KICACT 2016

Permanent URI for this collectionhttp://repository.kln.ac.lk/handle/123456789/15608

Browse

Author: search.filters.author.Timothy K. Shih

Search Results

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    A Working Group Construction Mechanism Based on Text Mining and Collaborative Filtering
    (Faculty of Computing and Technology, University of Kelaniya, Sri Lanka, 2016) Kasthuri Arachchi, S.P.; Zhen-Rong Chen; Irugalbandara, T.C.; Timothy K. Shih
    Massive Open Online Courses (MOOCs) are popular in E-learning research domain with the advance of internet technology (Sa'don, Alias, and Ohshima 2014). MOOCs easily provide higher education courses for registered users as well as institutions or teachers who can offer courses in order to join more students than traditional education. However, producing high-quality learning materials may cause increase time, cost and efforts. For the purpose of reusing materials and reducing the cost of re-creating materials, the Learning Object (LO) concepts have been introduced. The content management systems which used these LOs are called Learning Objects Repository (LOR). The stored LOs in the repository can be easily searched by users. In this paper we introduce a working group construction mechanism for users on LOR. The proposed mechanism uses text mining technique to analyse the similarity of groups to construct prototypes of working groups. Then find the users' preferences about LOs based collaborative filtering to optimize constructed prototypes. Hence users on LOR can find quickly and easily their interesting learning materials via relevant working groups. This mechanism reduces the consuming time for re-creating learning materials by improving the quality of production. This study is based on a Google MOOC FRA project (http://googleresearch.blogspot.tw/2015/03/announcing-google-mooc-focused-research.html). There are 3 parts of the system (Fig. 1 (a)) as: conversion tool between ELO (http://edxpdrlab.ncu.cc/), Course Builder, Open edX, and SCORM 2004; Authoring Tool for ELO; and Repository for ELO (Fig. 1 (b)). The user on the ELO repository can access the working groups which related to themselves and reduce the time consumed about re-creating learning materials and improving production quality.
  • Thumbnail Image
    Item
    Animal Behavior Video Classification by Spatial LSTM
    (Faculty of Computing and Technology, University of Kelaniya, Sri Lanka, 2016) Huy Q. Nguyen; Kasthuri Arachchi, S.P.; Maduranga, M.W.P.; Timothy K. Shih
    Deep learning which is basis for building artificial intelligent system is become a quite hot research area in recent years. Current deep neural network increase human recognition level of natural images even with huge dataset such as ImageNet. Among successful architectures, Convolution Neural Network (CNN) and Long Short-term Memory (LSTM) are widely used to build complex models because of their advantages. CNN reduces number of parameters compare to full connected neural net. Furthermore, it learns spatial features by sharing weights between convolution patch, which is not only help to improve performance but also extract similar features of input. LSTM is an improvement of Vanilla Recurrent Network (RNN). When processing with time-series data, RNN gradient has tend to vanish in training with backpropagation through time (BTT), while LSTM proposed to solve vanish problem. Therefore it is well suited for manage long-term dependencies. In other words, LSTM learn temporal features of time-series data. During this we study focused on creating an animal video dataset and investigating the way that deep learning system learn features with animal video dataset. We proposed a new dataset and experiments using two types of spatial-temporal LSTM, which allow us, discover latent information of animal videos. According to our knowledge of previous studies, no one has used this method before with animal activities. Our animal dataset created under three conditions; data must be videos. Thus, our network can learn spatial-temporal features, objects are popular animals like cats and dogs since it is easy to collect more data of them and the third is one video should have one animal but without humans or any other moving objects. Under experiments, we did the recognition task on Animal Behavior Dataset with two types of models to investigate its’ differences. The first model is Conv-LSTM which is an extend version of LSTM, by replacing all input and output connections of convolutional connections. The second model is Long-term Recurrent Convolutional Networks (LRCN), which proposed by Jeff Donahue. More layers of LSTM units can easily added to both models in order to make a deeper network. We did classification using 900 training and 90 testing videos and could reached the accuracy of 66.7% on recognition rate. Here we did not do any data augmentation. However in the future we hope to improve our accuracy rate using some of preprocessing steps such as flip, rotate video clips and collecting more data for the dataset.