DSpace community: 資訊創新與科技學系

宜蘭縣環境教育行動方案3.0

title: 宜蘭縣環境教育行動方案3.0

宜蘭縣環境教育行動方案2.0

title: 宜蘭縣環境教育行動方案2.0

結合長短期記憶模型與近端策略優化為基礎之策略增強式學習

title: 結合長短期記憶模型與近端策略優化為基礎之策略增強式學習

基於合作賽局理論的LED 照明調控機制

title: 基於合作賽局理論的LED 照明調控機制 abstract: 有鑑於物聯網技術的蓬勃發展，越來越多日常物品透過網際網路連線到雲端，例如智慧農場、智慧家庭和智慧電錶等，這些例子透過物聯網技術監測環境的變化，不僅達到節約能源的效果，也降低了人力與時間成本。在本論文的研究過程中，我們模擬智慧農場的能源管理情境，結合賽局理論與物聯網技術，設計了一套智慧能源分配調光系統。為了使調光系統更加準確並增加智慧農場節約能源的效率，此系統另外導入並整合低功耗通訊技術傳輸、合作賽局理論 Shapley value 計量方法、能源監控裝置等。利用本論文所發展的物聯網技術與系統整合的實驗設計，我們希望能讓管理者隨時隨地監控實驗環境，並讓系統即時偵測室內外部環境光源，同時達到給予植物所需光照及節約能源，實現最低成本能源消耗的理想。

An Algorithm based on Efficient Influence Maximization applied to Social Network

title: An Algorithm based on Efficient Influence Maximization applied to Social Network

A R routine to analyze Global Research Networks at the Individual Level

title: A R routine to analyze Global Research Networks at the Individual Level

A Routine in R to Extract Essential Fields from Patent Documents for Scientometric Analyses

title: A Routine in R to Extract Essential Fields from Patent Documents for Scientometric Analyses

A R Routine to Visualize Global IPC Code Maps

title: A R Routine to Visualize Global IPC Code Maps

Emergence of Global AIOT innovation ecosystem (2000-2019)

title: Emergence of Global AIOT innovation ecosystem (2000-2019)

以生成對抗網路為基礎的動態社群網路預測

title: 以生成對抗網路為基礎的動態社群網路預測

基於融合長短期記憶及深度類神經網路的股票預測

title: 基於融合長短期記憶及深度類神經網路的股票預測

A collaborative filtering recommendation system with dynamic time decay

title: A collaborative filtering recommendation system with dynamic time decay abstract: The collaborative filtering (CF) technique has been widely utilized in recommendation systems due to the precise prediction of users' interests. Most prior CF methods adapted overall ratings to make predictions by collecting preference information from other users. However, in real applications, people’s preferences usually vary with time; the traditional CF could not properly reveal the change in users’ interests. In this paper, we propose a novel CF-based recommendation, dynamic decay collaborative filtering (DDCF), which captures the preference variations of users and includes the concept of dynamic time decay. We extend the idea of human brain memory to specify the level of a user’s interests (i.e., instantaneous, short-term, or long-term). According to different interest levels, DDCF dynamically tunes the decay function based on users’ behaviors. The experimental results show that DDCF with the integration of the dynamic decay concept performs better than traditional CF. In addition, we conduct experiments on real-world datasets to demonstrate the practicability of the proposed DDCF.

Simulation analysis of the wild animal observation system

title: Simulation analysis of the wild animal observation system abstract: The Chinese white dolphin, along the west coast of Taiwan, is claimed as on the brink of distinction species. A top priority before entering the policy process of conservation is to have a well plan of accurate observation. Objective of this paper is to evaluate a proposed novel alternative, which applies the internet of thing (IoT) concept, in the observation of wild animal instead of the conventional way which is manpower intensive. The IoT observation System is consisted of the hydrophone laid on the sea floor, camera/IR sensor carried by Quadcopter, and land control center. Monte Carlo simulation was developed for getting insight of theinteraction between dolphin and IoT observation System. The result showed IoT observation System has a significant effect. The contribution this paper made is to prove that the IoT observation system is far more cost effectively than the manpower intensive boat watching method.

