Polyn has developed an Analog Neural Network Chip

Polyn Technology plans to introduce a novel Neuromorphic processor chip, based on analog electrical circuitry, unlike the standard digital neural networks. Lately, GlobalFoundries has completed the production of the first trial batch in a 55nm CMOS process. Alexander Timofeev, CEO and co-founder of Polyn Technology told Techtime that the chip’s commercial introduction is planned for November or December 2022. The company’s NASP (Neuromorphic Analog Signal Processing) technology had started as a mathematical development of the Chief Scientist and co-founder Dmitry Godovsky. He developed an equation that makes it possible to represent inferencing neural networks using analog elements such as operation amplifiers (Opamps) and resistors.

Based on that concept, Polyn was founded in 2019, and recruited Israeli developers from IBM, Intel, Tower Semiconductors, Cadance etc. The company successfully implemented the new representation by building Analog Neuromorphic Model composed of thousands of digital neurons using classic building blocks: Operational Amplifiers (OpAmp) and Resistors. Polyn’s first NASP chip contains 50,000 analog neurons. Timofeev estimates that its power consumption is 100 times better compared to a parallel digital neural network, and 1,000 times faster.

A combination of Digital Network and an Analog Processor

Since its foundation, Polyn raised $4.5 million, and currently it employs 22 employees. Although registered in UK and holds an office in London, Polyn’s headquarters is located in Caesarea, Israel. It is currently in the process of registering 21 different patents to protect its technology. The final product is a hybrid of a fixed analog network for pattern identification, and additional dynamic digital component in charge of interpreting patterns.

Alexander Timofeev, CEO and co-founder of Polyn Technology
Alexander Timofeev, CEO and co-founder of Polyn Technology

This component is also responsible for the addition of new components to the algorithm during updates and even to perform training at some level. Timofeev: “An average of 90% of the neural network is fixed and unchanged. Less the 10% is updated during operation. Our chip is hybrid: we transform 90% of the network to an analog circuit, and the rest remains digital. We can also adjust this ratio according to the specific needs.”

Three Chips Roamap

The first processor, NeuroSense, will be the first member of NASP family and is planned to reach the market by the end of 2022. Its power consumption is below 100µW, which make it perfect for low-consumption accessories such as smart watches and wearable devices. It will be marketed as a physical component and as an Intellectual Property (IP) offer. “NeuroSense will solve an existing problem within all smart watches: They have to constantly monitor all sensors, something that consume lots of energy.”

Polyn’s roadmap contains two more components: An Audio Processor, intended to segregate human voice from background voices, and is designed for the hearing aids market, first responders teams, radio communication devices and later also consumer applications. The third chip will focus on shock analysis for the industrial market.

What is the immunity level of the chips?

“We rely on existing, well proven production processes. This is why the production of the first chip is performed in CMOS 55nm technology. We perform 50 inferences per second. This is very low frequency, which makes the component highly immune to electromagnetic interferences. As any other neural network, it is a parallel network, meaning a fault in single neuron does not affect the final result. This is not a circuit that accumulates faults.”

Altair’s new business: AI DSP Engines

Photo above: Sony intelligent vision sensors IMX500 (left) and IMX501. Both include Altair’s DSP processor

Hod Hasharon-based Altair Semiconductor (owned by Sony) has secretly expanded its operations beyond the IoT sector and entered the Artificial Intelligence (AI) chips market. This came to light last month, when Sony announced new image sensors for smart control systems . The component is built of two chips embedded in stacked configuration inside a single package (Multi Chip Module) consisting of a Sony image sensor, and a DSP processor developed by Altair, which is responsible for a neuronal network inference operations.

This new family of smart image sensors is currently consists of two components: IMX500 and IMX501. When installed in a security camera, street camera, or other IoT devices, the logic circuit processes and sends only the inference itself to the network center. Thus, it saves considerable processing and communication resources and enables a given device to function as a smart sensor without compromising the privacy of the people being photographed.

A smart camera equipped with the visual-logic sensor can enumerate the number of people in the store and transmit the information without having to send their images to the cloud. It can discern congestion patterns in various complexes, and even track customer behavior in the store –  based only on analyzing their movements –  and without having to identify the customers themselves.

The images are sent back in a variety of configurations (see below): pure decoded information without visual elements, an image in various formats, or only the relevant visual area. From Sony’s point of view, this constitutes an entrance to a major market characterized by a very large growth. As far as Altair is concerned, this is a very surprising development, since so far the company has focused on communication solutions for IoT devices and not on the development of DSP or artificial intelligence processors.

Altair’s core activity is focused on IoT connectivity chips, with its flagship product being the ALT1250 chipset, which includes a modulator and a modem for supporting Cat-M1 standard and the NB-IoT standard. It features an RF front end circuit that supports all LTE bands, an RFIC circuit, a power management unit (PMU), memory, amplifier circuits, filters, an antenna switch, global navigation satellite system (GNSS), hardware-based security, an eSIM circuit and an internal micro controller unit (MCU) that allows customers to develop unique applications.

A new strategy for both Altair and Sony

Sony’s announcement positions it in a massive market and transforms it into ahybrid IoT-image-sensors player. The move can secure orders for Altair in very large quantities. However, it can also hint at a new Altair strategy that can develop in two interesting directions: the first is the integration of ALT1250 technologies into Sony’s future image sensors – alongside the recently unveiled AI processor.

The other direction is independent: integrating the artificial intelligence processor into its next-generation connectivity chip – a kind of ALT1250 reinforced with artificial intelligence. An IoT connectivity chip embedded with artificial intelligence has many advantages – from providing artificial intelligence to ‘dumb’ cameras – thus allowing enhanced communication management capabilities – and even enhancing the current-generation ALT1250 security system.