A5042345460- Sentiment Analysis and Opinion Mining
- Mental Health via Writing
- Topic Modeling
- Conducting polymers and applications
- Organic Light-Emitting Diodes Research
- Organic Electronics and Photovoltaics
- Machine Learning in Healthcare
- Misinformation and Its Impacts
- Advanced Text Analysis Techniques
- Adsorption and biosorption for pollutant removal
- Fuel Cells and Related Materials
- Chemical Synthesis and Reactions
- Spam and Phishing Detection
- Advanced Neural Network Applications
- Luminescence and Fluorescent Materials
- Emotion and Mood Recognition
- Extraction and Separation Processes
- Digital Mental Health Interventions
- Layered Double Hydroxides Synthesis and Applications
- Oxidative Organic Chemistry Reactions
- Synthesis and properties of polymers
- Membrane-based Ion Separation Techniques
- Polyoxometalates: Synthesis and Applications
- Chemical Synthesis and Characterization
- Membrane Separation Technologies
University of Manchester
2021-2024
Jiangxi Normal University
2024
Wuhan University
2024
Xi'an Polytechnic University
2024
Lanzhou University
2023
Nanchang Institute of Technology
2022
University of Chinese Academy of Sciences
2018-2022
Shandong Normal University
2022
Changsha University of Science and Technology
2020-2021
Shanghai Ocean University
2020
The latest large language models (LLMs) such as ChatGPT, exhibit strong capabilities in automated mental health analysis. However, existing relevant studies bear several limitations, including inadequate evaluations, lack of prompting strategies, and ignorance exploring LLMs for explainability. To bridge these gaps, we comprehensively evaluate the analysis emotional reasoning ability on 11 datasets across 5 tasks. We explore effects different strategies with unsupervised distantly supervised...
As an integral part of people's daily lives, social media is becoming a rich source for automatic mental health analysis.As traditional discriminative methods bear poor generalization ability and low interpretability, the recent large language models (LLMs) have been explored interpretable analysis on media, which aims to provide detailed explanations along with predictions in zero-shot or few-shot settings.The results show that LLMs still achieve unsatisfactory classification performance...
Mental health is a critical issue in modern society, and mental disorders could sometimes turn to suicidal ideation without adequate treatment. Early detection of from social content provides potential way for effective intervention. Recent advances pretrained contextualized language representations have promoted the development several domain-specific models facilitated downstream applications. However, there are no existing healthcare. This paper trains release two masked models, i.e.,...
Stress and depression detection on social media aim at the analysis of stress identification tendency from posts, which provide assistance for early mental health conditions. Existing methods mainly model states post speaker implicitly. They also lack ability to mentalise complex state reasoning. Besides, they are not designed explicitly capture class-specific features. To resolve above issues, we propose a Knowledge–aware Contrastive Network (KC-Net). In detail, first extract knowledge...
Mental illnesses are one of the most prevalent public health problems worldwide, which negatively influence people's lives and society's health. With increasing popularity social media, there has been a growing research interest in early detection mental illness by analysing user-generated posts on media. According to correlation between emotions illness, leveraging fusing emotion information developed into valuable topic. In this article, we provide comprehensive survey approaches media...
A key challenge for Emotion Recognition in Conversations (ERC) is to distinguish semantically similar emotions. Some works utilise Supervised Contrastive Learning (SCL) which uses categorical emotion labels as supervision signals and contrasts high-dimensional semantic space. However, fail provide quantitative information between ERC also not equally dependent on all embedded features the space, makes SCL inefficient. To address these issues, we propose a novel low-dimensional Cluster-level...
Here, we conveniently designed and synthesized a self-host thermally activated delayed fluorescence (TADF) emitter, which can not only form uniform thin film through wet-process, but also allow the subsequently deposition of electron transporting layer (ETL) by orthogonal solvent. By using this material as all-solution-processed multilayer TADF organic light emitting diodes (OLEDs) was successfully fabricated. The maximum current, power external quantum efficiencies nondoped device are 46.3...
