- Topic Modeling
- Natural Language Processing Techniques
- Nursing Diagnosis and Documentation
- Electronic Health Records Systems
- Recommender Systems and Techniques
- Human Mobility and Location-Based Analysis
- Advanced Image and Video Retrieval Techniques
- Domain Adaptation and Few-Shot Learning
- Digital Marketing and Social Media
- Embedded Systems Design Techniques
- Video Surveillance and Tracking Methods
- Parallel Computing and Optimization Techniques
- Interpreting and Communication in Healthcare
- Anomaly Detection Techniques and Applications
- Advanced Text Analysis Techniques
- Neural Networks and Applications
- Sentiment Analysis and Opinion Mining
- Speech Recognition and Synthesis
- Biometric Identification and Security
- Entrepreneurship Studies and Influences
- Advanced Image Fusion Techniques
- Face recognition and analysis
- Authorship Attribution and Profiling
- Machine Learning in Healthcare
- E-commerce and Technology Innovations
Anhui University
2024
Tibet University
2021-2023
Shandong Institute of Business and Technology
2023
Australian National University
2015-2019
Health Sciences and Nutrition
2019
Data61
2014-2019
Commonwealth Scientific and Industrial Research Organisation
2016-2019
Changchun University of Technology
2018
Eastern Liaoning University
2018
Shanghai University of Engineering Science
2014-2017
Over a tenth of preventable adverse events in health care are caused by failures information flow. These tangible clinical handover; regardless good verbal handover, from two-thirds to all this is lost after 3-5 shifts if notes taken hand, or not at all. Speech recognition and extraction provide way fill out handover form for proofing sign-off.The objective the study was recorded spoken annotated verbatim transcriptions, evaluations support research written natural language processing...
In named entity recognition, we often don't have a large in-domain training corpus or knowledge base with adequate coverage to train model directly.In this paper, propose method where, given data in related domain similar (but not identical) (NE) types and small amount of data, use transfer learning learn domain-specific NE model.That is, the novelty task setup is that assume just mismatch, but also label mismatch.
Word embeddings -- distributed word representations that can be learned from unlabelled data have been shown to high utility in many natural language processing applications. In this paper, we perform an extrinsic evaluation of five popular embedding methods the context four sequence labelling tasks: POS-tagging, syntactic chunking, NER and MWE identification. A particular focus paper is analysing effects task-based updating representations. We show when using as features, few several...
Abstract Objective We study the use of speech recognition and information extraction to generate drafts Australian nursing-handover documents. Methods Speech correctness clinicians’ preferences were evaluated using 15 recorder–microphone combinations, six documents, three speakers, Dragon Medical 11, five survey/interview participants. Information evaluation used 260 six-class classification for each word, two annotators, CRF++ conditional random field toolkit. Results A noise-cancelling...
Deep learning (DL) has been widely used to solve problems with success in speech recognition, visual object and detection for drug discovery genomics. Natural language processing achieved noticeable progress artificial intelligence. This gives an opportunity improve on the accuracy human-computer interaction of clinical informatics. However, due difference vocabularies context between a environment generic English, transplanting models directly from up-to-date methods real-world health care...
Nowadays, location based services (LBS) has become one of the most popular applications with rapid development mobile Internet technology. More and more research is focused on discovering required among massive information according to personalized behavior. In this paper, a collaborative filtering (CF) recommendation algorithm presented Location-aware Hidden Markov Model (LHMM). This approach includes three main stages. First, it clusters users by making pattern similarity calculation their...
Word embeddings -distributed word representations that can be learned from unlabelled data -have been shown to have high utility in many natural language processing applications.In this paper, we perform an extrinsic evaluation of four popular embedding methods the context sequence labelling tasks: part-of-speech tagging, syntactic chunking, named entity recognition, and multiword expression identification.A particular focus paper is analysing effects task-based updating representations.We...
