§01 推荐信模板
【个人评价】您认识申请人的方式、认识的时间和了解的程度;并请对申请人思想品德、道德修养方面加以介绍:
本人曾在XSX同学本科二年级时,担任其专业课的授课教师,之后也在几次访谈活动中与该生有过交流,对其较为了解。该生有着务实的作风,曾为了保证期末报告的详实性,主动向我咨询很多细节;他也在研究生阶段主动承担党支部任务,与我交流关于“导学思政”的看法。几次交流下来,该生具有较好的语言表达能力,态度诚恳,并且认真筹备每次活动,非常踏实。
【能力评价】对申请人学术水平、科研能力、研究成果、知识结构、外语水平等的介绍:
XSX同学在本科阶段和硕士阶段都取得了良好的课业成绩,对基础理论和专业知识都有着不错的掌握;在本科阶段,顺利通过四、六级英语考试,具备阅读英文文献的能力。该生对于自己的研究领域有着深入的了解,并且思维活跃,能够创新性地提出解决问题的办法,具有较好的科研能力。
【推荐意见】申请人的特长和弱点,从课程学习和从事科研工作(如有)的情况看,该申请人是否具有创新的潜力,有无继续培养的前途,对申请人申请攻读清华大学博士学位的意见:
XSX同学勤于思考,学习能力较强;性格开朗,具有良好的心态与科研热情;对于自己的科研方向和人生目标有着清晰的认知,具有较好的科研潜力。该生在成果转化方面还需要进一步探索和强化,本人也相信其在博士阶段可以做好。鉴于此,本人推荐许舒翔同学攻读清华大学博士学位。
§02 推荐信样例
一、Yulin Cai
Dear Dean Takao Someya,
It is my great pleasure and honor to recommend Yulin Cai for his doctoral application for the School of Engineering, the University of Tokyo. I came to know Yulin since I serve as Yulin’s instructor of his master’s course at Tsinghua University. When hearing that he is pursuing an oversea PhD position, I would like to highly recommend him for his excellent learning ability, impressive intellect of interdisciplinary research and outstanding innovation spirit.
Yulin attended my course Artificial Neural Network in the 2020 fall semester. This course comprised a series of neural network algorithms developed for pattern recognition as well as their mathematical principles, while in this course I also designed several challenging assignments for students to train their abilities of algorithm implementation. As a result, Yulin resolved all these problems with superior quality, and many of his assignments also enjoyed the highest performance among my class. In the last lecture of this course, Yulin presented a PID algorithm for network training which was inspired from the automatic control theory. By introducing differential terms into the iteration process of network training, this proposed algorithm is able to enjoy a faster convergence rate compared with conventional training algorithms (including SGD and Adam) on many frequently-used machine learning datasets without loss of model performance. From his behaviors in my class, I think Yulin demonstrated his extraordinary capability to study with enthusiasm and creativity.
Currently, Yulin is working on Vision Radar technologies served for autonomous driving. In this project, Yulin uses freeform optics design strategies to establish the frontend point cloud acquisition system, and implements an improved U-Transformer as the backend processing algorithm. Such a novel combination of optical technology and machine learning has outperformed state-of-the-art results in reconstruction accuracy. This remarkable work has been submitted to TPAMI, which is a well-known international journal of artificial intelligence and I am quite optimistic for the review result.
I strongly believe that Yulin will become a preeminent researcher in his field of interest according to the exceptional academic potential he has shown during his master’s study. Therefore, I will give Yulin my highest recommendation. Please feel free to contact me at zhuoqing@tsinghua.edu.cn if you have any further questions regarding Yulin’s personal characteristics and academic capability.
Sincerely,
Associate Professor
Department of Automation, Tsinghua University