Error protection and correction for block compressive sensing over the binary symmetric channels

title: Error protection and correction for block compressive sensing over the binary symmetric channels abstract: Protection scheme for compressed media form an inevitable part for data transmission. Here we choose block compressive sensing for image compression, and expect to transmit over the binary symmetric channels (BSC). Upon completion of compression at the encoder, we apply multiple description coding (MDC) for error resilience, and the polar codes for error correction. After reception of the possibly erroneous data from BSC, reverse operations for polar decoding and MDC compensation would be applied accordingly, and reconstructed image can be acquired. Simulation results have demonstrated the effectiveness for the protection schemes with the enhancements of the quality of reconstructed images.

Self-Organized Unstructured Network Architecture for Device and Service Deployment in Smart Home

title: Self-Organized Unstructured Network Architecture for Device and Service Deployment in Smart Home abstract: Whereas home access networks are generally built on wired hubs and wireless AP equipment, the deployment of home services is confined to a limited physical communication coverage of a home network. To get rid of local connectivity at home, we propose a self-organized network architecture on which neighbor devices establish direct hop-to-hop connections and relay data flows without resorting to underlying network infrastructures. Cameras, used as demonstration devices, are aligned in an array. Real-time media streams will be issued from any source cameras, sent via relaying cameras, and eventually received by a home gateway connecting with a management server that locates on a home network or the Internet.

Effectiveness evaluation of Internet of Things-aided firefighting by simulation

title: Effectiveness evaluation of Internet of Things-aided firefighting by simulation abstract: With the emerging Internet of Things technology, the world is facing rapid changes in all areas; firefighting is no exception. Conventional firefighting is a dangerous occupation which involves saving lives and property from fires. The skills of firefighting have not changed greatly over the years; hence, using the IoT to aid firefighters is a way to improve their performance. Due to the lack of research on implementing the IoT in the firefighting domain, the objective of this study was to use the quantitative method to gain insights into the usefulness of using the IoT as an aid to firefighting. A Monte Carlo simulation was developed for processing the detailed firefighting interactions in situations of uncertainty. After the verification of the simulation model, the results showed that the search time ratios of unmanned aerial vehicle (UAV) to conventional firefighting for various levels of severity of fire were 30.09, 26.69, and 22.24%. The search and rescue time ratios of UAV to conventional firefighting were 48.27, 35.95, and 31.87%. The most important of these statistics is that at least 50% of the time spent by firefighters on the scene of the fire can be reduced by using the Internet of Things. All of the above data were analyzed usingt test, which showed significant improvement when the Internet of Things was implemented in firefighting. The contribution of this study is to present quantitative results for proving the value of integrating the Internet of Things into firefighting.

Effectiveness analysis of an IoT mechanism in support of monitoring Chinese white dolphins by simulation model

title: Effectiveness analysis of an IoT mechanism in support of monitoring Chinese white dolphins by simulation model abstract: The population of the Chinese white dolphin is claimed to be critically endangered and is on the International Union for Conservation of Nature (IUCN) Red list. It is estimated that there are fewer than 100 individuals in the East Taiwan Strait, and the number is falling. The dolphin’s habitat has been seriously impacted by man-made pollution, such as industry contamination, fishing, and noise. To prevent extinction of the species, conservative action is vital. Prior to any such action, data on the dolphin are essential for decision makers. The current method of observing dolphins is the man-on-boat-watch approach, which is heavily dependent on manpower. Its performance is seriously affected by the weather, fatigue of those on board, and it is also risky and costly. An Internet of things (IoT) data collection mechanism concept is proposed for the purpose of observing dolphins with close watch. It consists of off-the-shelf products such as hydrophones, unmanned aerial vehicles (UAVs) and a specific command and control in search/detection for carrying out the observation task. A Monte Carlo simulation model was developed to analyze the effectiveness of the feasible alternatives, in which some factors are considered and analyzed for their significance. The simulation result showed that the IoT mechanism has an 8.5 times greater chance of availability in operation and at least 2 times more contact than the man-on-boat-watch method. The significant factor affecting the IoT mechanism’s effectiveness is the number of hydrophones and UAVs in the scenario. The great contribution made by this study is that it is the first analytical paper to reveal the effectiveness of an IoT mechanism in benefitting the observation of Chinese white dolphin. The limitation is that it uses off-the shelf products for the IoT mechanism instead of high-end products which could be more effective.

