Biography

Before university

I was born in Xinxiang, Henan, a picturesque city nestled in the heart of China’s east-middle region, alongside the majestic Yellow River, the second-longest river in the country. With a history dating back thousands of years, Xinxiang is a city that embodies the rich cultural heritage and vibrant spirit of China. Across the ages, its strategic location at the crossroads of major transportation routes has made it a vital center for trade and commerce. This vibrant convergence has cultivated Xinxiang into a melting pot of cultures, where a diverse population of people from various ethnicities and backgrounds contributes to enrich the city’s traditions, cuisines, and customs. I grew up and spent the first 18 years of my life there.

Bachelor

After successfully completing the extremly competitive Chinese national higher education entrance examination (Gaokao), I started the studies at École Centrale de Pékin, Beihang University, in 2013. According to my understandings, this institute represents an experimental Sino-French higher education cooperation project, seeking to integrate the Chinese university educational system with the French grande école educational system. It is a seven-year program that involves Bachelor’s, Master’s and French engineering education. The courses during the first three years include intensive French language acquisition as well as rigorous instruction in foundational mathematics, physics, and engineering. The majority of scientific courses are conducted in French. This period is similar to the “classe préparatoire” in France, which acts as an intensive preparatory course for admission to the grandes écoles.

Double degree in France

Following three years of Bachelor’s studies, in 2016, I was fortunate enough (maybe also due to my diligent work 🤔) to obtain the opportunity for a two-year exchange to École Centrale Paris (now CentraleSupélec). The first year of studies at Centrale is dedicated to advanced mathematical and physical knowledge, covering a broad spectrum of topics such as probabilities, statistics, quantum and statistical physics, thermal transfer, and more. In the subsequent year, students have the freedom to select courses according to their interests. I consider myself very fortunate to have chosen the Advanced Statistics course taught by Professor Christine Keribin and the Machine Learning course instructed by Professor Chloé-Agathe Azencott. These two courses have sparked my initial interests towards the domain of statistical machine learning. Beyond the academic parts, the initial experience of living in France is also deeply memorable, despite all the challenges of cultural integration at the beginning. The exchange at Centrale concluded in 2018, after which I returned to China to complete my Master’s education. This involved undertaking general courses, conducting a master research thesis and completing a mandatory six-month internship. I obtained my Master’s degree in June 2020 and commenced my PhD studies in Grenoble in November 2020.

Industrial experiences

Throughout my academic eductation journey, I have also made several tentative efforts to discover the industrial world.

During the summer of 2017, I conducted a two-month internship at Schlumberger in Beijing. This internship involved applying fluid mechanics knowledge, specifically the Bernoulli Equation, to analyze the velocity of oil flow within pipes. It was my first experience of applying theoretical knowledge from coursework to solve real-world problem.

In the summer of 2018, I returned to Schlumberger for a three-month internship. This time, my focus shifted to a data-driven project aimed at predicting the speed of drill bit using historical data. During this project, I have learned a lot about time-series clustering and prediction methods. Besides, through the experiences gained in this project, I came to recognize that real-world scenarios encompass far more complicated factors that can be addressed by simply applying theoretical models with ideal assumptions. Furthermore, this project held notable significance for me personally, as it represented my first formal research endeavor. It afforded me the opportunity to navigate the entire research process, from formulating the problem and crafting hypotheses, to exploring potential solutions and evaluating performance metrics.

From May 2019 to November 2019, I completed the six-month mandatory internship at a Hedge Funds start-up company in Paris. The internship project centered on applying statistical learning methods to analyze historical trading data and develop new trading strategies.

In the summer of 2020, I conducted a 3-month internship at a Chinese AI company dedicated to developing AI solutions for industrial applications. In this role, my internship projects focused on applying generative models, such as GANs, to forecast sequences of precipitation nowcasting radar maps.

Through my various industrial experiences, I have distilled several observations, which I summarize as follows. First, industrial companies have strong requirements for data analysis to improve their production efficiency. Second, integrating machine learning approaches into industrial problem-solving requires profound domain expertise. Collaboration between domain experts and data scientists is crucial for success. Third, numerous factors, such as the data bias, the specific industrial evaluation criteria, computational efficiency in terms of time and memory, and the reliability of predictions, should be carefully considered during the implementation of machine learning and deep learning methods.

PhD studies

From November 2020 to July 2024, I pursued the PhD degree in Inria Grenoble, under the supervision of Dr. Xavier Alameda-Pineda and Professor Laurent Girin. During my PhD studies, my research generally focused on applying deep probabilistic generative models for solving intricate scientific and engineering problems under un/semi-/weakly supervised configurations. Particularly, I applied a specific type of deep latent variable model designed for sequential data, referred to as dynamical variational auto-encoders (DVAEs), to tackle various audio and visual tasks, including multi-object tracking, single-channel audio source separation, and speech enhancement. The proposed approach involves initially pre-training a DVAE model with natural or synthetic signals to embed prior knowledge about the complex data patterns. Subsequently, this pre-trained model is integrated into an extended latent variable generative model (LVGM) to address the specific practical problem by applying the variational inference methodology. The proposed solutions are data-efficient and interpretable. For more details, please refer to my PhD Thesis as well as the presentation.

Both of my supervisors, Xavi and Laurent, are rigorous and respectable researchers. Their guidance and expertise have made the challenging journey of pursuing a PhD enjoyable. I have learned a lot from them, not only in terms of scientific knowledge and research methodologies but also in terms of their attitudes towards science and their invaluable personalities.

Beyond academia and science

Thanks for your strong interests about me that has led you to read until here 🤗. Now I am going to talk about the part of me that is ni scientific nor academic. The objective is to draw a more complete portrait of myself. As an individual, I generally possess a strong curiosity and deep concerns in a variety of subjects, including philosophy (both occidental and oriental), natural science, economics, history, psychology, and more. I love reading, thinking deeply, talking with people and trying to establish meaningful connections with others. Recently, I have discovered a useful tool, which is called MBTI test, that helped me a lot to understand the personalities of people around me. According to my test results, I am categorized as INFJ-A(Introverted, Intuitive, Feeling, Judging and Assertive).

Further, I enjoy a lot my life. I have a strong passion for sports🏃‍♀️ – running, cardio fitness, swimming, skiing and hiking (in the gorgeous moutains of Alpes). Engaging in these activities makes me feel truly alive. They provide me with a sense of primitive vitality that invigorates my spirit. Besides, I am into yoga and meditation 🧘‍♀️ to keep the balance of my life. Cooking is another thing that I enjoy a lot 🍜. Food has always been a source of cure for me in life. Preparing meals myself brings me back to reality and allows me to fully immerse myself in the present moment. When it comes to art, I take great pleasure in painting and I am practicing aquarelle 🎨. Music 🎵 also holds an indispensable place in my life. Arts allow me to experience the beauty and joy of life in a profound and meaningful way.