Kuan-Ting Lai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Kuan-Ting Lai has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
- Nehal Ahmad
Indian Institute of Technology Indore
3 shared publications - Amar Fadillah
Department of Electronic Engineering, National Taipei University of Technology, Taipei, Taiwan
2 shared publications Ming-Syan Chen
Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan
2 shared publications- Mohammad Tanveer
Indian Institute of Technology Indore
2 shared publications Chien-Hung Lai
Department of Electronic Engineering, National Taipei University of Technology, Taipei City, Taiwan
1 shared publication- Chin-Teng Lin
University of Technology Sydney
1 shared publication
Latest Publications
- 30 March 2026
- Published by Emerald Publishing in Data Technologies and Applications
- Vol. 60 (2) , 330-347
LawRAG: Indonesian legal document retrieval-augmented generation with specialized chunking and reranking strategies
Purpose: Enhancing Question-Answering accuracy on new law documents by using Retrieval-Augmented Generation (RAG)-Large Language Models . Design/methodology/approach: The proposed method is built on a RAG design for Indonesian new law documents, incorporating a specific data chunking method and a reranker model. Findings: The proposed method, which segments legal data by focusing on the document title, article, and paragraph, outperforms the sequential chunking method in terms of accuracy. Originality/value: This research demonstrates that the completeness of a sentence in the data set used during the retrieval phase of RAG significantly affects the accuracy of RAG responses. This is particularly true for legal documents, which are closely tied to the titles of articles and the wording of each legal provision. The relationship between the article titles and the content enhances the context-awareness of each data chunking process.
- 30 March 2026
- Published by Emerald Publishing in Data Technologies and Applications
- Vol. 60 (2) , 330-347
LawRAG: Indonesian legal document retrieval-augmented generation with specialized chunking and reranking strategies
Purpose: Enhancing Question-Answering accuracy on new law documents by using Retrieval-Augmented Generation (RAG)-Large Language Models . Design/methodology/approach: The proposed method is built on a RAG design for Indonesian new law documents, incorporating a specific data chunking method and a reranker model. Findings: The proposed method, which segments legal data by focusing on the document title, article, and paragraph, outperforms the sequential chunking method in terms of accuracy. Originality/value: This research demonstrates that the completeness of a sentence in the data set used during the retrieval phase of RAG significantly affects the accuracy of RAG responses. This is particularly true for legal documents, which are closely tied to the titles of articles and the wording of each legal provision. The relationship between the article titles and the content enhances the context-awareness of each data chunking process.
- Open Access
- 14 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. PP, 1
GoalNet: Goal Areas Oriented Pedestrian Trajectory Prediction
Predicting the future trajectories of pedestrians on the road is an important task for autonomous driving. The pedestrian trajectory prediction is affected by scene paths, pedestrian’s intentions and decision-making, which is a multi-modal problem. Relying solely on historical coordinates pedestrian data represents the most straightforward method for pedestrian trajectory prediction. Nevertheless, the accuracy achieved by this method is limited, primarily because it fails to account for the crucial scene paths impacting the pedestrian. Instead of predicting the future trajectory directly, we propose to use scene context and observed trajectory to predict the goal points first, and then reuse the goal points to predict the future trajectories. By leveraging the information from scene context and observed trajectory, the uncertainty can be limited to a few target areas, which represent the "goals" of the pedestrians. In this paper, we propose GoalNet, a new trajectory prediction neural network based on the goal areas of a pedestrian. Our network can predict both pedestrian’s trajectories and bounding boxes. The overall model is efficient and modular, and its outputs can be changed according to the usage scenario. Experimental results show that GoalNet significantly improves the previous state-of-the-art performance by 48.7% on the JAAD and 40.8% on the PIE dataset.
- 14 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. PP, 1
Open AccessGoalNet: Goal Areas Oriented Pedestrian Trajectory Prediction
Predicting the future trajectories of pedestrians on the road is an important task for autonomous driving. The pedestrian trajectory prediction is affected by scene paths, pedestrian’s intentions and decision-making, which is a multi-modal problem. Relying solely on historical coordinates pedestrian data represents the most straightforward method for pedestrian trajectory prediction. Nevertheless, the accuracy achieved by this method is limited, primarily because it fails to account for the crucial scene paths impacting the pedestrian. Instead of predicting the future trajectory directly, we propose to use scene context and observed trajectory to predict the goal points first, and then reuse the goal points to predict the future trajectories. By leveraging the information from scene context and observed trajectory, the uncertainty can be limited to a few target areas, which represent the "goals" of the pedestrians. In this paper, we propose GoalNet, a new trajectory prediction neural network based on the goal areas of a pedestrian. Our network can predict both pedestrian’s trajectories and bounding boxes. The overall model is efficient and modular, and its outputs can be changed according to the usage scenario. Experimental results show that GoalNet significantly improves the previous state-of-the-art performance by 48.7% on the JAAD and 40.8% on the PIE dataset.
