ANNIIP 2014 Abstracts

Short Papers
Short Papers
Paper Nr: 1
Title:

Applying Artificial Neural Networks to Promote Behaviour Change for Saving Residential Energy
Authors:

Yaqub Rafiq, Shen Wei, Robert Guest, Robert Stone and Pieter de Wilde
Abstract: In this paper Artificial Neural Networks (ANNs) is used to model effects of various human behaviour on energy consumption of the residential buildings in the UK. A virtual model of a bungalow has been developed in which various aspects of the, physical changes in the building such as wall and floor insulation, single and double glazing combined by the human behaviour aspects such as thermostat setting, various periods of door and/or window opening etc. are modelled using EnergyPlus software for evaluating energy consumption for a combination of scenarios. ANNs were then used to learn the effects of various human behaviours on energy consumption. The results demonstrated that the ANN is capable of learning the effects that changes in the human behaviour have in evaluating energy saving in residential buildings and it generated results very quickly for unseen cases.
Download
Paper Nr: 4
Title:

A Novel Neural Network Computing Based Way to Sensor and Method Fusion in Harsh Operational Environments
Authors:

Yuriy V. Shkvarko, Juan I. Yañez and Gustavo D. Martín del Campo
Abstract: We address a novel neural network computing-based approach to the problem of near real-time feature enhanced fusion of remote sensing (RS) imagery acquired in harsh sensing environments. The novel proposition consists in adapting the Hopfield-type maximum entropy neural network (MENN) computational framework to solving the RS image fusion inverse problem. The feature enhanced fusion is performed via aggregating the descriptive experiment design with the variational analysis (VA) inspired regularization frameworks that lead to an adaptive procedure for proper adjustments of the MENN synaptic weights and bias inputs. We feature on the considerably speeded-up implementation of the MENN-based RS image fusion and verify the overall image enhancement efficiency via computer simulations with real-world RS imagery.
Download
Paper Nr: 5
Title:

New Learning Rules for Three-layered Feed-forward Neural Networks based on a General Learning Schema
Authors:

Christina Klüver and Jürgen Klüver
Abstract: We propose a general schema for learning rules in neural networks, the General Enforcing Rule Schema (GERS), from which we infer new simplified learning rules for in particular supervised learning of three-layered feed-forward networks, namely Enforcing Rule Supervised ERS and ERS2. These rules are comparatively simpler than the established rule of Backpropagation and their performance is at least equivalent. The new rules are compared with Backpropagation in different experiments; an additional comparison is performed with real data for the prediction of parcel delivering.
Download
Paper Nr: 6
Title:

Localization of Visual Codes in the DCT Domain Using Deep Rectifier Neural Networks
Authors:

Péter Bodnár, Tamás Grósz, László Tóth and László G. Nyúl
Abstract: The reading process of visual codes consists of two steps, localization and data decoding. This paper presents a novel method for QR code localization using deep rectifier neural networks, trained directly in the JPEG DCT domain, thus making image decompression unnecessary. This approach is efficient with respect to both storage and computation cost, being convenient, since camera hardware can provide JPEG stream as their output in many cases. The structure of the neural networks, regularization, and training data parameters, like input vector length and compression level, are evaluated and discussed. The proposed approach is not exclusively for QR codes, but can be adapted to Data Matrix codes or other two-dimensional code types as well.
Download
Paper Nr: 8
Title:

'Mind Reading': Hitting Cognition by Using ANNs to Analyze fMRI Data in a Paradigm Exempted from Motor Responses
Authors:

José Paulo Marques dos Santos, Luiz Moutinho and Miguel Castelo-Branco
Abstract: The main goal of the present study is to launch the foundations of a pipeline for fMRI-based human behavior classification, addressing however some particularities of cognitive processes. While studying cognition, much of the experiments with fMRI use devices to record subjects’ responses, which recruits the participation of the motor cortex. Although the influence of this aspect may be reduced in subtractive univariate analyses methods, it may negatively interfere in multivariate methods. The fMRI data here used is exempted of motor responses. Subjects were asked to form impressions about persons, objects, and brands, but their thoughts were not recorded by devices. The feedforward backpropagation artificial neural network was used. With this procedure it was possible to correctly classify above randomness. The analysis of the hidden nodes reveals the extensive participation of the fusiform gyri and lateral occipital cortex in this cognitive process, corroborating the critical participation of these structures during classification in the natural brain.
Download
Paper Nr: 12
Title:

Traveling Salesman Problem Solutions by Using Passive Neural Networks
Authors:

Andrzej Luksza and Wieslaw Sienko
Abstract: Presented in this paper numeric experiments on random, relative large travelling salesman problems, show that the passive neural networks can be used as an efficient, dynamic optimization tool for combinatorial programming. Moreover, the passive neural networks, when implemented in VLSI technology, could be a basis for structure of bio-inspired processors, for real-time optimizations.
Download
Paper Nr: 13
Title:

