, and hardware-constrained wireless cognition. To date, work on the integration of wireless connectivity and neuromorphic systems has been very limited, including only a specific implementation introduced for biomedical applications.