Andreas K. Fidjeland

Accelerated simulation of spiking neural networks using GPUs

Spiking neural network simulators provide environments in which to implement and experiment with models of biological brain structures. Simulating large-scale models is computationally expensive, however, due to the number and interconnectedness of neurons in the brain. Furthermore, where such simulations are used in an embodied setting, the simulation must be real-time in order to be useful. In this paper we present a platform (nemo) for such simulations which achieves high performance on parallel commodity hardware in the form of graphics processing units (GPUs). This work makes use of the Izhikevich neuron model which provides a range of realistic spiking dynamics while being computationally efficient. Learning is facilitated through spike-timing dependent synaptic plasticity. Our GPU kernel can deliver up to 550 million spikes per second using a single device. This corresponds to a real-time simulation of around 55 000 neurons under biologically plausible conditions with 1000 synapses per neuron and a mean firing rate of 10 Hz.

NeMo: A Platform for Neural Modelling of Spiking Neurons Using GPUs

Simulating spiking neural networks is of great interest to scientists wanting to model the functioning of the brain. However, large-scale models are expensive to simulate due to the number and interconnectedness of neurons in the brain. Furthermore, where such simulations are used in an embodied setting, the simulation must be real-time in order to be useful. In this paper we present NeMo, a platform for such simulations which achieves high performance through the use of highly parallel commodity hardware in the form of graphics processing units (GPUs). NeMo makes use of the Izhikevich neuron model which provides a range of realistic spiking dynamics while being computationally efficient. Our GPU kernel can deliver up to 400 million spikes per second. This corresponds to a real-time simulation of around 40 000 neurons under biologically plausible conditions with 1000 synapses per neuron and a mean firing rate of 10 Hz.

iSpike: a spiking neural interface for the iCub robot

This paper presents iSpike: a C++ library that interfaces between spiking neural network simulators and the iCub humanoid robot. It uses a biologically inspired approach to convert the robots sensory information into spikes that are passed to the neural network simulator, and it decodes output spikes from the network into motor signals that are sent to control the robot. Applications of iSpike range from embodied models of the brain to the development of intelligent robots using biologically inspired spiking neural networks. iSpike is an open source library that is available for free download under the terms of the GPL.

A Neurally Controlled Computer Game Avatar With Humanlike Behavior

This paper describes the NeuroBot system, which uses a global workspace architecture, implemented in spiking neurons, to control an avatar within the Unreal Tournament 2004 (UT2004) computer game. This system is designed to display humanlike behavior within UT2004, which provides a good environment for comparing human and embodied AI behavior without the cost and difficulty of full humanoid robots. Using a biologically inspired approach, the architecture is loosely based on theories about the high-level control circuits in the brain, and it is the first neural implementation of a global workspace that has been embodied in a complex dynamic real-time environment. NeuroBots humanlike behavior was tested by competing in the 2011 BotPrize competition, in which human judges play UT2004 and rate the humanness of other avatars that are controlled by a human or a bot. NeuroBot came a close second, achieving a humanness rating of 36%, while the most human human reached 67%. We also developed a humanness metric that combines a number of statistical measures of an avatars behavior into a single number. In our experiments with this metric, NeuroBot was rated as 33% human, and the most human human achieved 73%.

Scalable acceleration of inductive logic programs

Inductive logic programming systems are an emerging and powerful paradigm for machine learning which can make use of background knowledge to produce theories expressed in logic. They have been applied successfully to a wide range of problem domains, from protein structure prediction to satellite fault diagnosis. However, their execution can be computationally demanding. We introduce a scalable FPGA-based architecture for executing inductive logic programs, such that the execution speed largely increases linearly with respect to the number of processors. The architecture contains multiple processors derived from Warrens Abstract Machine, which has been optimised for hardware implementation using techniques such as instruction grouping and speculative assignment. The effectiveness of the architecture is demonstrated using the mutagenesis data set containing 12000 facts of chemical compounds.

A neuronal global workspace for human-like control of a computer game character

This paper describes a system that uses a global workspace architecture implemented in spiking neurons to control an avatar within the Unreal Tournament 2004 (UT2004) computer game. This system is designed to display human-like behaviour within UT2004, which provides a good environment for comparing human and embodied AI behaviour without the cost and difficulty of full humanoid robots. Using a biologically-inspired approach, the architecture is loosely based on theories about the high level control circuits in the brain, and it is the first neural implementation of a global workspace that is embodied in a dynamic real time environment. At its current stage of development the system can navigate through UT2004 and shoot opponents. We are currently completing the implementation and testing in preparation for the human-like bot competition at CIG 2011 in September.

Customisable Multi-Processor Acceleration of Inductive Logic Programming

Parallel approaches to Inductive Logic Programming (ILP) are adopted to address the computational complexity in the learning process. Existing parallel ILP implementations build on conventional general-purpose processors. This paper describes a different approach, by exploiting user-customisable parallelism available in advanced reconfigurable devices such as Field-Programmable Gate Arrays (FPGAs). Our customisable parallel architecture for ILP has three elements: a customisable logic programming processor, a multi-processor for parallel hypothesis evaluation, and an architecture generation framework for creating such multi-processors. Our approach offers a means of achieving high performance by producing parallel architectures adapted both to the problem domain and to specific problem instances.

Customising parallelism and caching for machine learning

Inductive logic programming is an attractive and expressive paradigm for machine learning. A drawback of inductive logic programs is their demanding computational requirements. We present of FPGA-based multi-processor architecture aimed at fast execution of such programs. The architecture exploits both coarse-grained parallelism at the query level, and fine-grained parallelism in the unification algorithm. Instructions are not required, and the components are customised for a hypothesis space referring only to ground unit clauses in the background knowledge. It also benefits from a distributed memory hierarchy, with a method for including background knowledge to eliminate instructions. The effectiveness of this architecture is demonstrated using a large organic chemistry data set. The proposed architecture is faster and smaller than our previous design based on multiple instruction processors. A single customised processor at 38 MHz can run 9 times faster than a Pentium 4 processor at 1.8GHz; a Xilinx XCV2000E device can accommodate 24 processors running in parallel.

Three Tools for the Real-Time Simulation of Embodied Spiking Neural Networks Using GPUs

This paper presents a toolbox of solutions that enable the user to construct biologically-inspired spiking neural networks with tens of thousands of neurons and millions of connections that can be simulated in real time, visualized in 3D and connected to robots and other devices. NeMo is a high performance simulator that works with a variety of neural and oscillator models and performs parallel simulations on either GPUs or multi-core processors. SpikeStream is a visualization and analysis environment that works with NeMo and can construct networks, store them in a database and visualize their activity in 3D. The iSpike library provides biologically-inspired conversion between real data and spike representations to support work with robots, such as the iCub. Each of the tools described in this paper can be used independently with other software, and they also work well together.

Archlog: High-Level Synthesis of Reconfigurable Multiprocessors for Logic Programming

This paper presents Archlog, a language and framework for designing multiprocessor architectures in the logic programming domain. Our goal is to enable application developers in areas such as machine learning and cognitive robotics to produce high-performance designs for reconflgurable devices, without detailed knowledge of hardware development. The Archlog framework provides a high level of abstraction, enabling rapid system generation while supporting high performance. In this paper we present the Archlog language and its library-based compilation framework, which makes use of a customisable logic programming processor. The system generates multiple designs, with different trade-offs in the use of reconfigurable logic and embedded memories. An implementation of a multiprocessor for the machine learning system Progol on a 40MHz XC2V6000 FPGA is 10 times faster than a 2GHz Pentium 4 processor