- 2 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28 (3) , 1173-1184
Retinal Blood Vessel Tracking and Diameter Estimation via Gaussian Process With Rider Optimization Algorithm
Retinal blood vessels structure analysis is an important step in the detection of ocular diseases such as diabetic retinopathy and retinopathy of prematurity. Accurate tracking and estimation of retinal blood vessels in terms of their diameter remains a major challenge in retinal structure analysis. In this research, we develop a rider-based Gaussian approach for accurate tracking and diameter estimation of retinal blood vessels. The diameter and curvature of the blood vessel are assumed as the Gaussian processes. The features are determined for training the Gaussian process using Radon transform. The kernel hyperparameter of Gaussian processes is optimized using Rider Optimization Algorithm for evaluating the direction of the vessel. Multiple Gaussian processes are used for detecting the bifurcations and the difference in the prediction direction is quantified. The performance of the proposed Rider-based Gaussian process is evaluated with mean and standard deviation. Our method achieved high performance with the standard deviation of 0.2499 and mean average of 0.0147, which outperformed the state-of-the-art method by 6.32%. Although the proposed model outperformed the state-of-the-art method in normal blood vessels, in future research, one can include tortuous blood vessels of different retinopathy patients, which would be more challenging due to large angle variations. We used Rider-based Gaussian process for tracking blood vessels to obtain the diameter of retinal blood vessels, and the method performed well on the “STrutred Analysis of the REtina (STARE) Database,” accessed on Oct. 2020 ( https://cecas.clemson.edu/ ahoover/stare/ ). To the best of our knowledge, this experiment is one of the most recent analysis using this type of algorithm.
- 2 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28 (3) , 1173-1184
Retinal Blood Vessel Tracking and Diameter Estimation via Gaussian Process With Rider Optimization Algorithm
Retinal blood vessels structure analysis is an important step in the detection of ocular diseases such as diabetic retinopathy and retinopathy of prematurity. Accurate tracking and estimation of retinal blood vessels in terms of their diameter remains a major challenge in retinal structure analysis. In this research, we develop a rider-based Gaussian approach for accurate tracking and diameter estimation of retinal blood vessels. The diameter and curvature of the blood vessel are assumed as the Gaussian processes. The features are determined for training the Gaussian process using Radon transform. The kernel hyperparameter of Gaussian processes is optimized using Rider Optimization Algorithm for evaluating the direction of the vessel. Multiple Gaussian processes are used for detecting the bifurcations and the difference in the prediction direction is quantified. The performance of the proposed Rider-based Gaussian process is evaluated with mean and standard deviation. Our method achieved high performance with the standard deviation of 0.2499 and mean average of 0.0147, which outperformed the state-of-the-art method by 6.32%. Although the proposed model outperformed the state-of-the-art method in normal blood vessels, in future research, one can include tortuous blood vessels of different retinopathy patients, which would be more challenging due to large angle variations. We used Rider-based Gaussian process for tracking blood vessels to obtain the diameter of retinal blood vessels, and the method performed well on the “STrutred Analysis of the REtina (STARE) Database,” accessed on Oct. 2020 ( https://cecas.clemson.edu/ ahoover/stare/ ). To the best of our knowledge, this experiment is one of the most recent analysis using this type of algorithm.