A small variation in mobile hardware and software can potentially cause a significant heterogeneity or the sensor data each device collects. For example, microphone accelerometer sensors on different devices respond very differently to same audio motion phenomena. Other factors, like instantaneous computational load smartphone, key behavior sampling rates fluctuate, further polluting data. When sensing are deployed unconstrained real-world conditions, examples of sharply lower classification...
Detecting rotated ships is difficult in optical remote sensing images due to the challenges of complex scenes. Existing advanced ship detectors are typically anchor-based algorithms that require plenty predefined anchors. However, use anchors brings three critical problems: 1) a large number bring huge amount calculation; 2) attributes (e.g., size and aspect ratios) designed via <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ad hoc</i>...
The latest large language models (LLMs) such as ChatGPT, exhibit strong capabilities in automated mental health analysis. However, existing relevant studies bear several limitations, including inadequate evaluations, lack of prompting strategies, and ignorance exploring LLMs for explainability. To bridge these gaps, we comprehensively evaluate the analysis emotional reasoning ability on 11 datasets across 5 tasks. We explore effects different strategies with unsupervised distantly supervised...
Depressive symptoms identification on social media aims to identify posts from expressing of depression. This can be beneficial for developing mental health support systems and understanding the The Patient Health Questionnaire-9 (PHQ-9) is an instrument that healthcare professionals widely use assess monitor However, most existing models only consider capturing semantic information posts, without considering PHQ-9 descriptive related symptoms. In addition, they are not devised capture...
Sentiment analysis and emotion detection are important research topics in natural language processing (NLP) benefit many downstream tasks. With the widespread application of large models (LLMs), researchers have started exploring LLMs based on instruction-tuning field sentiment analysis. However, these only focus single aspects affective classification tasks (e.g. sentimental polarity or categorical emotions), overlook regression strength intensity), which leads to poor performance The main...
This paper proposes a robust obstacle detection and recognition method for driver assistance systems. Unlike existing methods, our aims to detect recognize obstacles on the road rather than all in view. The proposed involves two stages aiming at an increased quality of results. first stage is locate positions road. In order accurately on-road obstacles, we propose based U-V disparity map generated from stereo vision system. algorithm makes use V-disparity that provides good representation...
Suicide is one of the leading causes death worldwide. At same time, widespread use social media has led to an increase in people posting their suicide notes online. Therefore, designing a learning model that can aid detection online great importance. However, current methods cannot capture both local and global semantic features. In this paper, we propose transformer-based named TransformerRNN, which effectively extract contextual long-term dependency information by using transformer encoder...
Background In recent years, the COVID-19 pandemic has brought great changes to public health, society, and economy. Social media provide a platform for people discuss health concerns, living conditions, policies during epidemic, allowing policymakers use this content analyze emotions attitudes decision-making. Objective The aim of study was deep learning–based methods understand on topics related in United Kingdom through comparative geolocation text mining analysis Twitter. Methods Over...
Pretrained language models have been used in various natural processing applications. In the mental health domain, domain-specific are pretrained and released, which facilitates early detection of conditions. Social posts, e.g., on Reddit, usually long documents. However, there no for long-sequence modeling domain. This paper conducts continued pretraining to capture context health. Specifically, we train release MentalXLNet MentalLongformer based XLNet Longformer. We evaluate classification...
LLMs have transformed NLP and shown promise in various fields, yet their potential finance is underexplored due to a lack of thorough evaluations the complexity financial tasks. This along with rapid development LLMs, highlights urgent need for systematic evaluation benchmark LLMs. In this paper, we introduce FinBen, first comprehensive open-sourced benchmark, specifically designed thoroughly assess capabilities domain. FinBen encompasses 35 datasets across 23 tasks, organized into three...
Multimodal fake news detection has garnered significant attention due to its profound implications for social security. While existing approaches have contributed understanding cross-modal consistency, they often fail leverage modal-specific representations and explicit discrepant features. To address these limitations, we propose a Inverse Attention Network (MIAN), novel framework that explores intrinsic discriminative features based on content advance detection. Specifically, MIAN...