Enhanced Appearance-based Finger Detection and Tracking Using Finite State Machine Control

title: Enhanced Appearance-based Finger Detection and Tracking Using Finite State Machine Control abstract: Real-time finger detection and tracking systems have been growing rapidly in the past decade. Among those methods, Appearance-based and Model-based methods have produced excellent results. However, the occlusion issue is one of the main challenges in this field. In this study, we address this issue by considering the repeating-finite gestures of a guitar-strumming or a hand puppet and, represent using a Finite State Machine model. Also we proposed a novel finger pose tracking system using FSM Model combining with the appearance-based method.. The proposed system consists of two parts: FSM-FT builder creates the FSM hand, and the FSM-FT runner controls the FSM-FT system. Empirically, we conducted an experimental study involving one sample repeating hand gesture and our approach achieved a significance recognition rate of 82% in the testing phase.

Image-adaptive robust transmission for block compressed sensing

title: Image-adaptive robust transmission for block compressed sensing abstract: Compressed media are vulnerable to channel errors during transmission, and protection of compressed media would be much required. In this paper, we employ block compressed sensing (BCS) for compression. We apply the inherent characteristics of original image content to aid the data protection performance for BCS. Image blocks are classified into active and smooth ones, and different protection schemes are applied. With active blocks, we protect with multiple description coding (MDC), while with smooth blocks, we perform the least absolute shrinkage and selection operator (LASSO). Both schemes can be worked together or be performed independently for data protection. Simulation results have pointed out the enhancements with image-adaptive classification of blocks for error resilient transmission of block compressed sensing.

Multi-purpose watermarking with QR code applications

title: Multi-purpose watermarking with QR code applications abstract: Abstract: Watermarking plays an important role in the applications of copyright protection and data authentication. In this paper, we apply the commonly used discrete cosine transform (DCT), and embed two watermarks into color images. For the first watermark we choose the QR code, and for the second one we employ the binary random sequence. With the classification of watermarking, robust watermarking is suitable for copyright protection, while fragile watermarking aims at authentication purposes. We embed the two watermarks into the low and mid frequency bands, respectively. When intentional attacks are applied, the first watermark can be extracted, while no attack is applied, both watermarks can survive. Simulations have demonstrated the capabilities of multi-purpose watermarking with our method.

Image recognition approach for expediting chinese cafeteria checkout process

title: Image recognition approach for expediting chinese cafeteria checkout process abstract: One of the common running themes in modern-day Chinese cafeterias is the hold up in foot traffic in queueing due to checkout. We find out that this bottleneck is caused by the staff requiring extra time to look up the prices of those miscellaneous entrees and calculating the total due amount during checkout. In this paper, this issue is addressed by introducing real-time image recognition techniques into this process. By using a webcam taking live video feed at the checkout desk with the image recognition model outputs the total due amount simultaneously, we are able to eliminate the need to perform manual price calculations. Additionally, the nutrition facts of the meal can also be calculated and displayed to the customers based on the detected entrees. The image recognition model is based on YOLOv3 with 575 entree-catered plate images involved in model training, validation, and testing. The transfer learning technique is also incorporated to speed up the training process. Experimental results show that the recognition accuracy of individual entree is around 70% and that of the entire plate is roughly 63%. With the advanced training with a larger dataset, we believe that the accuracy can be increased, and applying the approach during the checkout will become more practicable.

Learning the classroom automation preferences with low user intervention

title: Learning the classroom automation preferences with low user intervention abstract: With the affordable Internet of Things (IoT) devices, the number of smart classrooms are increasing. There are researches on how to incorporate the IoT technology into the pedagogy. We put the emphasis on classroom automation which enables the teacher to flexibly configure the smart devices without coding. It is achieved by a framework on top of the physical IoT network. In the framework, the automation process is modeled as a state transition engine. The teacher only needs to signal the engine to take a few system state snapshots as the preferences. Once the preference model is derived by the learning process, an event would trigger the engine to compute the suggested system states from this model. Then the automation process invokes the predefined actions to reach the target system states. The framework allows the engineer to provide the basic functions to configure the system, while keeping the user intervention low at providing the training data. In addition to describing the example applications of the framework, a simple use case is also simulated to demonstrate how to design a learning mechanism for this framework.

