Experience

• Participated in the scientific research project “Based on machine learning, the recognition of quantum state image and the regression analysis of magnetic field value are realized”.
  • May. 2019 - Oct. 2019
  • Tianjin, China
  • Advisor: Ding Liu
  • Based on the Pytorch framework, I built a deep residual network (ResNet), and conducted image classification experiments on Cifar10 and Cifar100 datasets respectively.Then continuously deepen the network to observe the model performance changes, find out resNet network deepening caused by the problem, and put forward the improvement plan.
• China Undergraduate Mathematical Contest in Modeling (CUMCM)
  • Sep. 2018
  • Tianjin, China
  • In the intelligent RGV scheduling problem, I was mainly responsible for programming, and I was also responsible for the discussion and establishment of the model with another person. For a given intelligent machining system, we set up optimization models of single objective and multi-objective under three conditions – single operation, double operation and possible failure, and discussed dynamic scheduling of RGV.
• Mathematical Contest In Modeling / Interdisciplinary Contest In Modeling （MCM/ICM）
  • Jun. 2019
  • Tianjin, China
  • In the Louvre problem, I was mainly responsible for programming, and I was also responsible for the discussion and establishment of the model with another person.Aiming at the shortest evacuation time, we establish optimization models to obtain the emergency evacuation plan of the Louvre. With the minimum evacuation time as the optimization goal, we established the emergency evacuation model of the Louvre and obtained the evacuation plan. The evacuation model consists of three parts, including a moving time model based on cellular automaton, a time model of going up and down stairs and a waiting time model based on queuing theory.
• China Students Service Outsourcing Innovation and Enterpreneurship Competition
  • May. 2019
  • Wuxi, jiangsu province, China
  • Our team took campus life as the starting point, based on source separation technology, paddlepaddles frame designed a product that can improve the quality of campus life and improve work efficiency – classroom denoising application based on voice source separation. In the team, I was mainly responsible for the construction of the software front-end platform and the speech during the competition. The front-end is developed based on the Android platform, with simple operation, humanized design and easy man-machine interaction. Users can obtain the de-noise frequency through the server by recording.
• MLA 2019
  • Nov. 2019
  • Tianjin, China
  • Attended a conference called The 17th China conference on machine learning and its applications,where I listened to the presentation of professors from universities all over China.