Search

A current bottleneck that prevents Machine Learning (ML) from being successful outside of a few restricted fields such as chess playing and highway driving is its impairment in ap- propriately using context to infer the meaning of what it is observing. This paper describes techniques to allow ML systems to derive meaning from context, derived from how the human cortex works. In particular, this paper shows how the multiplication a horizontal vector representing a Sparse Distributed Representation (SDR) of patterns in sensory data by a vertical vector representing a SDR of patterns in context data followed by a pattern recognition operation on the result- ing matrix results in the integration of relevant context and in the output of data containing meaning.

In this paper, the author examines the differences in handling complexity between the human brain and artificial intelligences. It will be shown how the current modularity and tendency to reduce complexity of most current AIs is a limit to their potential to reach artificial general intelligence (AGI). Finally, techniques to address these limitations and to properly address complexity are ad- dressed.

Over the years, and with the emergence of various technological innovations, the relevance of automatic learning methods has increased exponentially, and they now play a key role in society. More specifically, Deep Learning (DL), with the ability to recognize audio, image, and time series predictions, has helped to solve various types of problems. This paper aims to introduce a new theory, Hierarchical Temporal Memory (HTM), that applies to stock market prediction. HTM is based on the biological functions of the brain as well as its learning mechanism. The results are of significant relevance and show a low percentage of errors in the predictions made over time. It can be noted that the learning curve of the algorithm is fast, identifying trends in the stock market for all seven data universes using the same network. Although the algorithm suffered at the time a pandemic was declared, it was able to adapt and return to good predictions. HTM proved to be a good continuous learning method for predicting time series datasets., This work is funded by “FCT—Fundação para a Ciência e Tecnologia” within the R&D Units Project Scope: UIDB/00319/2020. The grant of R.S. is supported by the European Structural and Investment Funds in the FEDER component, through the Operational Competitiveness and Internalization Programme (COMPETE 2020). [Project n. 039479. Funding Reference: POCI-01-0247- FEDER-039479].

With the development of intrusion detection, a number of the intelligence algorithms (e.g., artificial neural networks) are introduced to enhance the performance of the intrusion detection systems. However, many intelligence algorithms should be trained before being used, and retrained regularly, which is not applicable for continuous online learning and analyzing. In this paper, a new online intrusion scenario discovery framework is proposed and the intelligence algorithm HTM (Hierarchical Temporal Memory) is employed to improve the performance of the online learning ability of the system. The proposed framework can discover and model intrusion scenarios, and the constructed model keeps evolving with the variance of the data. Additionally, a series of data preprocessing methods are introduced to enhance its adaptability to the noisy and twisted data. The experimental results show that the framework is effective in intrusion scenario discovery, and the discovered scenario is more concise and accurate than our previous work.

This paper introduces a two-stage approach to the detection of people eating and/or drinking for the purposes of surveillance of daily life. With the sole use of wearable accelerometer sensor attached to somebody’s (man or a woman) wrists, this two-stage approach consists of feature extraction followed by classification. At the first stage, based on the limb’s three dimensional kinematics movement model and the Extended Kalman Filter (EKF), the realtime arm movement features described by Euler angles are extracted from the raw accelerometer measurement data. In the latter stage, the Hierarchical Temporal Memory (HTM) network is adopted to classify the extracted features of the eating/drinking activities based on the space and time varying property of the features, by making use of the powerful modelling capability of HTM network on dynamic signals which is varying with both space and time. The proposed approach is tested through the real eating and drinking activities using the three dimensional accelerometers. Experimental results show that the EKF and HTM based two-stage approach can perform the activity detection successfully with very high accuracy.

In the Hierarchical Temporal Memory (HTM) paradigm the effect of overlap between inputs on the activation of columns in the spatial pooler is studied. Numerical results suggest that similar inputs are represented by similar sets of columns and dissimilar inputs are represented by dissimilar sets of columns. It is shown that the spatial pooler produces these results under certain conditions for the connectivity and proximal thresholds. Following the discussion of the initialization of parameters for the thresholds, corresponding qualitative arguments about the learning dynamics of the spatial pooler are discussed., {"references":["B. Bobier and M.Wirth, \"Content-based image retrieval using hierarchical\ntemporal memory,\" in Proc. 16th ACM Int. Conf. on Multimedia, 2008,\npp. 925-928.","P. Gabrielsson, R. Konig, and U. Johansson, \"Evolving hierarchical\ntemporal memory-based trading models,\" in EvoApplications\n2013-Applications of Evolutionary Computing, Vienna, April 3-5,\n2013.","J. Hawkins, S. Ahmad, and D. Dubinsky, \"Hierarchical temporal memory\nincluding HTM cortical learning algorithms,\" Numenta, Redwood City,\nCA, Tech. Rep. ver. 0.2.1, 2011.","D.O. Hebb, \"The first stage of perception: growth of the assembly,\" in\nThe Organization of Behavior, New York, Wiley, 1949, intro. and ch. 4,\npp. xi-xix, 60-78.","D. Maltoni, \"Pattern recognition by hierarchical temporal memory,\" DEIS\nUniv. Bologna, Tech. Rep., pp. 1-46, Apr. 13, 2011.","V. Mountcastle, \"The columnar organization of the neocortex,\" Brain, vol.\n120(4), pp. 701-722, 1997.","A.J. Perea, J.E. Merono, and M.J. Aguilera, \"Application of Numenta\nhierarchical temporal memory for land-use classification,\" S. Afr. J. Sci.,\nvol. 105, pp. 370-375, Sept./Oct. 2009.","J. Thornton and A. Srbic, \"Spatial pooling for greyscale images,\" Int. J.\nMach. Learn. & Cyber., vol. 4, pp. 207-216, 2013.","J. van Doremalen and L. Boves, \"Spoken digit recognition using a\nhierarchical temporal memory,\" Interspeech, pp. 2566-2569, Brisbane,\nSept. 22-26, 2008."]}

The main intention of this thesis is to give the basic information about the implementation of a node of one of the neural network algorithms. The main purpose of this thesis is to design, implement and analyze the node of the HTM (Hierarchical Temporal Memory) algorithm suggested by Jeff Hawkins [1]. In this document, a design implementation of HTM algorithm node based on Verilog hardware description language and MATLAB programming language is given. The node of HTM algorithm is implemented using Xilinx Spartan-3e FPGA (Field Programmable Gate Array) kit. The simulation results obtained with Xilinx ISE (Integrated Software Environment) 10.1 software are also provided.