- Open Access
Deep Learning for Brain Age Estimation: A Systematic Review
Over the years, Machine Learning models have been successfully employed on neuroimaging data for accurately predicting brain age. Deviations from the healthy brain aging pattern are associated to the accelerated brain aging and brain abnormalities. Hence, efficient and accurate diagnosis techniques are required for eliciting accurate brain age estimations. Several contributions have been reported in the past for this purpose, resorting to different data-driven modeling methods. Recently, deep neural networks (also referred to as deep learning) have become prevalent in manifold neuroimaging studies, including brain age estimation. In this review, we offer a comprehensive analysis of the literature related to the adoption of deep learning for brain age estimation with neuroimaging data. We detail and analyze different deep learning architectures used for this application, pausing at research works published to date quantitatively exploring their application. We also examine different brain age estimation frameworks, comparatively exposing their advantages and weaknesses. Finally, the review concludes with an outlook towards future directions that should be followed by prospective studies. The ultimate goal of this paper is to establish a common and informed reference for newcomers and experienced researchers willing to approach brain age estimation by using deep learning models
Open AccessDeep Learning for Brain Age Estimation: A Systematic Review
Over the years, Machine Learning models have been successfully employed on neuroimaging data for accurately predicting brain age. Deviations from the healthy brain aging pattern are associated to the accelerated brain aging and brain abnormalities. Hence, efficient and accurate diagnosis techniques are required for eliciting accurate brain age estimations. Several contributions have been reported in the past for this purpose, resorting to different data-driven modeling methods. Recently, deep neural networks (also referred to as deep learning) have become prevalent in manifold neuroimaging studies, including brain age estimation. In this review, we offer a comprehensive analysis of the literature related to the adoption of deep learning for brain age estimation with neuroimaging data. We detail and analyze different deep learning architectures used for this application, pausing at research works published to date quantitatively exploring their application. We also examine different brain age estimation frameworks, comparatively exposing their advantages and weaknesses. Finally, the review concludes with an outlook towards future directions that should be followed by prospective studies. The ultimate goal of this paper is to establish a common and informed reference for newcomers and experienced researchers willing to approach brain age estimation by using deep learning models
- 21 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 17 (2) , 2213-2224
AI Wings: An AIoT Drone System for Commanding ArduPilot UAVs
Recent advances in drone and artificial intelligence (AI) technologies have enabled many innovative applications, such as package delivery, reconnaissance, and search and rescue, to name a few. In this article, we propose AI wings, an artificial intelligence of things (AIoT) drone system for commanding multiple unmanned aerial vehicles and deploying AI models. We integrated ArduPilot with the Android mobile platform, which equips DIY drones with AI computing power and 4G/5G connectivity. Embedded control software is developed to cooperate with the AI Wings cloud. Users can easily convert ArduPilot drones into AIoT drones using Android phones, and connect to a cloud server to create their own Internet of Drones. Our cloud server is also integrated with the drone simulation software AirSim for simulating drone missions in virtual reality (VR) worlds. The virtual simulation enables users to test software/hardware configurations as well as train AI models. Moreover, to ensure secure communication, we propose an authentication protocol based on elliptic-curve cryptography with pseudoidentities and time freshness check. In summary, AI Wings provides a cloud server for commanding drone fleets securely, software/hardware design for AIoT drones, and VR simulation for training and testing AI models. Users can install the AI models on the drones directly. To test the system, we built an experimental medical drone service, which delivers an automated external defibrillator to people with a sudden cardiac attack in the shortest time possible.
- 21 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 17 (2) , 2213-2224
AI Wings: An AIoT Drone System for Commanding ArduPilot UAVs
Recent advances in drone and artificial intelligence (AI) technologies have enabled many innovative applications, such as package delivery, reconnaissance, and search and rescue, to name a few. In this article, we propose AI wings, an artificial intelligence of things (AIoT) drone system for commanding multiple unmanned aerial vehicles and deploying AI models. We integrated ArduPilot with the Android mobile platform, which equips DIY drones with AI computing power and 4G/5G connectivity. Embedded control software is developed to cooperate with the AI Wings cloud. Users can easily convert ArduPilot drones into AIoT drones using Android phones, and connect to a cloud server to create their own Internet of Drones. Our cloud server is also integrated with the drone simulation software AirSim for simulating drone missions in virtual reality (VR) worlds. The virtual simulation enables users to test software/hardware configurations as well as train AI models. Moreover, to ensure secure communication, we propose an authentication protocol based on elliptic-curve cryptography with pseudoidentities and time freshness check. In summary, AI Wings provides a cloud server for commanding drone fleets securely, software/hardware design for AIoT drones, and VR simulation for training and testing AI models. Users can install the AI models on the drones directly. To test the system, we built an experimental medical drone service, which delivers an automated external defibrillator to people with a sudden cardiac attack in the shortest time possible.
