Rujing Yao

ORCID: 0009-0001-2738-5786
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Topic Modeling
  • Advanced Text Analysis Techniques
  • Machine Learning and Data Classification
  • Biomedical Text Mining and Ontologies
  • Imbalanced Data Classification Techniques
  • Constraint Satisfaction and Optimization
  • Machine Learning and Algorithms
  • Natural Language Processing Techniques
  • Rough Sets and Fuzzy Logic
  • Artificial Intelligence in Law
  • Legal Education and Practice Innovations
  • Multimodal Machine Learning Applications
  • Geographic Information Systems Studies
  • Data Quality and Management
  • Data Management and Algorithms
  • Gaussian Processes and Bayesian Inference
  • scientometrics and bibliometrics research
  • Adversarial Robustness in Machine Learning
  • Human Motion and Animation
  • 3D Surveying and Cultural Heritage
  • Anomaly Detection Techniques and Applications
  • Comparative and International Law Studies
  • Misinformation and Its Impacts
  • Augmented Reality Applications
  • Expert finding and Q&A systems

Nankai University
2022-2025

Tianjin University
2020-2024

Zhejiang Lab
2019

Jilin Agricultural University
2016-2017

10.1109/tkde.2025.3530916 article EN IEEE Transactions on Knowledge and Data Engineering 2025-01-01

Large Language Models (LLMs) have achieved impressive results across numerous domains, yet they experience notable deficiencies in legal question-answering tasks. LLMs often generate generalized responses that lack the logical specificity required for expert advice and are prone to hallucination, providing answers appear correct but unreliable. Retrieval-Augmented Generation (RAG) techniques offer partial solutions address this challenge, existing approaches typically focus only on semantic...

10.48550/arxiv.2502.07912 preprint EN arXiv (Cornell University) 2025-02-11

The rise of large language models has opened new avenues for users seeking legal advice. However, often lack professional knowledge, which can lead to questions that omit critical information. This deficiency makes it challenging traditional question-answering systems accurately identify users' actual needs, resulting in imprecise or generalized In this work, we develop a system called Intelligent Legal Assistant, interacts with precisely capture their needs. When user poses question, the...

10.48550/arxiv.2502.07904 preprint EN arXiv (Cornell University) 2025-02-11

10.1007/s41060-020-00231-3 article EN International Journal of Data Science and Analytics 2020-08-12

Evaluating academic papers and groups is important in scholar evaluation literature retrieval. However, current indices, which pay excessive attention to the citation number rather than importance unidirectionality, are relatively simple. This study proposes new indices for groups. First, an improved PageRank (PR) algorithm introducing proposed obtain a citation-based paper index (CPI) via pre-ranking fine-tuning strategy. Second, evaluate paper’s influence inside outside its research field,...

10.1177/01655515221105038 article EN Journal of Information Science 2022-07-21

Literature analysis facilitates researchers better understanding the development of science and technology. The conventional literature focuses on topics, authors, abstracts, keywords, references, etc., rarely pays attention to content papers. In field machine learning, involved methods (M) datasets (D) are key information in extraction mining M D useful for discipline algorithm recommendation. this paper, we propose a novel entity recognition model, called MDER, constructe from papers PAKDD...

10.1109/bigdata47090.2019.9006262 article EN 2021 IEEE International Conference on Big Data (Big Data) 2019-12-01

Weighting strategy prevails in machine learning. For example, a common approach robust learning is to exert low weights on samples which are likely be noisy or quite hard. This study summarizes another less-explored strategy, namely, perturbation. Various incarnations of perturbation have been utilized but it has not explicitly revealed. Learning with called and systematic taxonomy constructed for this study. In our taxonomy, divided the basis targets, directions, inference manners,...

10.1145/3644391 article EN ACM Transactions on Knowledge Discovery from Data 2024-02-03

There are representative models for spatial topological relations of Region connection calculus and intersections model. The purpose this paper is to study the multiple simple regions. In paper, 4-intersection matrix model extended into a 2n-intersection function represent exclusiveness completeness these proved. Moreover, also gives detailed discussion on properties applications from special new perspective. And among three regions in applied forecast impact typhoon specified target areas....