Unequal error protection for compressed sensing with polar codes

title: Unequal error protection for compressed sensing with polar codes abstract: Compressed sensing is famous for its compression performances under error-free transmission. It would be helpful to apply polar codes for compressed sensing data to cope with the transmission over error-prone channels. Parts of data are more important than others, hence, we apply polar codes with unequal protection capabilities to compressed sensing coefficients, for the error resilient transmission over the binary symmetric channels. Simulations have presented the enhancements and the potential use for practical applications.

Error control and regression for block compressed sensing of color images

title: Error control and regression for block compressed sensing of color images abstract: Transmission of multimedia contents with block compressed sensing (BCS) would be of great value for practical applications. Compressed data are vulnerable to channel errors during transmission, and effective means to protect them would be helpful to the decoding procedures. We choose multiple description coding (MDC) for error control in advance, and apply the least absolute shrinkage and selection operator (LASSO) for training the received data to alleviate the effect caused by channel errors. We also consider the inherent characteristics between color planes of original image to enhance the reconstruction. Simulations have depicted the advantages with these error control schemes for BCS.

Exploring the Use of Machine Learning Techniques for Content-Based Fake News Detection

title: Exploring the Use of Machine Learning Techniques for Content-Based Fake News Detection

The Role of Technology Policy for Knowledge Transfer in Building Emergent Sector

title: The Role of Technology Policy for Knowledge Transfer in Building Emergent Sector abstract: This research considers the interactions between knowledge transfer and technology policy and investigate how technology policy could help technology latecomer countries in maintain its economic competitiveness. We take Taiwan’s attempt to develop Artificial Intelligent sector as an example. Taiwan was one of the first newly-industrialised countries noted for maintaining exceptionally high growth rates and rapid industrialisation between the early 1960s and 1990s (Ash and Greene 2007, Hsu and Perkins 2001, Hong 1997, White and Gray 1988). In the 21st century, the successful development of information technology (IT) and semiconductor industries has played an important role in transforming Taiwan into an advanced emerging economy (Breznitz 2007, Fuller and Rubinstein 2013). In the past three decades, the efforts of the Taiwanese government have included support for and promotion of the emerging sectors such as biotechnology with the hope of upgrading the economic structure of Taiwan into a more advanced knowledge-based economy. However, Taiwan has continuously lost its leading technological status in an era of globalization. In recent years, the government in Taiwan has tried hard to provide incentives to build connections with the US (with special attention paid to the Silicon Valley) to seek the possibility of copying its successful development experience. As an intensively knowledge-based sector, the artificial intelligent sector is an attractive starting point for relatively small economies wanting to build high-value industries. The artificial intelligent sector originated from academic research, which relies heavily on scientific research, the kind that is routinely carried out in the public research sector (Bartholomew, 1997; Carlsson 2010). Previous studies of the biotechnology innovation system in Taiwan laid the foundation for this research by focusing on conflicts between government organisations (Wong, 2005) and the mutual interactions between governance policies and the development of the biotechnology innovation system (Chung, 2011). It is based on a theoretical focus, using the framework of a sectoral innovation system, the concept of knowledge transfer, and analysis of the implementation of technology policies. Empirically it integrated multi-perspective approaches with mixed qualitative-quantitative data gathering from several electronic databases, official publications. Finally, the project will explore the roles academic collaboration play in the networks evolving over time. These analyses will cover few decades to permit a dynamic and longitudinal perspective analysis. These results will be instructive for designing policy incentives to further enhance knowledge transfer between the US and Taiwan. Ultimately, this study hopes to find out how Taiwan can maintain its economic competitiveness in an era of globalization and technological change through developing its artificial intelligent sector.

Proximity and Cluster Effects: The Case of Emerging Biotechnology Innovation Networks

