Search

BackgroundThe widespread use of electronic health records in the clinical and biomedical fields makes the removal of protected health information (PHI) essential to maintain privacy. However, a significant portion of information is recorded in unstructured textual forms, posing a challenge for deidentification. In multilingual countries, medical records could be written in a mixture of more than one language, referred to as code mixing. Most current clinical natural language processing techniques are designed for monolingual text, and there is a need to address the deidentification of code-mixed text. ObjectiveThe aim of this study was to investigate the effectiveness and underlying mechanism of fine-tuned pretrained language models (PLMs) in identifying PHI in the code-mixed context. Additionally, we aimed to evaluate the potential of prompting large language models (LLMs) for recognizing PHI in a zero-shot manner. MethodsWe compiled the first clinical code-mixed deidentification data set consisting of text written in Chinese and English. We explored the effectiveness of fine-tuned PLMs for recognizing PHI in code-mixed content, with a focus on whether PLMs exploit naming regularity and mention coverage to achieve superior performance, by probing the developed models’ outputs to examine their decision-making process. Furthermore, we investigated the potential of prompt-based in-context learning of LLMs for recognizing PHI in code-mixed text. ResultsThe developed methods were evaluated on a code-mixed deidentification corpus of 1700 discharge summaries. We observed that different PHI types had preferences in their occurrences within the different types of language-mixed sentences, and PLMs could effectively recognize PHI by exploiting the learned name regularity. However, the models may exhibit suboptimal results when regularity is weak or mentions contain unknown words that the representations cannot generate well. We also found that the availability of code-mixed training instances is essential for the model’s performance. Furthermore, the LLM-based deidentification method was a feasible and appealing approach that can be controlled and enhanced through natural language prompts. ConclusionsThe study contributes to understanding the underlying mechanism of PLMs in addressing the deidentification process in the code-mixed context and highlights the significance of incorporating code-mixed training instances into the model training phase. To support the advancement of research, we created a manipulated subset of the resynthesized data set available for research purposes. Based on the compiled data set, we found that the LLM-based deidentification method is a feasible approach, but carefully crafted prompts are essential to avoid unwanted output. However, the use of such methods in the hospital setting requires careful consideration of data security and privacy concerns. Further research could explore the augmentation of PLMs and LLMs with external knowledge to improve their strength in recognizing rare PHI.

The driving behaviors of electric vehicle (EV) and hybrid electric vehicle (HEV) drivers have received considerable attention in the literature. The use of image recognition in combination with GPS and driving data has emerged as a popular approach to improving driver safety. However, such methods often generate sensitive personal information, including driver images, names, and GPS locations, which may risk the safety of the driver’s privacy. To address this issue, a privacy-preserving approach for identifying driver behavior characteristics is necessary. To achieve this, we utilize on-board diagnostic (OBD) interface vehicle-mounted devices to collect and analyze data from an electronic control unit (ECU), thereby collecting only onboard data for data processing and analysis. In this work, we propose deep learning models such as long short-term memory (LSTM) networks and gated recurrent units (GRUs), which enable the learning of specific behavior patterns and represent the state-of-the-art model to classify and predict driving behavior that potentially leads to dangerous accidents. The predictions were then converted into alarm signals and transmitted to the dashboard of vehicle. Our experiments showed that our proposed system model achieved excellent results with an excellent kappa score of 96.5%, demonstrating that it can accurately identify unique driving behaviors in a privacy-preserving manner.