10.1049/cje.2017.07.005 article EN Chinese Journal of Electronics 2017-09-01

As the scale, shape and function of engineering construction projects become more complex, design building systems has complex.In order to help students better acquire course resource information, effectively demonstrate teaching process, understand curriculum learn effectively, this paper designs an Auto CAD network architecture system.We analyze compare model based on system, introduce cluster analysis management.An was constructed through collaborative among professionals technology.We...

10.14733/cadaps.2020.s2.145-157 article EN Computer-Aided Design and Applications 2020-01-24

The knowledge contained in academic literature is interesting to mine. Inspired by the idea of molecular markers tracing field biochemistry, three named entities, namely, methods, datasets and metrics are used as AI for literature. These entities can be trace research process described bodies papers, which opens up new perspectives seeking mining more valuable information. Firstly, entity extraction model this study extract from large-scale Secondly, original papers traced markers....

10.48550/arxiv.2011.00518 preprint EN other-oa arXiv (Cornell University) 2020-01-01

This study focuses on a reverse question answering (QA) procedure, in which machines proactively raise questions and humans supply the answers. procedure exists many real human-machine interaction applications. However, crucial problem is answer understanding. The existing solutions have relied mandatory option term selection to avoid automatic these led unnatural human-computer negatively affected user experience. To this end, current proposes novel deep understanding network, called...

10.48550/arxiv.1912.00398 preprint EN other-oa arXiv (Cornell University) 2019-01-01

Weighting strategy prevails in machine learning. For example, a common approach robust learning is to exert lower weights on samples which are likely be noisy or quite hard. This study reveals another undiscovered strategy, namely, compensating. Various incarnations of compensating have been utilized but it has not explicitly revealed. Learning with called compensation and systematic taxonomy constructed for this study. In our taxonomy, divided the basis targets, directions, inference...

10.48550/arxiv.2107.11921 preprint EN other-oa arXiv (Cornell University) 2021-01-01

Large-scale, high-quality data are considered an essential factor for the successful application of many deep learning techniques. Meanwhile, numerous real-world tasks still have to contend with lack sufficient amounts data. Additionally, issues such as model robustness, fairness, and trustworthiness also closely related training Consequently, a huge number studies in existing literature focused on aspect tasks. Some typical optimization techniques include augmentation, logit perturbation,...

10.48550/arxiv.2310.16499 preprint EN other-oa arXiv (Cornell University) 2023-01-01

Today, big data plays an important role. Electronic map has great research value as the core part of application data. retrieval is a hot spot in field electronic map. This paper makes some improvement to defect traditional ZZL algorithm that same tail character string initialized every time when we retrieve something so performance reduced. We introduce clustering analysis and propose improved algorithm. The experimental results show efficiency by 37.1% on average compared with geographic...

10.22323/1.299.0055 article EN cc-by-nc-nd 2017-07-17

Literature analysis facilitates researchers better understanding the development of science and technology. The conventional literature focuses on topics, authors, abstracts, keywords, references, etc., rarely pays attention to content papers. In field machine learning, involved methods (M) datasets (D) are key information in extraction mining M D useful for discipline algorithm recommendation. this paper, we propose a novel entity recognition model, called MDER, constructe from papers PAKDD...

10.48550/arxiv.1911.13096 preprint EN other-oa arXiv (Cornell University) 2019-01-01

Literature analysis facilitates researchers to acquire a good understanding of the development science and technology. The traditional literature focuses largely on metadata such as topics, authors, abstracts, keywords, references, etc., little attention was paid main content papers. In many scientific domains science, computing, engineering, methods datasets involved in papers published those carry important information are quite useful for domain well algorithm dataset recommendation. this...

10.48550/arxiv.2010.13583 preprint EN other-oa arXiv (Cornell University) 2020-01-01

With the development of economy and continuous progress society, computer has become an indispensable part People's Daily work life, making life more convenient improving efficiency industrial production.Computers are important in all kinds industries product design as well.The addition computers not only saves labor costs improves customer satisfaction, but also quality performance to a certain extent efficiency.It can be seen that process make detailed plans, understand current trend...

10.14733/cadaps.2023.s1.151-161 article EN Computer-Aided Design and Applications 2022-06-06
Coming Soon ...