Neuromorphic Computing Technology - Research Paper Example

Nueuromorphic Computing Nueuromorphic Computing The neuromorphic computing refers to a technology that has been developed by Scientist from the Freje University of Berlin, the Beratsain Center Berlin and the Heidelberg University. This new technology that is still being worked on by the scientists from the three above named institutions works on the basis of parallel data processing. With this new technology, the computational work is done by the neurons that are made using silicon technology (Schmuker, Pfeil &. Nawrot, 2014). The silicon neurons are put on a computer chip from where they are able to carry puts parallel data processing. In this prototype that was built by the scientists, the neurons are linked together in a way that makes them appear and operate like the brain cells. When data is then fed into the assembly, the neurons are able to work in a parallel manner in solving the problem. This is unlike the normal working in various devices which is usually non parallel. Due to the ability of the assembly to function in a parallel manner it can be used to solve complex computing problems (Schmuker, Pfeil &. Nawrot, 2014). According to Schmuker, Pfeil &. Nawrot (2014) the assembly that was made is able to classify data with different features and even recognizes data which is hand written. Moreover, it was able to distinguish plant species on the basis of flowering characteristics. This is facilitated by the parallel working of the neurons. Schmuker, Pfeil &. Nawrot (2014) indicates that the design of the network architecture was based and inspired by the nervous system of insects that processes odor. The system is highly optimized for processing complex chemicals in a parallel manner. The chip with neurons that was used in the study by the scientists was manufactured by the Kirchhoff Institute for Physics in Heidelberg University. Characteristics of the Technology In order to understand the working of the network architecture that was developed by these scientists, it is imperative to first understand the characteristics of silicon neurons as they are the building blocks that were used it the new architecture. According to Merolla et al. (2011) silicon neurons are made by combining the principles of silicon engineering with the neuron physiological principle. When this is done an analog integrated circuit is produced. The analog circuit that is produced possesses functional characteristics similar to those of the human nerve cells (Indiveri, Chicca & Douglas, 2009). The silicon neuron is therefore able to effectively control the discharge and dynamics of the ions in a similar way that the real nerve neurons do. The silicon neurons that have been produced are used in various technological and cyber applications (Arena, Fortuna, Frasca & Patane, 2009). According Brüderle (2011), such neurons are able to operate in real time and consume little power. Several neurons can be fabricated on a single silicon chip. The silicon neurons have been used in various electronic and computer systems (Renaud S, et al, 2010). An example of an area where it has been used is in the control of robots (Indiveri G, et al., 2011). In the robots, they are spiked to be able to control the movement of the robots. The technology that was used in parallel computing is purely based on silicon neurons. The technology was found to be able to classify data and even recognize handwritten data. This therefore means that the technology can be used in various technical devices. Examples of areas where the technology can be used to solve complex tasks include smartphones and super computers. This technology can be effectively used in super computers that are built on the principle of human brain (Indiveri et al., 2011). This will effectively help in carrying out or computing complex problems. However, it will be necessary to determine the required architecture of the human brain like super computers to be developed. It is worth noting that even though this can be done in the future, it will not be easy since there are no silicon neurons that are identical (Merolla, Arthur, Alvarez, Bussat & Boahen, 2014). Therefore, it will be hard to assemble the different silicon neurons to work in solving complex tasks in a parallel manner (Gao, Benjamin & Boahen, 2012). Should this be done, the resulting human brain-like computer will be very effective. If this technology is used in super computers, it can have both positive and negative impacts on cyber security. The technology can improve cyber security through its ability to process data in real time and compute complex data. The technology can be used in helping determine which incoming data is valid and the data that is invalid or is coming from cyber attackers. The ability to compute data in a parallel manner coupled with the ability to distinguish data based on the features and characteristics will make it possible to improve cyber security (Choudhary S, et al. 2012). Despite the advantages using this parallel silicon neuron computation technology to improve cyber security, there is also the possibility of this technology being used by cyber attackers to further their own interests. Due to the ability of the technology to carry out the function of distinguishing data based on features, cyber attackers may use the technology in devices so as to be able to distinguish between the data that they want from their target and those that do not want. This, coupled with the ability of the technology to process complex tasks in real time may facilitate cyber attacks.The attacks that result from the use this technology by the cyber attackers will be complex and occur in real time. In order to counter the cyber threats, the pillars of IA/IS can be used (Benjamin et al., 2014). The five pillars of AI/IS can effectively be used to monitor and prevent the threats that may result from the use of technology by cyber attackers in carrying out cyber attacks. The five aspects of the pillar that can be used to respond to the cyber attacks that may result from the use of this technology are availability of information, information integrity, authentication, information confidentiality, and non-repudiation. Availability aspect can be considered by ensuring that should the technology is used to come up with a brain like supercomputer, information that is processed by the system is only available to those who need it and when it is needed. The second aspect is integrity which entails ensuring that the system and the data is sound and unimpaired. This can be done by ensuring that no intrusion can occur to the information being processed by the silicon neuron based super computer. The third aspect is authentication. This can easily be achieved because of the ability of the silicon neurons to differentiate data features, This ability can be used is setting up authentication system in the devices where they are used so as to ensure information contained is only accessed by the right users. The fourth aspect that should be considered is confidentiality. In order to ensure that there is confidentiality of data, it will be imperative that the ability of the technology to differentiate data based on features is used to block any intrusion into the system. The final aspect is non repudiation. This will be important as it will ensure that data originality is ensured. In this aspect, the differentiating ability of the technology will be vital as it will help in determining data that has been altered and data that has not be interfered with. This will ensure that there is no modification of data or information that is received by the system.  Reference Arena,P. , Fortuna,l., Frasca, M. & Patane, L. (2009). Learning anticipation via spiking networks: Application to navigation control. Neural Networks, IEEE Transactions. 20(2) pp. 202–216. Benjamin, B et al. (2014). Neurogrid: A mixed-analog-digital multichip system for large-scale neural simulations. Proceedings of the IEEE. 102(5) pp. 699–716. Brüderle D, et al. (2011) A comprehensive workflow for general-purpose neural modeling with highly configurable neuromorphic hardware systems. Biol Cybern 104(4-5):263–296. Choudhary S, et al. (2012). Silicon neurons that compute. Proceedings of the IEEE. 22(2) pp 121–128. Gao,P., Benjamin, B. & Boahen, K.(2012). Dynamical system guided mapping of quantitative neuronal models onto neuromorphic hardware. IEEE Trans. on Circuits and Systems. 59 (10), pp. 2383–2394. Indiveri G, Chicca E,Douglas R .(2009). Artificial cognitive systems: From VLSI networks of spiking neurons to neuromorphic cognition. Cognit Comput 1:119–127. Indiveri G, et al. (2011). Neuromorphic silicon neuron circuits. Front Neurosci 5:73. Merolla P, et al. (2011) A digital neurosynaptic core using embedded crossbar memory with 45pJ per spike in 45nm. Custom Integrated Circuits Conference (CICC), IEEE (IEEE, Piscataway, NJ), pp 1–4. Merolla, P., Arthur, J., Alvarez, R., Bussat, M. & Boahen, K. (2014). A multicast tree router for multichip neuromorphic systems. Circuits 4and Systems I: Regular Papers, IEEE Transactions.61 (3) pp. 820–833. Renaud S, et al. (2010) PAX: A mixed hardware/software simulation platform for spiking neural networks. Neural Netw 23(7):905–916. Schmuker, M., Pfeil, T. &. Nawrot, P. (2014). A neuromorphic network for generic multivariate data classification. PNAS, published ahead of print January 27, doi:10.1073/pnas.1303053111 Read More