Rapid Progress

Bawendi is an expert in the field of Perovskites, but the Nobel Prize he received in 2023 was not for his achievements in Perovskites, but for his innovation in the chemical preparation methods of Quantum Dots.

Quantum Dots are tiny semiconductor particles with unique optical and electronic properties, which can be applied in fields such as LEDs, Infrared Detection, and Solar Cells.

Hearing that the other party chose Perovskites, Bawendi smiled: "Light, I'm not sure about your experimental skills, because your paper didn't really show your talent in this area.

To publish a paper in the field of Perovskites, the experimental requirements are very high. You need to be able to achieve results that others cannot, or observe phenomena that others cannot, and be able to theorize them."

Bawendi continued: "Of course, if you want to publish papers in top journals, I know this is very important for Chinese students, because you need these proofs to demonstrate your abilities so that you can get better treatment when you return to China.

From this perspective, Perovskites are indeed a very good direction. Every year, there are no less than ten papers on Perovskites in Nature alone, plus Science. This is indeed a direction where it is easy to achieve results.

I'm just not sure about your talent in doing experiments."

Chen Yuanguang was brimming with confidence. This was nothing compared to the project Pangu gave him: "I think I can."

Bawendi stared into his eyes for a while, then nodded: "Okay, if you think you can, then let's get to work, young man!"

After Bawendi sent him the list of papers and relevant textbooks, he let him study on his own. This is the MIT style; the advisor assumes you are a genius.

"Has Chen been to the lab?" Bawendi asked.

An Indian-American student said: "I haven't seen him."

"Alright." Bawendi was a little puzzled.

When Chen Yuanguang emailed him again, hoping he would recommend some new papers, Bawendi replied with an email asking: "Do you have any ideas about the paper?"

"Roughly, yes. I plan to study how to use Machine Learning methods to predict the high-throughput situation of anti-solvents on Perovskite stability," Chen Yuanguang replied.

After reading it, Bawendi realized that Chen Yuanguang had still taken the path he was familiar with, the path of computational chemistry.

Machine Learning is great. In 2016, AlphaGo emerged and defeated Go nine-dan Li Shishi, which was hailed as the first year of Machine Learning.

Two years have passed since then, and everyone feels that they should combine artificial intelligence with their own major, just like twenty years ago when there was "Internet+" everywhere.

Now it's "Artificial Intelligence+", but artificial intelligence talents are hard to find. Big companies like Google, Amazon, and FB offer sky-high prices for talents in the field of artificial intelligence, and some startup companies offer even higher prices.

It is too difficult to recruit a talent who understands artificial intelligence and also understands chemistry, let alone combining the two.

Even Bawendi couldn't find such talent. Chen Yuanguang's appearance gave Bawendi some ideas, but he didn't expect the other party to adapt so quickly and find a combination point after only a month.

"Light, come to my office tomorrow, and we'll talk about the specific research direction," Bawendi said directly on the phone after seeing Chen Yuanguang's reply, no longer sending emails.

The next day, "Light, tell me roughly about your idea." Bawendi handed the coffee to Chen Yuanguang.

"My idea is that we can combine automated characterization, chemical robot synthesis technology, and Machine Learning to explore how the choice of anti-solvent affects the intrinsic stability of Perovskites.

For example, we can choose different terminals to combine, such as MAPbI3, CsPbI3, and CsPbBr3, etc., to synthesize some combinatorial libraries, and each library will have its own unique combination.

I expect that we can synthesize more than a thousand components in total, and then each library will be made twice using two different anti-solvents: Toluene and Chloroform.

After synthesis, Photoluminescence Spectroscopy Analysis will be performed automatically every 5 minutes for a period of time. How long is appropriate needs to be determined through experiments.

Finally, Non-negative Matrix Factorization will be used to plot the time and composition-dependent optoelectronic properties.

By using this workflow for each library, we can find how the choice of anti-solvent affects the intrinsic stability of Perovskites.

In fact, it may be a dynamic process." Chen Yuanguang roughly explained his idea.

This approach is very similar to the previous directed evolution of proteins, but the difference is that experiments are done first, using chemical robots to set fixed procedures, and then machine learning is used for analysis.

Bawendi nodded repeatedly as he listened, thinking to himself that Light was indeed a good prospect, no wonder Levitt sounded resentful when he called him: "Great idea, Light, I support you.

This depends on the final result. If the result is good, it can be published in Nature or Science.

If the result is not ideal, it can still be published in JACS or a Nature sub-journal. For scholars in the field of chemistry who are behind the times, they will still give face to the most popular artificial intelligence algorithms nowadays."

Chen Yuanguang pondered that the main reason was that the internet industry offered such good salaries. Those who could combine perovskite and machine learning either hadn't grown up yet, as artificial intelligence had only been popular for two years, or they wouldn't come to pursue a doctorate at all and had already gone to the industry to make big money.

It was only because this big money was nothing to him that he had the opportunity to study here.

"Professor, also, our lab doesn't have graphics cards. I need to buy the latest Nvidia graphics cards for computation," Chen Yuanguang mentioned.

Bawendi said helplessly, "It seems I'm also outdated. Buy them. Write me an application form, and after I sign it, give it to Susan and have her purchase them."

"It's okay, the result is not great, but not bad either. Publishing in Nature might be a bit difficult, but publishing a JACS paper is more than enough," Bawendi said after staring at the results for a long time.

Chen Yuanguang said, "JACS is fine. While I was working on this project, I also worked on another project. This project can definitely be published in Nature."

Bawendi was shocked, "Tell me about it."

Chen Yuanguang: "I built a cross-attribute deep learning framework on GitHub, which can mainly be used for predictive analysis of small material data.

I mainly used some existing data from the college to train this model. First, I built a large dataset, and then based on this large dataset, I built small datasets with different attributes and their models.

Through the framework refined from these models, we can directly input physical attributes as its computational and experimental dataset, and finally get the results of its other properties."

To put it simply, it was digging another layer deeper into the topic of perovskite prediction and producing a more general result.