- 26 June 2022
- Published by Springer Nature
- p. 815-828
Deep Residual Network-Based Sentiment Analysis of Amazon Cell Phone Reviews
- 26 June 2022
- Published by Springer Nature
- p. 815-828
Deep Residual Network-Based Sentiment Analysis of Amazon Cell Phone Reviews
- 20 April 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 32 (3) , 1148-1159
TAGNet: Triplet-Attention Graph Networks for Hashtag Recommendation
Hashtag is an important advertising tool and a must-have feature for social media nowadays. In the past, many hashtag recommendation methods have been proposed from different perspectives of images, texts, and users. However, most previous works consider neither the mutual influence between multi-modalities, nor the visual similarity between images. In this paper, we devise a novel model, named Triplet-Attention Graph Network (TAGNet). The rationale behind our method is that visually similar images share some common hashtags. Therefore, we build an image graph, and apply a new Aggregated Graph Convolution (AGC) module to propagate information in a collective way. Furthermore, it is noted that text and user also have rich content information within posts, and we hence propose a Triplet Attention (TA) module to incorporate multi-modalities into node features. Experiments on the large-scale dataset collected from Instagram show that TAGNet achieved significant improvement in recall rate over the best state-of-the-art method.
- 20 April 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 32 (3) , 1148-1159
TAGNet: Triplet-Attention Graph Networks for Hashtag Recommendation
Hashtag is an important advertising tool and a must-have feature for social media nowadays. In the past, many hashtag recommendation methods have been proposed from different perspectives of images, texts, and users. However, most previous works consider neither the mutual influence between multi-modalities, nor the visual similarity between images. In this paper, we devise a novel model, named Triplet-Attention Graph Network (TAGNet). The rationale behind our method is that visually similar images share some common hashtags. Therefore, we build an image graph, and apply a new Aggregated Graph Convolution (AGC) module to propagate information in a collective way. Furthermore, it is noted that text and user also have rich content information within posts, and we hence propose a Triplet Attention (TA) module to incorporate multi-modalities into node features. Experiments on the large-scale dataset collected from Instagram show that TAGNet achieved significant improvement in recall rate over the best state-of-the-art method.
- 13 January 2021
- Published by Emerald Publishing in Data Technologies and Applications
- Vol. 55 (1) , 1-18
Who is the boss? Identifying key roles in telecom fraud network via centrality-guided deep random walk
Purpose Telecommunication (telecom) fraud is one of the most common crimes and causes the greatest financial losses. To effectively eradicate fraud groups, the key fraudsters must be identified and captured. One strategy is to analyze the fraud interaction network using social network analysis. However, the underlying structures of fraud networks are different from those of common social networks, which makes traditional indicators such as centrality not directly applicable. Recently, a new line of research called deep random walk has emerged. These methods utilize random walks to explore local information and then apply deep learning algorithms to learn the representative feature vectors. Although effective for many types of networks, random walk is used for discovering local structural equivalence and does not consider the global properties of nodes. Design/methodology/approach The authors proposed a new method to combine the merits of deep random walk and social network analysis, which is called centrality-guided deep random walk. By using the centrality of nodes as edge weights, the authors' biased random walks implicitly consider the global importance of nodes and can thus find key fraudster roles more accurately. To evaluate the authors' algorithm, a real telecom fraud data set with around 562 fraudsters was built, which is the largest telecom fraud network to date. Findings The authors' proposed method achieved better results than traditional centrality indices and various deep random walk algorithms and successfully identified key roles in a fraud network. Research limitations/implications The study used co-offending and flight record to construct a criminal network, more interpersonal relationships of fraudsters, such as friendships and relatives, can be included in the future. Originality/value This paper proposed a novel algorithm, centrality-guided deep random walk, and applied it to a new telecom fraud data set. Experimental results show that the authors' method can successfully identify the key roles in a fraud group and outperform other baseline methods. To the best of the authors' knowledge, it is the largest analysis of telecom fraud network to date.