In named entity recognition, we often don't have a large in-domain training corpus or knowledge base with adequate coverage to train model directly. this paper, propose method where, given data in related domain similar (but not identical) (NE) types and small amount of data, use transfer learning learn domain-specific NE model. That is, the novelty task setup is that assume just mismatch, but also label mismatch.
Precision livestock management requires animal traceability and disease trajectory, for which discriminating between or re-identifying individual animals is of significant importance. Existing re-identification (re-ID) methods are mostly proposed persons vehicles, compared with extraordinarily more challenging to be re-identified because subtle visual differences individuals. In this paper, we focus on image-based re-ID yaks (Bos grunniens), indispensable in local husbandry economy...
Tibetan medicinal materials play a significant role in culture. However, some types of share similar shapes and colors, but possess different properties functions. The incorrect use such may lead to poisoning, delayed treatment, potentially severe consequences for patients. Historically, the identification ellipsoid-like herbaceous has relied on manual methods, including observation, touching, tasting, nasal smell, which heavily rely technicians' accumulated experience are prone errors. In...
During clinical handover, clinicians exchange information about the patients and state of management. To improve care safety quality, both handover its documentation have been standardized. Speech recognition entity extraction provide a way to help health service providers follow these standards by implementing process as structured form, whose headings guide narrative, proofing sign-off automatically filled-out form. In this paper, we evaluate such systems. The form considers sections...
This paper describes our approach, called EPUTION, for the open trial of SemEval- 2018 Task 2, Multilingual Emoji Prediction. The task relates to using social media — more precisely, Twitter with its aim predict most likely associated emoji a tweet. Our solution this text classification problem explores idea transfer learning adapting classifier based on users’ tweeting history. experiments show that user-adaption method improves results by than 6 per cent macro-averaged F1. Thus, provides...
Entrepreneurial network plays an important role in entrepreneurship activities. In recent years, entrepreneurial becomes hot spot the area of Entrepreneurship. With support "social theory", exchange theory" and cognitive research has been developed to micro-and macro-levels gradually. The purposes this paper is analyses last ten proposes framework network, deeply default existing research, then propose prospects about future. And we hope that Literature Review On Network will give some...
An improved fruit fly optimization algorithm based on knowledge memory (KM-FOA) is proposed for solving continuous functions problems. It featured with mechanisms devised the concerned problems: (1) Direction vectors, a novel method, defined as knowledge. Each needs to learn and memorize direction most obvious concentration changes in search space iteratively; (2) we embedded vectors guide procedure of choosing food. Fruit swarm can optimize roads basis And this approach achieves goal...
With the expansion of mobile Internet, location-based services have become a hot spot inInternet industry.To improve accuracy and efficiency discovery, researchers in web recommendation area are still busy looking for method.In this paper, it proposes LCAMSP model (Location Context Awareness Mobile Service Prediction, LCAMSP) under Internet environment, aims to meet exact personal requirements users' current location preference.Then, similar grouping is also important thread predict...
Yaks (Bos grunniens) are the most important domestic animals for people living at high altitudes. In order to implement precise livestock management yaks, it is of significant importance automatically identify, keep track of, and monitor yaks. Traditional animal identification methods such as ear tags, tattoos, RFID based suffer from problems like infection, maintenance cost, inefficiency or sensor failure. Existing biometric-based muzzle prints, iris patterns, retinal vascular patterns...
Automatic segmentation and overlapping bigrams are the most common methods for overcoming lack of explicit word boundaries in Chinese text. Past studies have compared their effectiveness, but findings been equivocal site search has little studied. We compare representatives two approaches using a 465,000 page crawl test queries applicable to university context. 503 pairs result sets were judged by 56 students.
A bstract N ow adays, location based services (LB S) has becom e one of the m ost popular applications w ith rapid developm ent obile Internet environm ent.M ore and research is focused on discovering required am ong assive inform ation according to personalized behavior.In this paper,a collaborative filtering (CF) recom endation algorithm presented Location-aw are H idden M arkov odel(LH ).This approach includes three ain stages.First,itclusters users by aking a pattern sim ilarity...