title: Proximity and Cluster Effects: The Case of Emerging Biotechnology Innovation Networks abstract: Literature maintains that proximity has strong impacts on enhancing interactive learning and innovation in the clusters (Howells, 2002). Spatial proximity (distance) can be important sources of knowledge spillover, but other dimensions of proximities (cognitive, institutional, organizational and social dimensions) are equally important (Boschma, 2005; 15. Boschma, R., and K. Frenken. 2010; Omobhude and Chen 2019). However, what is less clear is how do interactions occur in the networks to develop linkage between actors and how do cluster effects enhance the collaborations in emerging high-tech sectors? To explore the associations that regional cluster brought to enhance the formation of R&D networks in the emerging high-tech sector, this paper examines the R&D collaboration network of biotechnology industry in Taiwan between 1998 and 2017. Combining more than 50 interviews and applying social network analysis on a longitudinal dataset gathered from financial reports of 180 emerging biotechnology firms who have initiated public offering (IPO) in Taiwan, this paper aims to explore the R&D collaboration networks between the actors in the innovation system to understand whether cluster effects would enhance the R&D collaborations in the high-tech science-based sectors. Comparing the networks of the firms in this group, a shift from relative sparseness in 1998-2007 to connectedness in 2008-2017 can be readily observed. The finding of this paper suggest that while the nascent sector stays in a small size, geographical proximity is not the most important factor to determinate the networking establishments between the actors in the innovation system. In contrast, the technological distance, the fit of specialties and the mutual complementary of the business is the key factor to drive the formation of collaboration networks and alliances in the biotechnology sector- a science-based sector. To further enhance the collaboration network in a nascent science-based sector, cluster effects through policy intervention attempting to stimulate the collaboration networks between the actors may not be the mostly efficient enhancement. Instead, the strength of local knowledge base, shorten the technological distance, and enhance the mutual complementary between the actors would be most important enhancements to strengthen the local collaboration networks and the knowledge transfer in the networks. Future technology policy to promote emerging sectors needs to focus on building the capabilities of the local sector, taking into account the distinct structural features of local innovation context, rather than copying policy models from the successful experiences from other sectors or from other countries.

Using Camera Array to Detech Elderly Falling and Distribute Alerting Media for Smart Home Care

title: Using Camera Array to Detech Elderly Falling and Distribute Alerting Media for Smart Home Care

CarView Plus: A Social-Based Live Media Distribution Platform in Vehicular Networks

title: CarView Plus: A Social-Based Live Media Distribution Platform in Vehicular Networks

The Design of Real-Time Digital Clothing Projection System

title: The Design of Real-Time Digital Clothing Projection System

The Application of Game with Artificial Intelligence

title: The Application of Game with Artificial Intelligence

見領袖探勘在社群網路中之研究與應用

title: 見領袖探勘在社群網路中之研究與應用

A computational investigation of atrial fibrillation treatment using HIFU energy source

title: A computational investigation of atrial fibrillation treatment using HIFU energy source abstract: Atrial fibrillation (AF) is the most prevalent arrhythmia of the heart, originating usually from ectopic atrial activity of the pulmonary veins. In recent technology, RF ablation has been part of clinical practice for more than two decades and has become an important treatment option for most clinically relevant cardiac arrhythmias [1]. The electrical isolation by ablation of the pulmonary veins (PVs) in the left atrium (LA) of the heart has been proven as an effective cure of atrial fibrillation (AF). The advantage of HIFU is that it could cause deep tissue lesions without damaging intervening tissues and prevent from thrombosis. Thus, it may lead to reducing the risk of stroke. The computer model uses the Pressure Acoustics, Frequency Domain interface to model the stationary acoustic field. The wave equation solved is the homogeneous Helmholtz equation in 2D axisymmetric cylindrical coordinates. With calculated acoustic pressure field, the generated heat combines with the Pennes’ Bioheat Transfer equation to solve the temperature field. The use of extracorporeal HIFU ablation has a problem of vibration motion of heart and it causes focal energy unfocused. The results showed the impact of the thickness of traversing medium to the focal area and heating power with pulse time on the focal area. The varying thickness is to simulate vibration of heart motion. Frequency ranging from 1 MHz to 4 MHz and different geometrical configurations are studied. The preliminary computer modeling and test are feasible. There are in consistent.

基於頻繁項目集之增量協同過濾推薦

title: 基於頻繁項目集之增量協同過濾推薦 abstract: 近年來「推薦系統」的技術越來越蓬勃發展，透過協同過濾的方式，根據用戶的興趣或購買行為，對物品的「評分」或「偏好」，向用戶進行推薦，此技術也廣泛應用在各個領域，例如：電影、書籍、美食…等。傳統的協同過濾是透過用戶相似的偏好，去預測你個人的偏好，進一步把其他跟你相似的人所喜愛的物品推薦給你，達到個人化的推薦效果。然而，對於新進的用戶，該如何進行有效的推薦，是一個值得思索的問題。因此本研究運用使用者對電影的喜好，透過樹的建構，將每筆數據依序插入樹中。當找出頻繁項目集後，再利用協同過濾，找出相似的使用者，進一步去做推薦。本文也將對於新使用者的加入，透過增量式挖掘，將採用規範序列樹Canonical-order tree (Can Tree)方法，將數據庫的數據都儲存於Can Tree中，再以FP-growth的方式對Can Tree進行探勘，找出頻繁項目集。接著利用協同過濾的方式，計算使用者之間的相似性，找出要對使用者推薦的項目集清單，最後使用預測函數進行推薦。

