Johan Sveholm

Temporal Sequences of Patterns with an Inverse Function Delayed Neural Network

A network based on the Inverse Function Delayed (ID) model which can recall a temporal sequence of patterns, is proposed. The classical problem that the network is forced to make long distance jumps due to strong attractors that have to be isolated from each other, is solved by the introduction of the ID neuron. The ID neuron has negative resistance in its dynamics which makes a gradual change from one attractor to another possible. It is then shown that a network structure consisting of paired conventional and ID neurons, perfectly can recall a sequence.

A knowledge-based support method for autonomous service operations after disasters

After the 2011 earthquake off the Pacific coast of Tohoku, the importance of network services, like IP phone and e-mail, as a mean of communication in an emergency was hugely increased, but are likely to be discontinued in these situations. If that happens, network administrators have to repair the network and restart the services promptly. It is desirable that novice administrators also take part in network recovery operations, because expert administrators are not always stationed all day long. In this paper, we propose a knowledge-based support method for autonomous service operations in emergency situations. We use the Active Information Resource based Network Management System (AIR-NMS) to reduce the burden on administrators and to enable even novice administrators to operate network services. Finally, we show the effectiveness of the proposed method through experiments using a prototype system.

Multiagent–based processing and integration of system data

This paper presents a multiagent–based ETL (Extract, Transform, Load) unit for the processing and integration of system operational data in order to improve its value. Operational data plays a vital role in managing and optimising systems. Although KDD (Knowledge Discovery and Data Mining) techniques and concepts have long existed, it is only now that we are seeing real applications being extended onto network and systems management. However, the massive data pre–processing (e.g. feature extraction and data integration) which is needed prior to putting KDD tools in action, is still limiting the extent of exploitation. We propose and design the multiagent–based ETL unit which uses Support Vector Machine and Natural Language Processing techniques to efficiently extract information features from operational data. The unit uses an mSPIDER algorithm to discover INclusion Dependencies (INDs) which are used to integrate data across its peers within the system. We demonstrate efficiency of the unit and the used approaches using operational data from a mailing system.

On-demand numerosity reduction for object learning

In Internet of Things, softwares shall enable their host objects (everyday-objects) to monitor other objects, take actions, and notify humans while using some form of reasoning. The ever changing nature of real life environment necessitates the need for these objects to be able to generalize various inputs inductively in order to play their roles more effectively. These objects shall learn from stored training examples using some generalization algorithm. In this paper, we investigate training sets requirements for object learning and propose a Stratified Ordered Selection (SOS) method as a means to scale down training sets. SOS uses a new instance ranking scheme called LO ranking. Everyday-objects use SOS to select training subsets based on their capacity (e.g. memory, CPU). LO ranking has been designed to broaden class representation, achieve significant reduction while offering same or near same analytical results and to facilitate faster on-demand subset selection and retrieval for resource constrained objects. We show how SOS outperforms other methods using well known machine learning datasets.

Numerosity Reduction for Resource Constrained Learning

When coupling data mining (DM) and learning agents, one of the crucial challenges is the need for the Knowledge Extraction (KE) process to be lightweight enough so that even resource (e.g., memory, CPU etc.) constrained agents are able to extract knowledge. We propose the Stratified Ordered Selection (SOS) method for achieving lightweight KE using dynamic numerosity reduction of training examples. SOS allows for agents to retrieve differentsized training subsets based on available resources. The method employs ranking-based subset selection using a novel Level Order (LO) ranking scheme. We show representativeness of subsets selected using the proposed method, its noise tolerance nature and ability to preserve KE performance over different reduction levels. When compared to subset selection methods of the same category, the proposed method offers the best trade-off between cost, reduction and the ability to preserve performance.

Network Management System Based on Activated Knowledge Resource

Nowadays a lot of organizations work and act based on network systems. Because delay of detection and recovery of network failure causes degradation of faith and profits of the organization, the necessity and importance of stable network management and operation with high availability are increased. However, to realize the efficient support of the management task of the administrators, various knowledge and information with respect to the managed network have to be utilized and integrated in the system. In this paper, we propose the practical design method and implementation of a network management system based on activated knowledge resource in a distributed network environment. The effectiveness of our system is evaluated by experiments conducted on an experimental network environment. The activated knowledge resource oriented design is shown to reduce the burden for administrators by the support to utilize and manage various information and knowledge in an autonomic way.

Recalling Temporal Sequences of Patterns Using Neurons with Hysteretic Property

Further development of a network based on the Inverse Function Delayed (ID) model which can recall temporal sequences of patterns, is proposed. Additional advantage is taken of the negative resistance region of the ID model and its hysteretic properties by widening the negative resistance region and letting the output of the ID neuron be almost instant. Calling this neuron limit ID neuron, a model with limit ID neurons connected pairwise with conventional neurons enlarges the storage capacity and increases it even further by using a weightmatrix that is calculated to guarantee the storage after transforming the sequence of patterns into a linear separation problem. The networks tolerance, or the models ability to recall a sequence, starting in a pattern with initial distortion is also investigated and by choosing a suitable value for the output delay of the conventional neuron, the distortion is gradually reduced and finally vanishes.