A5035214368- Topic Modeling
- Speech and dialogue systems
- Natural Language Processing Techniques
- Multimodal Machine Learning Applications
- Immune Cell Function and Interaction
- IL-33, ST2, and ILC Pathways
- Energy and Environment Impacts
- Video Analysis and Summarization
- Misinformation and Its Impacts
- IoT and Edge/Fog Computing
- Microplastics and Plastic Pollution
- Heavy Metal Exposure and Toxicity
- Animal testing and alternatives
- Immunotoxicology and immune responses
- AI in Service Interactions
- Air Quality and Health Impacts
- Autophagy in Disease and Therapy
- IoT Networks and Protocols
- Data Stream Mining Techniques
- Natural product bioactivities and synthesis
- Effects and risks of endocrine disrupting chemicals
- biodegradable polymer synthesis and properties
- Advanced Nanomaterials in Catalysis
- Phosphorus and nutrient management
- Machine Learning and Data Classification
Xi’an University of Posts and Telecommunications
2024
Chinese Academy of Sciences
2019-2023
Peking University
2021-2023
Institute of Computing Technology
2019-2023
China National Center for Food Safety Risk Assessment
2022
Binzhou Medical University
2021
Binzhou University
2020-2021
University of Chinese Academy of Sciences
2021
Northeastern University
2019
Huazhong University of Science and Technology
2019
Configuration tuning is vital to optimize the performance of database management system (DBMS). It becomes more tedious and urgent for cloud databases (CDB) due diverse instances query workloads, which make administrator (DBA) incompetent. Although there are some studies on automatic DBMS configuration tuning, they have several limitations. Firstly, adopt a pipelined learning model but cannot overall in an end-to-end manner. Secondly, rely large-scale high-quality training samples hard...
Abstract The extensive existing of microplastics (MPs) in the ecosystem have increased considerable attention concerning their potential adverse effects, toxicities and underlying mechanism MPs are still scarce. To explore effect on cardiac tissue Wistar rats unravel pyroptosis oxidative stress process cardiomyocytes injury, 32 male Wister were divided into control group three model groups, which exposed to 0.5 mm PS at 0.5, 5 50 mg/L for 90 days. Results revealed that could damage structure...
Document Grounded Conversations is a task to generate dialogue responses when chatting about the content of given document. Obviously, document knowledge plays critical role in Conversations, while existing models do not exploit this kind effectively enough. In paper, we propose novel Transformer-based architecture for multi-turn grounded conversations. particular, devise an Incremental Transformer encode utterances along with related documents. Motivated by human cognitive process, design...
Audio-Visual Scene-Aware Dialog (AVSD) is a task to generate responses when chatting about given video, which organized as track of the 8th System Technology Challenge (DSTC8). There are two challenges in this task: 1) making effective interaction among different modalities; 2) better understanding dialogues and generating informative responses. To tackle challenges, we propose universal multimodal transformer introduce multi-task learning method learn joint representations modalities well...
Recent studies in dialogue state tracking (DST) leverage historical information to determine states which are generally represented as slot-value pairs. However, most of them have limitations efficiently exploit relevant context due the lack a powerful mechanism for modeling interactions between slot and history. Besides, existing methods usually ignore imbalance problem treat all slots indiscriminately, limits learning hard eventually hurts overall performance. In this paper, we propose...
Zekang Li, Jinchao Zhang, Zhengcong Fei, Yang Feng, Jie Zhou. Proceedings of the 59th Annual Meeting Association for Computational Linguistics and 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.
Audio-Visual Scene-Aware Dialog (AVSD) is a task to generate responses when chatting about given video, which organized as track of the 8th System Technology Challenge (DSTC8). To solve task, we propose universal multimodal transformer and introduce multi-task learning method learn joint representations among different modalities well informative fluent responses. Our extends natural language generation pre-trained model dialogue task. system achieves best performance in both objective...
Document Grounded Conversations is a task to generate dialogue responses when chatting about the content of given document. Obviously, document knowledge plays critical role in Conversations, while existing models do not exploit this kind effectively enough. In paper, we propose novel Transformer-based architecture for multi-turn grounded conversations. particular, devise an Incremental Transformer encode utterances along with related documents. Motivated by human cognitive process, design...
Black carbon (BC) is an important component of atmospheric PM 2.5 and the second largest contributor to global warming. 1,4-naphthoquinone-coated BC (1,4 NQ-BC) a secondary particle with great research value, so we chose 1,4 NQ-BC as object. In our study, mitochondria lysosomes were selected targets confirm whether they impaired by NQ-BC, label free proteomics technology, fluorescent probes, qRT-PCR western blots used investigate mechanism toxicity. We found 494 differentially expressed...
