Tianyu Yang
Doctoral Researcher
Research Interest
Tianyu’s current research focuses on the development of multi-modal document processing system for intelligent and automatic data ingestion. Besides, he is also interested in the general generative model and reinforcement learning.
His primary research interest can be summarized as:
- Natural Language Processing
- Reinforcement Learning
- Variational Inference
- Multi-modal Large Language Models
- Information Extraction
SHORT CV
12/2024 – Present
Ph.D Student
University of Göttingen, Germany
08/2022 – 07/2024
Doctoral Researcher
Technichal University of Darmstadt, Germany
09/2019 – 06/2022
M.Sc. Computer Science
Civil Aviation University of China, China
09/2015 – 07/2019
B.Sc. Aircraft Manufacturing Engineering
Civil Aviation University of China, China
SELECTED PUBLICATIONS
Robust Utility-Preserving Text Anonymization Based on Large Language Models
Yang, Tianyu, Xiaodan Zhu, and Iryna Gurevych.
arXiv preprint arXiv:2407.11770 (2024).
(PDF)
Dior-CVAE: Pre-trained Language Models and Diffusion Priors for Variational Dialog Generation
Tianyu Yang, Thy Thy Tran, and Iryna Gurevych
In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 4718–4735, Singapore. Association for Computational Linguistics.
(PDF)
HTKG: Deep keyphrase generation with neural hierarchical topic guidance
Zhang, Yuxiang, Tao Jiang, Tianyu Yang, Xiaoli Li, and Suge Wang.
In Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval, pp. 1044-1054. 2022.
(PDF)
Hyperbolic deep keyphrase generation
Zhang, Yuxiang, Tianyu Yang, Tao Jiang, Xiaoli Li, and Suge Wang
In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, pp. 521-536. Cham: Springer International Publishing, 2022.
(PDF)
Boosting KG-to-Text Generation via Multi-granularity Graph Representations
Yang, Tianyu, Yuxiang Zhang, and Tao Jiang
In 2022 International Joint Conference on Neural Networks (IJCNN), pp. 1-9. IEEE, 2022.
(PDF)
Order-guided deep neural network for emotion-cause pair prediction
Fan, Wei, Yuexuan Zhu, Ziyun Wei, Tianyu Yang, W. H. Ip, and Yuxiang Zhang
Applied Soft Computing 112 (2021): 107818.
(PDF)
NEWS
Launch event of the Interdisciplinary Center for Machine Learning and Data Analysis (IZMD)
The Interdisciplinary Center for Machine Learning and Data Analysis (IZMD) has officially launched at the University of Wuppertal, on March 25th, 2019. Our group is excited to be closely associated with this newly founded center, and Bela Gipp is looking [...]
Funding granted for a 3-year, interdisciplinary research project
The Heidelberger Akademie der Wissenschaften has awarded us 3-year research grant for the project „Fake News and Collective Decision Making: Rapid Automated Assessment of Media Bias“ The project aims to automatically identify specific instances of media bias, thus ultimately helping [...]
Article about our group member Felix Petersen
An article about our group member Felix Petersen appeared in campus.kn. Felix won several prizes at Jugend forscht and will complete his Bachelor in Computer Science at the age of 19. Full article about Felix is published in campus.kn.
Article in Südkurier
The Südkurier reports on our Decentralized Trusted Timestamping project Originstamp.org.
German supreme court cites our research in its ruling on dashcam recordings
If you have ever browsed viral videos, chances are, you have come across accident footage recorded by dashcams. Dashboard cameras, or Dashcams, have become increasingly wide-spread among motorists. Especially in countries where drivers are more likely to ignore the rules [...]
Talk on blockchain technologies at the ‘Bodensee Mittelstand 4.0’ talk series
Bela Gipp will give a talk on blockchain technologies and visions at the event 'Blockchain – eine Kette an Möglichkeiten für Unternehmen' organized by the Wirtschaftsförderung Bodenseekreis and the competence centre cyberLAGO. https://www.youtube.com/watch?v=0Fxb6QgHIDg
Project ‘Open Legal Data’ wins faculty4makers Challenge 2017/18
The project openlegaldata.io founded by our group member (doctoral researcher) Malte Schwarzer has won the 1st price of this year’s faculty4makers challenge. The competition was organized by Technische Universität Berlin for DIY-projects that students carry out in their spare time outside of the [...]
Article about Media Bias research in uni’kon
An article about our research on the automated identification of media bias appeared in uni'kon #68. For further information and contact information refer to Felix Hamborg's profile page or see here.