結合動作偵測與直播影像品質調整之動態網路頻寬分配機制

title: 結合動作偵測與直播影像品質調整之動態網路頻寬分配機制 abstract: 隨著物聯網發展的成熟度逐日攀升，與物聯網感測裝置部署的普及化，未來將有越來越多的感測裝置，譬如溫溼度感測器，PM2.5感測器，超音波感測器以及網路攝影機，負責收集資料並交由閘道器整合並送往雲端平台，供雲端平台進行資料的計算與分析，因此在一個家庭網路中，閘道器在整合與轉送至雲端前必須考慮到網路傳輸的頻寬，避免在進行感測器資料的傳輸過程中，因網路資源的相互競爭而造成網路壅塞。本文提出一在小型物聯網基礎下根據動作偵測來動態調整直播監控錄像畫質的機制，此套機制比起一成不變的單一畫質直播，實現節省網路資源之目的。

The Matters of Deep Reinforcement Learning Toward Practical AI Applications

title: The Matters of Deep Reinforcement Learning Toward Practical AI Applications

Simulation analysis of IoT_based police force effectiveness on hostage rescue

title: Simulation analysis of IoT_based police force effectiveness on hostage rescue abstract: Even the fast spread Internet of Thing (IoT) technology has penetrated many fields, there are huge parts of society that is still far behind the people’s expectation in using new technology for public safety and risk reduction, such as the issue of resolving the crime scene by police or fire fighting. This paper aims at proposing an alternative for using IoT in support of police operation in rescue the kidnapped civilian by criminal. Models were developed for a specific scenario. A simulation was used to run for gathering the simulated data in order to prove the effectiveness of the IoT aided police operations. The result has shown the proposed alternative does improve the effectiveness in executing the rescue mission.

一種基於演化的新推薦系統

title: 一種基於演化的新推薦系統 abstract: 由於對用戶興趣的精確預測，矩陣分解（Matrix factorization, MF）技術已被廣泛用於推薦系統中。先前基於 MF 的方法通過從用戶和項目中提取潛在因子來調整總體評級以進行推薦。然而，在實際應用中，人們的偏好通常隨時間而變化；傳統的基於 MF 的方法無法正確捕捉用戶興趣的變化。在本文中，通過將遞歸神經網絡（recurrent neural network, RNN）結合到 MF 中，我們開發了一種新穎的推薦系統 M-RNN-F，以有效地描述用戶隨時間的偏好演變。提出了一種學習模型來捕捉進化模式並預測未來的用戶偏好。實驗結果顯示，M-RNN-F 的性能優於其他最先進的推薦演算法。此外，我們在現實世界數據集上進行實驗，以證明其實用性。

A Biomedical Way for Personality Assessment

title: A Biomedical Way for Personality Assessment

Error resilience for block compressed sensing with multiple-channel transmission

title: Error resilience for block compressed sensing with multiple-channel transmission abstract: Compressed sensing is well known for its superior compression performance, in existing schemes, in lossy compression. Conventional research aims to reach a larger compression ratio at the encoder, with acceptable quality reconstructed images at the decoder. This implies looking for compression performance with error-free transmission between the encoder and the decoder. Besides looking at compression performance, we applied block compressed sensing to digital images for robust transmission. For transmission over lossy channels, error propagation or data loss can be expected, and protection mechanisms for compressed sensing signals are required for guaranteed quality of the reconstructed images. We propose transmitting compressed sensing signals over multiple independent channels for robust transmission. By introducing correlations with multiple-description coding, which is an effective means for error resilient coding, errors induced in the lossy channels can effectively be alleviated. Simulation results presented the applicability and superiority of performance, depicting the effectiveness of protection of compressed sensing signals.