A good open-domain chatbot should avoid presenting contradictory responses about facts or opinions in a conversational session, known as its consistency capacity.However, evaluating the capacity of is still challenging.Employing human judges to interact with chatbots on purpose check their capacities costly and low-efficient, difficult get rid subjective bias.In this paper, we propose Addressing Inquiries History (AIH), an efficient practical framework for evaluation.At conversation stage,...
The present manuscript introduces a method for evaluating and forecasting the quality of vehicle-to-infrastructure (V2I) communication channels in urban settings. This precisely classifies predicts channel levels V2I scenarios based on long-range (LoRa) technology. approach aims to accurately classify predict scenarios. concept scoring was first introduced, offering more precise description compared traditional packet reception rate (PRR) assessments. In assessment model gated recurrent unit...
Phosphorous is vital for life, including plants, animals, and human beings it an essential component in agricultural fertilizers phosphorous-based chemicals. Phosphate ores, mainly the form of apatite collophanite, are non-renewable natural resources phosphorus. Froth flotation has been used phosphate industry more than half a century as primary technique upgrading phosphate. Nevertheless, does not produce satisfactory performance beneficiation many cases, especially very fine coarse...
Recent studies in dialogue state tracking (DST) leverage historical information to determine states which are generally represented as slot-value pairs. However, most of them have limitations efficiently exploit relevant context due the lack a powerful mechanism for modeling interactions between slot and history. Besides, existing methods usually ignore imbalance problem treat all slots indiscriminately, limits learning hard eventually hurts overall performance. In this paper, we propose...
As a kind of new expression elements, Internet memes are popular and extensively used in online chatting scenarios since they manage to make dialogues vivid, moving, interesting. However, most current dialogue researches focus on text-only tasks. In this paper, we propose task named as \textbf{M}eme incorporated \textbf{O}pen-domain \textbf{D}ialogue (MOD). Compared previous tasks, MOD is much more challenging it requires the model understand multimodal elements well emotions behind them. To...
Environmental pollution is a risk factor for adverse birth outcomes, especially preterm (PTB) and early-term (ETB). It has been revealed that exposure to fine particulate matter (PM2.5) during pregnancy increase the prevalence of PTB. However, relationship between PM2.5 ETB not elucidated. In high-risk pregnancies, whether will bring higher PTB than in normal pregnancies still unclear, susceptible window obscure. Therefore, it worthy assessing on identifying windows pregnant women.This paper...
Black carbon (BC) correlates with the occurrence and progression of atherosclerosis other cardiovascular diseases. Increasing evidence has demonstrated that BC could impair vascular endothelial cells, but underlying mechanisms remain obscure. It is known IL-33 exerts a significant biological role in disease, little about molecular regulation expression at present. We first found significantly increased mRNA EA.hy926 cells concentration time-dependent manner, we conducted this study to...
This paper introduces the Tenth Dialog System Technology Challenge (DSTC-10). edition of DSTC focuses on applying end-to-end dialog technologies for five distinct tasks in systems, namely 1. Incorporation Meme images into open domain dialogs, 2. Knowledge-grounded Task-oriented Dialogue Modeling Spoken Conversations, 3. Situated Interactive Multimodal 4. Reasoning Audio Visual Scene-Aware Dialog, and 5. Automatic Evaluation Moderation Open-domain Systems. describes task definition, provided...
In conversational machine comprehension, it has become one of the research hotspots integrating history information through question reformulation for obtaining better answers. However, existing models are trained only using supervised labels annotated by annotators without considering any feedback from this paper, we propose a novel Answer-Supervised Question Reformulation (ASQR) model enhancing comprehension with reinforcement learning technology. ASQR utilizes pointer-copy-based as an...
Nowadays, open-domain dialogue models can generate acceptable responses according to the historical context based on large-scale pre-trained language models. However, they generally concatenate history directly as model input predict response, which we named flat pattern and ignores dynamic information flow across utterances. In this work, propose DialoFlow model, in introduce a mechanism flow, design three training objectives capture dynamics utterances by addressing semantic influence...
Abstract Yttrium is a typical heavy rare earth element with widespread use in numerous sectors. Only one previous study has indicated that yttrium the potential to cause developmental immunotoxicity (DIT). Therefore, there remains paucity of evidence on DIT yttrium. This aimed explore nitrate (YN) and self‐recovery YN‐induced DIT. Dams were treated 0, 0.2, 2, 20 mg/kg bw/day YN by gavage during gestation lactation. No significant changes found innate immunity between control YN‐treated...
Recently, variational auto-encoder (VAE) based approaches have made impressive progress on improving the diversity of generated responses. However, these methods usually suffer cost decreased relevance accompanied by improvements. In this paper, we propose a novel multimodal response generation framework with exemplar augmentation and curriculum optimization to enhance First, unlike existing VAE-based models that approximate simple Gaussian posterior distribution, present mixture...