- 13 January 2021
- Published by Emerald Publishing in Data Technologies and Applications
- Vol. 55 (1) , 1-18
Who is the boss? Identifying key roles in telecom fraud network via centrality-guided deep random walk
Purpose Telecommunication (telecom) fraud is one of the most common crimes and causes the greatest financial losses. To effectively eradicate fraud groups, the key fraudsters must be identified and captured. One strategy is to analyze the fraud interaction network using social network analysis. However, the underlying structures of fraud networks are different from those of common social networks, which makes traditional indicators such as centrality not directly applicable. Recently, a new line of research called deep random walk has emerged. These methods utilize random walks to explore local information and then apply deep learning algorithms to learn the representative feature vectors. Although effective for many types of networks, random walk is used for discovering local structural equivalence and does not consider the global properties of nodes. Design/methodology/approach The authors proposed a new method to combine the merits of deep random walk and social network analysis, which is called centrality-guided deep random walk. By using the centrality of nodes as edge weights, the authors' biased random walks implicitly consider the global importance of nodes and can thus find key fraudster roles more accurately. To evaluate the authors' algorithm, a real telecom fraud data set with around 562 fraudsters was built, which is the largest telecom fraud network to date. Findings The authors' proposed method achieved better results than traditional centrality indices and various deep random walk algorithms and successfully identified key roles in a fraud network. Research limitations/implications The study used co-offending and flight record to construct a criminal network, more interpersonal relationships of fraudsters, such as friendships and relatives, can be included in the future. Originality/value This paper proposed a novel algorithm, centrality-guided deep random walk, and applied it to a new telecom fraud data set. Experimental results show that the authors' method can successfully identify the key roles in a fraud group and outperform other baseline methods. To the best of the authors' knowledge, it is the largest analysis of telecom fraud network to date.
- 23 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 6 (6) , 10888-10901
AnyCharge: An IoT-Based Wireless Charging Service for the Public
Mobile phones have become a necessity of modern life and are used for various tasks including audio calls, text messaging, web surfing, video streaming, gaming, and payments. Most of these tasks consume significant battery power, which makes public charging services increasingly important. Therefore, many public places and private stores have started to provide free charging services. However, there are several issues that restrict the popularity of public charging service. First, people cannot easily find charging spots. Second, there is no effective way to monitor charging status and manage the chargers. Third, a free charging service increases business expenses. To address these issues, we developed an IoT-based wireless charging service system, which is called AnyCharge. There are five major components of AnyCharge: a Wi-Fi enabled wireless charger, an IoT gateway, a cloud-based management platform, a secure Wi-Fi auto-connection algorithm, and a mobile app. The charger is connected to an IoT gateway through Wi-Fi using our secure auto-connection algorithm, and the gateways are linked to the cloud server using MQTT. The administrators can monitor and control chargers using the management platform. In addition, Android and iOS apps have been created to allow users to locate free chargers and find the shortest route to the nearest charging spot. We initiated a large-scale experimental deployment in Taiwan, Thailand, Singapore, and Japan. More than 200 chargers have been installed in different places, including restaurants, hospitals, telecom stores, and hotels. Statistics show that there is a strong demand for public charging services.
- 23 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 6 (6) , 10888-10901
AnyCharge: An IoT-Based Wireless Charging Service for the Public
Mobile phones have become a necessity of modern life and are used for various tasks including audio calls, text messaging, web surfing, video streaming, gaming, and payments. Most of these tasks consume significant battery power, which makes public charging services increasingly important. Therefore, many public places and private stores have started to provide free charging services. However, there are several issues that restrict the popularity of public charging service. First, people cannot easily find charging spots. Second, there is no effective way to monitor charging status and manage the chargers. Third, a free charging service increases business expenses. To address these issues, we developed an IoT-based wireless charging service system, which is called AnyCharge. There are five major components of AnyCharge: a Wi-Fi enabled wireless charger, an IoT gateway, a cloud-based management platform, a secure Wi-Fi auto-connection algorithm, and a mobile app. The charger is connected to an IoT gateway through Wi-Fi using our secure auto-connection algorithm, and the gateways are linked to the cloud server using MQTT. The administrators can monitor and control chargers using the management platform. In addition, Android and iOS apps have been created to allow users to locate free chargers and find the shortest route to the nearest charging spot. We initiated a large-scale experimental deployment in Taiwan, Thailand, Singapore, and Japan. More than 200 chargers have been installed in different places, including restaurants, hospitals, telecom stores, and hotels. Statistics show that there is a strong demand for public charging services.
- 22 March 2019
- Published by Springer Nature
- p. 150-162
An Efficient and Resource-Aware Hashtag Recommendation Using Deep Neural Networks
- 22 March 2019
- Published by Springer Nature
- p. 150-162
An Efficient and Resource-Aware Hashtag Recommendation Using Deep Neural Networks