Simplified Information Acquisition Method to Improve Prediction Performance: Direct Use of Hidden Neuron Outputs and Separation of Information Acquisition and Use Phase
Authors:

Ryotaro Kamimura
Abstract: In this paper, we propose a new type of information-theoretic method to improve prediction performance in supervised learning with two main technical features. First, the complicated procedures to increase information content is replaced by the direct use of hidden neuron outputs. We realize higher information by directly changing the outputs from hidden neurons. In addition, we have had difficulty in increasing information content and at the same time decreasing errors between targets and outputs. To cope with this problem, we separate information acquisition and use phase learning. In the information acquisition phase, the auto-encoder tries to acquire information content on input patterns as much as possible. In the information use phase, information obtained in the phase of information acquisition is used to train supervised learning. The method is a simplified version of actual information maximization and directly deals with the outputs from neurons. We applied the method to the protein classification problem. Experimental results showed that our simplified information acquisition method was effective in increasing the real information content. In addition, by using the information content, prediction performance was greatly improved.
Download
Paper Nr: 15
Title:

A Comparison of Two Fitting Functions for Sacadic Pulse Component Mathematical Modelling
Authors:

Rodolfo García-Bermúdez, Camilo Velázquez, Fernando Rojas, Roberto Becerra, Michel Velazquez, Liliana López and Luis Velázquez
Abstract: An accepted model for the saccade signal of ocular motor neurons comprises two components in the form of a pulse and a step. In this contribution, an assessment of two fitting functions for the saccadic pulse component is made, in order to obtain a reduced set of descriptors that could be used for the early diagnosis of ataxia. Results show that both models have achieved to describe the waveform of the saccadic pulse signal, revealing higher performance of Gauss series over the gamma function.
Download
Paper Nr: 2
Title:

Input-output Characteristics of LIF Neuron with Dynamic Threshold and Short Term Synaptic Depression
Authors:

Mikhail Kiselev
Abstract: We consider a model of leaky integrate-and-fire neuron with dynamic threshold and a very simple realization of short term synaptic depression mechanism. Model simplicity makes possible creation of very large networks on its basis. Required general properties of these networks can be obtained due to the appropriate selection of neuron parameters. Knowledge of the dependence of neuron firing frequency on presynaptic activity for various neuron parameters is crucial for this selection. Since this dependence cannot be obtained in an exact analytical form we describe the process of building its empirical approximation using the multiple adaptive regression splines algorithm. This methodology can be used for other neuron models.
Download
Paper Nr: 9
Title:

Neural Network Based Complex Visual Information Processing: Face Detection and Recognition
Authors:

Vaclav Zacek, Eva Volná and Jaroslav Zacek
Abstract: This paper focuses on the issue of detecting and recognizing faces. The work is divided into three main categories. The first part is about detection of faces in constrained conditions. The second part focuses on creation of a different recognition approach. The third one is about the test with robotic devices. However mobile devices (such as robots, small CCD cameras or cheaper cell phones) have many limitations i.e. images quality or very limited computing performance. With respect to limitations the system manages two substantial parts. The first one is responsible for detecting a face in an image. The second one is responsible for calculating the information featured in a face image and recognition of that information. The system is able to process faces in real-time with minimal computation performance and to use minimal space for storing its data. The proposed system was tested on a face database. We have used a FDDB benchmark for an exact comparison.
Download
Paper Nr: 11
Title:

Neural Networks Based Local Weather Prediction System
Authors:

Ján Adamčák, Rudolf Jakša and Ján Liguš
Abstract: In this paper we describe how to build a fully autonomous system for collection, prediction and presentation of single-position meteorological data - the local weather prediction system. By employing nonlinear statistics with neural network predictor on meteorological time-series data we were able to achieve good results for the one-day weather prediction. This novel local statistical approach to weather prediction is different compare to standard methods which are based on the air mass movement modelling. Main objective of this paper is to describe whole system for local weather prediction including technology, software, methods and parameters, and also experimental results.
Download
Paper Nr: 16
Title:

Rete-ECA: A Rule-based System for Device Control
Authors:

Rachel Lee and Sang-Young Cho
Abstract: This paper propose a new rule match algorithm, called Rete-ECA, based on Rete for context-aware device control environment. The Rete-ECA exploits the natural Event-Condition-Action feature of device control situations. This enables the implemtation of the Rete-ECA algorithm to perform better than the Rete algorithm with smaller size and more flexibility. The Rete-ECA system is evaluated using a mouse-tracking environment. Rete-ECA consumes about 2% of the clock ticks consumed by original Rete.
Download