Yingrui Ji

Yingrui Ji

ABOUT ME

Yingrui Ji is studying as a Ph.D student at University of Chinese Academy of Sciences in 2022. Her main research interests are in optimizing computer vision model training/inference, high-performance computing, high-performance deep learning systems, and enhancement for classifying low-quality image samples.

For additional details, see her Full CV.

EDUCATION BACKGROUND

  • Ph.D in Machine Learning and Computer Vision, Aerospace Information Research Institute, Chinese Academy of Sciences, 09/2022 - 06/2026
  • M.S. in Computer Science and Technology, Institute of Computing Technology (ICT), Chinese Academy of Sciences and Dalian Ocean University Joint Training, 09/2019 - 06/2022

PUBLICATIONS

  • Yingrui Ji, Vijaya Sindhoori Kaza, Nishanth Artham, Tianyang Wang, Deep Active Learning with Manifold-Preserving Trajectory Sampling. (Submitted to ICASSP 2025)

  • Zeyu Wang, Yizhuo Chang, Yingrui Ji, Zhongruo Wang, Yuwang Wang, Zhigang Li, Yiqing Shen, CausalHalEval: A Benchmark for Evaluating the Hallucination of LLMs from the Perspective of Casual Reasoning and Inference Capabilities. (Submitted to AAAI 2025)

  • Jiayi Guo, Zan Chen, Yizhuo Chang, Yingrui Ji, Daqin Luo, Liyun Zhang, Zhongruo Wang, Zhigang Li, Yiqing Shen, UniAutoML: A Human-Centered Framework for Unified Discriminative and Generative AutoML with Large Language Models. (Submitted to AAAI 2025)

  • Yingrui Ji, Yao Zhu, Zhigang Li, Jiansheng Chen, Yunlong Kong, Jingbo Chen. 2023. Advancing Out-of-Distribution Detection through Data Purification and Dynamic Activation Function Design. IEEE Transactions on Circuits and Systems for Video Technology(IEEE TCSVT)[J]. (Under Review) PDF

  • Zijie Ding, Yingrui Ji, Yan Gan, Yuwen Wang, Yukun Xia. 2023. Current Status and Development Trends of Technology, Methods, and Application Fields of Human-computer Intelligent Interaction: Bibliometric Research[J]. (available accept) PDF

  • Yukun Xia, Yingrui Ji, Yan Gan, Zijie Ding. 2023. Applying Ming furniture features to modern furniture design using deep learning. Artificial Intelligence, Social Computing and Wearable Technologies. PDF

  • Yan Gan，Yingrui Ji, Shuo Jiang, Xinxiong Liu, Zhipeng Feng, Yao Li, Yuan Liu. 2021. Integrating aesthetic and emotional preferences in social robot design: An affective design approach with Kansei engineering and a deep convolutional generative adversarial network. International journal of industrial ergonomics[J]. PDF

  • Shang H, Duan X, Li F, …… Yingrui Ji et al. 2021. Many-core acceleration of the first-principles all-electron quantum perturbation calculations. Computer Physics Communications[J]. PDF

HONORS AND AWARDS

  • 2019 The Tenth Blue Bridge Cup National Software and Information Technology Professional Talent Competition Liaoning Division 1st.
  • 2018 The Ninth Blue Bridge Cup National Software and Information Technology Professional Talent Competition Liaoning Division 1st
  • 2017 The Eighth Blue Bridge Cup National Software and Information Technology Professional Talent Competition Liaoning Division 2nd
  • 2017 The 4th College Student Mobile Application Development Competition Provincial 2nd
  • 2017 11th iCAN International Innovation and Entrepreneurship Competition Liaoning Division 3th
  • 2021 Second Class Scholarship for Parallel Software Group, Institute of Computing Technology, Chinese Academy of Sciences
  • 2022 Third Class Academic Scholarship of Parallel Software Group, Institute of Computing Technology, Chinese Academy of Sciences
  • 2022 Second Class Academic Scholarship of Dalian Ocean University

INTERN EXPERIENCE

  • 2023.07 - Present $\quad$ Qiyuan Lab
    We refine the four datasets, use the evaluation algorithm to reduce the noise in these datasets, and clear the ID data in the OOD data set to ensure that the evaluation results are more reliable.
    • Responsible for small algorithm model optimization and datasets optimization for Out-of-Distribution Detection.
  • 2022.07 - 2022.12 $\quad$ NXP Semiconductor Corporation
    Complete the conversion of the eiq-deepview model in the eiq-toolkit into the form required by the user, which is convenient for the user’s subsequent development, and at the same time output the model result in the form of a picture. Optimize the training code of eiq-toolkit, reduce redundancy, help the group train and test new project use cases and write test files, submit and solve bugs in the updated version of eiq-toolkit during the test.
    • Convert commercial models, optimize training codes, and test new version use cases.