Visiting Researcher
We are happy to announce that Dr. Tomas Foltynek of the Mendel University in Brno, Czechia, joined our research group for a 6-month research stay. Tomas is an expert for topics related to academic integrity and is especially interested in [...]
Talk on the immutable research data trail at conference on Blockchain for Science
Bela Gipp gave a talk about the immutable research data trail at the 1st international Conference on Blockchain for Science, Research and Knowledge Creation held on 5-6 November in Berlin. https://www.youtube.com/watch?v=HUNVXc_2uJc
The eCoach teaching innovation project receives funding from Wikimedia Germany, the Stifterverband, and the Volkswagen Stiftung
eCoach, an abbreviation for Electronic Coding Assigments Checker, is a teaching innovation project initiated at the Department of Computer and Information Science at the University of Konstanz. In 2017, the Department of Computer and Information Science provided seed funding to the eCoach project. The [...]
Our blockchain-based Dashcam App in the news
An article about the Dashcam App in BTC-ECHO based on OriginStamp project. Read the article here. To learn more about the Dashcam-APP, you may want to read our paper that we’ve published in 2016: B. Gipp, J. Kosti, and C. Breitinger, [...]
Interview: Blockchain technology research in “Die Welt” & “NDR Info”
Click here to open the online version of the interview in "Die Welt". Bela Gipp was also interviewed by NDR Info (audio download).
New DFG Project about math retrieval
The DFG, Germany’s organization for research funding, has awarded our group a 3-year research grant for our project: “Methods and Tools to Advance the Retrieval of Mathematical Knowledge from Digital Libraries for Search-, Recommendation- and Assistance-Systems” (Click here for more [...]
Best Student Paper Award at the JCDL 2017
Felix Hamborg and Norman Meuschke received the Best Student Paper Award for their paper Matrix-based News Aggregation: Exploring different News Perspectives; at the A* ranked Joint Conference on Digital Libraries (JCDL) 2017.
Blockchain Innovation: Collaboration with the International Blockchain Competition
Bela Gipp will serve as the University partner for the world’s largest ‘Blockchain Competition’ organized by the Swiss investor group Lakeside Partners AG. The call for entries is now open. Organizers are looking forward to receiving many submissions describing innovative [...]
Research cooperation with Human Bios International AG on blockchain-based security technologies
The Human Bios International AG is announcing a research cooperation with our group to explore the use of blockchain technology for security applications. In the scope of the cooperation, Human Bios is providing research stipends for up to four PhD positions within [...]
Norman Meuschke represents ACM SIGIR at the 50th Turing Award Celebration
The Special Interest Group on Information Retrieval (SIGIR) within the Association for Computing Machinery (ACM) has selected Norman Meuschke as one of ten students to represent SIGIR at the 50th ACM Turing Award Celebration in San Francisco, on June 23 - [...]
Talk at the department of computer science at the TUM Research Center Garching
Bela Gipp will give a talk on “Building Applications on the Blockchain of Crypto Currencies”, at the Department of Computer Science, TUM Forschungszentrum Garching on November 5th, 2016, at 11:00 in lecture hall #2. Everyone is welcome to attend.
Article about the research internship at NII in Tokyo
Felix Hamborg shared his experience of the research internship at National Institute of Informatics (NII) in Tokyo in the uni'kon magazine. In the article, he explains which opportunities the internship program offers, both in conducting research and exploring Japan. To read the full [...]
Our blockchain research on Spektrum.de
The German magazine Spektrum der Wissenschaft (PDF) covered our research on blockchain technologies (for instance, used in the cryptocurrency Bitcoin).
Paper about the link-based recommendations covered in Wikimedia Research Newsletter
The Wikimedia Research Newsletter, a monthly online overview of recent academic research about Wikipedia and other Wikimedia projects, reports on our paper "Evaluating Link-based Recommendations for Wikipedia". We will present the paper at the ACM/IEEE Joint Conference on Digital Libraries in Newark, NJ, US in [...]
Improved math rendering in Wikipedia
Wikipedia is using a new approach for rendering mathematical formulae - as of May 31st. This approach was developed by our group member Moritz Schubotz. Bitmap images representing formulae, were replaced with a combination of MathML (the web standard for [...]
CitePlag on public radio
The national public radio broadcasters Deutschlandfunk and Deutschlandradio Kultur covered our collaborative research on Plagiarism Prevention and Detection and our prototype system CitePlag.
Talk at the University of Copenhagen on March 4th
Bela Gipp will give a public talk on “New Approaches for Intellectual Property Protection” at the University of Copenhagen, Denmark on March 4, 2016, 16:00-17:00. The event is jointly organized by the Centre for Communication and Computing and the Royal School of [...]
