Carlos Zamarreño-Ramos

On spike-timing-dependent-plasticity, memristive devices, and building a self-learning visual cortex.

In this paper we present a very exciting overlap between emergent nanotechnology and neuroscience, which has been discovered by neuromorphic engineers. Specifically, we are linking one type of memristor nanotechnology devices to the biological synaptic update rule known as spike-time-dependent-plasticity (STDP) found in real biological synapses. Understanding this link allows neuromorphic engineers to develop circuit architectures that use this type of memristors to artificially emulate parts of the visual cortex. We focus on the type of memristors referred to as voltage or flux driven memristors and focus our discussions on a behavioral macro-model for such devices. The implementations result in fully asynchronous architectures with neurons sending their action potentials not only forward but also backward. One critical aspect is to use neurons that generate spikes of specific shapes. We will see how by changing the shapes of the neuron action potential spikes we can tune and manipulate the STDP learning rules for both excitatory and inhibitory synapses. We will see how neurons and memristors can be interconnected to achieve large scale spiking learning systems, that follow a type of multiplicative STDP learning rule. We will briefly extend the architectures to use three-terminal transistors with similar memristive behavior. We will illustrate how a V1 visual cortex layer can assembled and how it is capable of learning to extract orientations from visual data coming from a real artificial CMOS spiking retina observing real life scenes. Finally, we will discuss limitations of currently available memristors. The results presented are based on behavioral simulations and do not take into account non-idealities of devices and interconnects. The aim of this paper is to present, in a tutorial manner, an initial framework for the possible development of fully asynchronous STDP learning neuromorphic architectures exploiting two or three-terminal memristive type devices. All files used for the simulations are made available through the journal web site1.

An Event-Driven Multi-Kernel Convolution Processor Module for Event-Driven Vision Sensors

Event-Driven vision sensing is a new way of sensing visual reality in a frame-free manner. This is, the vision sensor (camera) is not capturing a sequence of still frames, as in conventional video and computer vision systems. In Event-Driven sensors each pixel autonomously and asynchronously decides when to send its address out. This way, the sensor output is a continuous stream of address events representing reality dynamically continuously and without constraining to frames. In this paper we present an Event-Driven Convolution Module for computing 2D convolutions on such event streams. The Convolution Module has been designed to assemble many of them for building modular and hierarchical Convolutional Neural Networks for robust shape and pose invariant object recognition. The Convolution Module has multi-kernel capability. This is, it will select the convolution kernel depending on the origin of the event. A proof-of-concept test prototype has been fabricated in a 0.35 μm CMOS process and extensive experimental results are provided. The Convolution Processor has also been combined with an Event-Driven Dynamic Vision Sensor (DVS) for high-speed recognition examples. The chip can discriminate propellers rotating at 2 k revolutions per second, detect symbols on a 52 card deck when browsing all cards in 410 ms, or detect and follow the center of a phosphor oscilloscope trace rotating at 5 KHz.

Multicasting Mesh AER: A Scalable Assembly Approach for Reconfigurable Neuromorphic Structured AER Systems. Application to ConvNets

This paper presents a modular, scalable approach to assembling hierarchically structured neuromorphic Address Event Representation (AER) systems. The method consists of arranging modules in a 2D mesh, each communicating bidirectionally with all four neighbors. Address events include a module label. Each module includes an AER router which decides how to route address events. Two routing approaches have been proposed, analyzed and tested, using either destination or source module labels. Our analyses reveal that depending on traffic conditions and network topologies either one or the other approach may result in better performance. Experimental results are given after testing the approach using high-end Virtex-6 FPGAs. The approach is proposed for both single and multiple FPGAs, in which case a special bidirectional parallel-serial AER link with flow control is exploited, using the FPGA Rocket-I/O interfaces. Extensive test results are provided exploiting convolution modules of 64 × 64 pixels with kernels with sizes up to 11 × 11, which process real sensory data from a Dynamic Vision Sensor (DVS) retina. One single Virtex-6 FPGA can hold up to 64 of these convolution modules, which is equivalent to a neural network with 262 × 103 neurons and almost 32 million synapses.