Grid faults are found to be one of the major issues in renewable energy systems, particularly in wind energy conversion systems (WECS) connected to the grid via back-to-back (BTB) converters. Under such faulty grid conditions, the system requires an effective regulation of the active (P) and reactive (Q) power to accomplish low voltage ride through (LVRT) operation in accordance with the grid codes. In this paper, an improved finite-control-set model predictive control (FCS-MPC) scheme is proposed for a PMSG based WECS to achieve LVRT ability under symmetrical and asymmetrical grid faults, including mitigation of DC-link voltage fluctuation. With proposed predictive control, optimized switching states for cost function minimization with weighing factor (WF) selection guidelines are established for robust BTB converter control and reduced cross-coupling amid P and Q during transient conditions. Besides, grid voltage support is provided by grid side inverter control to inject reactive power during voltage dips. The effectiveness of the FCS-MPC method is compared with the conventional proportional-integral (PI) controller in case of symmetrical and asymmetrical grid faults. The simulation and experimental results endorse the superiority of the developed FCS-MPC scheme to diminish the fault effect quickly with lower overshoot and better damping performance than the traditional controller.

To achieve successful investments, in addition to financial expertise and knowledge of market information, a further critical factor is an individual’s personality. Decisive people tend to be able to quickly judge when to invest, while calm people can analyze the current situation more carefully and make appropriate decisions. Therefore, in this study, we developed a multimodal personality-recognition system to understand investors’ personality traits. The system analyzes the personality traits of investors when they share their investment experiences and plans, allowing them to understand their own personality traits before investing. To perform system functions, we collected digital human behavior data through video-recording devices and extracted human behavior features using video, speech, and text data. We then used data fusion to fuse human behavior features from heterogeneous data to address the problem of learning only one-sided information from a single modality. Through several experiments, we demonstrated that multimodal (i.e., three different signal inputs) personality trait analysis is more accurate than unimodal models. We also used statistical methods and questionnaires to evaluate the correlation between the investor’s personality traits and risk tolerance. It was found that investors with higher openness, extraversion, and lower neuroticism personality traits took higher risks, which is similar to research findings in the field of behavioral finance. Experimental results show that, in a case study, our multimodal personality prediction system exhibits high performance with highly accurate prediction scores in various metrics.

Zero-emission transportation is currently a public priority, especially in big cities. For this reason, the use of electric vehicles (EVs) is receiving much attention. To facilitate the adoption of EVs, a proper charging infrastructure together with energy management is essential. This article proposes a design guideline for a direct current (DC) charging station with bipolar properties. A bipolar system can convert a two-wire system into three wires in a microgrid system with a neutral line. The configuration of the bipolar system supports different loads; therefore, the unbalanced operation is inherent to the system. The proposed bipolar DC charging station (CS) has a three-level balancing converter that reduces the step-down effort chargers. Moreover, this paper proposes the continuous-control-set model predictive control (CCS-MPC)-based balancing strategy that allows the handling of different output loads while keeping the neutral-line voltage efficiently regulated with improved dynamic performance compared to a traditional controller. Stability and parameter robustness analyses are also performed for the control parameter selection. To ensure the performance of the proposed method, both simulation and experimental results are presented and compared with those obtained from the traditional methods.

Aroma and taste have long been considered important indicators of quality coffee. Specialty coffee, that is, coffee from a single estate, farm, or village in a coffee-growing region, in particular, has a unique aroma that reflects the coffee-producing region. In order to enable the traceability of coffee origin, in this study we have developed an e-nose system to discriminate the aroma of freshly roasted coffee in different production regions. In the case study, we employed the e-nose system to experiment with various machine learning models for recognizing several collected coffee beans such as coffees from Yirgacheffe and Kona. Additionally, our contribution also includes the development of a method to create an aromatic digital fingerprint of a specific coffee bean to identify its origin. The experimental results show that the developed e-nose system achieves good recognition performance for coffee aroma recognition. The extracted digital fingerprints have great potential to be stored in an extensible coffee aroma database similar to a comprehensive library of specific coffee bean aroma characteristics, for traceability and reconfirmation of their origin.