On neuromorphic spiking architectures for asynchronous STDP memristive systems

Neuromorphic circuits and systems techniques have great potential for exploiting novel nanotechnology devices, which suffer from great parametric spread and high defect rate. In this paper we explore some potential ways of building neural network systems for sophisticated pattern recognition tasks using memristors. We will focus on spiking signal coding because of its energy and information coding efficiency, and concentrate on Convolutional Neural Networks because of their good scaling behavior, both in terms of number of synapses and temporal processing delay. We propose asynchronous architectures that exploit memristive synapses with specially designed neurons that allow for arbitrary scalability as well as STDP learning. We present some behavioral simulation results for small neural arrays using electrical circuit simulators, and system level spike processing results on human detection using a custom made event based simulator.

A 32

This paper describes a convolution chip for event-driven vision sensing and processing systems. As opposed to conventional frame-constraint vision systems, in event-driven vision there is no need for frames. In frame-free event-based vision, information is represented by a continuous flow of self-timed asynchronous events. Such events can be processed on the fly by event-based convolution chips, providing at their output a continuous event flow representing the 2-D filtered version of the input flow. In this paper we present a 32 × 32 pixel 2-D convolution event processor whose kernel can have arbitrary shape and size up to 32 × 32. Arrays of such chips can be assembled to process larger pixel arrays. Event latency between input and output event flows can be as low as 155 ns. Input event throughput can reach 20 Meps (mega events per second), and output peak event rate can reach 45 Meps. The chip can be configured to discriminate between two simulated propeller-like shapes rotating simultaneously in the field of view at a speed as high as 9400 rps (revolutions per second). Achieving this with a frame-constraint system would require a sensing and processing capability of about 100 K frames per second. The prototype chip has been built in 0.35 CMOS technology, occupies 4.3 × 5.4 and consumes a peak power of 200 mW at maximum kernel size at maximum input event rate.

\,\times\,

Most scene segmentation and categorization architectures for the extraction of features in images and patches make exhaustive use of 2D convolution operations for template matching, template search, and denoising. Convolutional Neural Networks (ConvNets) are one example of such architectures that can implement general-purpose bio-inspired vision systems. In standard digital computers 2D convolutions are usually expensive in terms of resource consumption and impose severe limitations for efficient real-time applications. Nevertheless, neuro-cortex inspired solutions, like dedicated Frame-Based or Frame-Free Spiking ConvNet Convolution Processors, are advancing real-time visual processing. These two approaches share the neural inspiration, but each of them solves the problem in different ways. Frame-Based ConvNets process frame by frame video information in a very robust and fast way that requires to use and share the available hardware resources (such as: multipliers, adders). Hardware resources are fixed- and time-multiplexed by fetching data in and out. Thus memory bandwidth and size is important for good performance. On the other hand, spike-based convolution processors are a frame-free alternative that is able to perform convolution of a spike-based source of visual information with very low latency, which makes ideal for very high-speed applications. However, hardware resources need to be available all the time and cannot be time-multiplexed. Thus, hardware should be modular, reconfigurable, and expansible. Hardware implementations in both VLSI custom integrated circuits (digital and analog) and FPGA have been already used to demonstrate the performance of these systems. In this paper we present a comparison study of these two neuro-inspired solutions. A brief description of both systems is presented and also discussions about their differences, pros and cons.

32 Pixel Convolution Processor Chip for Address Event Vision Sensors With 155 ns Event Latency and 20 Meps Throughput

This paper presents a serializer/deserializer scheme for asynchronous address event representation (AER) bit-serial interchip communications. Each serial AER (sAER) link uses four wires: a micro strip pair for low voltage differential signaling (LVDS) and two handshaking lines. Each event is represented by a 32-bit word. Two extra preamble bits are used for alignment. Transmission clock is embedded in the data using Manchester encoding. As opposed to conventional LVDS links, the presented approach allows to stop physical communication between data events, so that no “comma” characters need to be transmitted during these pauses. As soon as a new event needs to be transmitted, the link recovers immediately thanks to a built-in control voltage memorization circuit. As a result, power consumption of the serializer and deserializer circuits is proportional to data event rate. The approach is also highly tolerant to clock jitter, due to the asynchronous nature and the Manchester encoding. A chip test prototype has been fabricated in standard 0.35 μm CMOS including a pair of Serializer and Deserializer circuits. Maximum measured event transmission rate is 15 Meps (mega events per second) for 32-bit events, with a maximum bit transmission speed of 670 Mbps (mega bits per second).

Comparison between Frame-Constrained Fix-Pixel-Value and Frame-Free Spiking-Dynamic-Pixel ConvNets for Visual Processing

This paper presents a low power switchable current mode driver/receiver I/O pair for high speed serial transmission of asynchronous address event representation (AER) information. The sparse nature of AER packets (also called events) allows driver/receiver bias currents to be switched off to save power. The on/off times must be lower than the bit time to minimize the latency introduced by the switching mechanism. Using this technique, the link power consumption can be scaled down with the event rate without compromising the maximum system throughput. The proposed technique has been implemented on a typical push/pull low voltage differential signaling (LVDS) circuit, but it can easily be extended to other widely used current mode standards, such as current mode logic (CML) or low-voltage positive emitter-coupled logic (LVPECL). A proof of concept prototype has been fabricated in 0.35 μm CMOS incorporating the proposed driver/receiver pair along with a previously reported switchable serializer/deserializer scheme. At a 500 Mbps bit rate, the maximum event rate is 11 Mevent/s for 32-bit events. In this situation, current consumption is 7.5 mA and 9.6 mA for the driver and receiver, respectively, while differential voltage amplitude is ±300 mV. However, if event rate is lower than 20-30 Kevent/s, current consumption has a floor of 270 μA for the driver and 570 μA for the receiver. The measured ON/OFF switching times are in the order of 1 ns. The serial link could be operated at up to 710 Mbps bit rate, resulting in a maximum 32-bit event rate of 15 Mevent/s . This is the same peak event rate as that obtained with the same SerDes circuits and a non-switched driver/receiver pair.

An Instant-Startup Jitter-Tolerant Manchester-Encoding Serializer/Deserializer Scheme for Event-Driven Bit-Serial LVDS Interchip AER Links

This paper presents the design and simulation of a serial AER LVDS communication link. It converts data from classical AER parallel bus with a 4-phase handshaking protocol into a bit stream which is transmitted serially into a single LVDS wire. At the receiver side data from the LVDS cable are transformed back to a parallel AER bus and handshaking signals are also properly managed. The link has been designed in a 90 nms technology. Extensive simulations have been performed demonstrating that the link can operate at a speed of 1 Gbps for all the technology corners, exhibiting a power consumption of 27.8 mW for the transmitter and 12.3 mW for the receiver. In the simulation the transmission channel was modelled as a 50 cm cat5E UTP cable, connected to the AER chip through 5 cm PCB traces modelled as a coupled microstrip transmission line. The design has been completed up to the layout level and has been submitted for fabrication. The transmitter and the receiver take up an area of 311times148 mum2 and 300x148 mum2 respectively.

A

This paper summarizes how Convolutional Neural Networks (ConvNets) can be implemented in hardware using Spiking neural network Address-Event-Representation (AER) technology, for sophisticated pattern and object recognition tasks operating at mili second delay throughputs. Although such hardware would require hundreds of individual convolutional modules and thus is presently not yet available, we discuss methods and technologies for implementing it in the near future. On the other hand, we provide precise behavioral simulations of large scale spiking AER convolutional hardware and evaluate its performance, by using peformance figures of already available AER convolution chips fed with real sensory data obtained from physically available AER motion retina chips. We provide simulation results of systems trained for people recognition, showing recognition delays of a few miliseconds from stimulus onset. ConvNets show good up scaling behavior and possibilities for being implemented efficiently with new nano scale hybrid CMOS/nonCMOS technologies.