Recently, an emerging application field through Twitter messages and algorithmic computation to detect real-time world events has become a new paradigm in the field of data science applications. During a high-impact event, people may want to know the latest information about the development of the event because they want to better understand the situation and possible trends of the event for making decisions. However, often in emergencies, the government or enterprises are usually unable to notify people in time for early warning and avoiding risks. A sensible solution is to integrate real-time event monitoring and intelligence gathering functions into their decision support system. Such a system can provide real-time event summaries, which are updated whenever important new events are detected. Therefore, in this work, we combine a developed Twitter-based real-time event detection algorithm with pre-trained language models for summarizing emergent events. We used an online text-stream clustering algorithm and self-adaptive method developed to gather the Twitter data for detection of emerging events. Subsequently we used the Xsum data set with a pre-trained language model, namely T5 model, to train the summarization model. The Rouge metrics were used to compare the summary performance of various models. Subsequently, we started to use the trained model to summarize the incoming Twitter data set for experimentation. In particular, in this work, we provide a real-world case study, namely the COVID-19 pandemic event, to verify the applicability of the proposed method. Finally, we conducted a survey on the example resulting summaries with human judges for quality assessment of generated summaries. From the case study and experimental results, we have demonstrated that our summarization method provides users with a feasible method to quickly understand the updates in the specific event intelligence based on the real-time summary of the event story.

In recent years, many countries have provided promotion policies related to renewable energy in order to take advantage of the environmental factors of sufficient sunlight. However, the application of solar energy in the power grid also has disadvantages. The most obvious is the variability of power output, which will put pressure on the system. As more grid reserves are needed to compensate for fluctuations in power output, the variable nature of solar power may hinder further deployment. Besides, one of the main issues surrounding solar energy is the variability and unpredictability of sunlight. If it is cloudy or covered by clouds during the day, the photovoltaic cell cannot produce satisfactory electricity. How to collect relevant factors (variables) and data to make predictions so that the solar system can increase the power generation of solar power plants is an important topic that every solar supplier is constantly thinking about. The view is taken, therefore, in this work, we utilized the historical monitoring data collected by the ground-connected solar power plants to predict the power generation, using daily characteristics (24 h) to replace the usual seasonal characteristics (365 days) as the experimental basis. Further, we implemented daily numerical prediction of the whole-point power generation. The preliminary experimental evaluations demonstrate that our developed method is sensible, allowing for exploring the performance of solar power prediction.

The risk of supply chain disruption is usually related to daily disturbances in supply chain operations (e.g., demand fluctuations) and some emergency risks, such as earthquakes and epidemic outbreaks. During a crisis, companies need agility to quickly find new suppliers and open auxiliary sales channels to meet customer needs and remain competitive. However, identifying “event” is one of the most difficult challenges of current decision support systems. If the system encounters an emergency, it is usually unable to promptly notify users of the warning to avoid risks. A sensible solution is to incorporate the real-time event-monitoring system into SCM (i.e., supply chain management) in order to share emergency information in the early stage for preemptive management in the supply chain. On the other hand, in order to process confidential supply chain data with other members, the SCM infrastructure requires secure data sharing. The blockchain-based SCM system can improve the transparency of traceability to ensure that the supply chain system provides high-quality products and protects data privacy and security. The view is taken; therefore, in this work, we combined a method of real-time event detection using collected Twitter data and blockchain technology for event monitoring to improve the visibility of the supply chain system and take preemptive measures for risk avoidance. The experiments show some interesting results and potentials for future work in the field of the agile supply chain.

In this paper, we address a big-data analysis method for estimating the driving range of an electric vehicle (EV), allowing drivers to overcome range anxiety. First, we present an estimating approach to project the life of battery pack for 1600 cycles (i.e., 8 years/160 000 km) based on the data collected from a cycle-life test. This approach has the merit of simplicity. In addition, it considers several critical issues that occur inside battery packs, such as the dependence of internal resistance and the state-of-health. Subsequently, we describe our work on driving pattern analysis of an EV, using a machine-learning approach, namely growing hierarchical self-organizing maps, to cluster the collected EV big data. This paper contains the analysis of energy consumption and driving range estimation for EVs, including powertrain simulation and driving behavior analysis. The experimental results, including both simulating battery degradation and analysis of driving behaviors, demonstrate a feasible solution for improving driving range estimation by the EV big data.