【推荐1】Developing new materials requires significant time and labor, but some chemists are now hopeful that AI could one day shoulder much of this burden. In a new study in the Journal of the American Chemical Society, a team used a popular AI model, ChatGPT, to perform one particularly time-consuming task: searching scientific literature. With that data, they built a second tool, a new model to predict experimental results.

Reports from previous studies offer a vast amount of information that chemists need, but finding and analyzing the most relevant details can be laborious. For example, those interested in designing highly porous, crystalline (晶体) metal-organic frameworks (MOFs) — which have potential applications in areas such as clean energy—must sort through hundreds of scientific papers describing a variety of experimental conditions. Researchers have previously attempted to let AI take over this task; however, the language processing models they used required significant technical expertise , and applying them to new topics meant changing the program. Omar Yaghi and colleagues wanted to see if the next generation of language models, which includes ChatGPT, could offer a more accessible, flexible way to extract information.

To analyze text from scientific papers, the team gave ChatGPT prompts (提示符) it through three processes intended to identify and summarize the experimental information the papers contained. The researchers carefully constructed these prompts to minimize the model’s tendency to make up responses, and to ensure the best responses possible.

When tested on 228 papers describing MOF syntheses (合成), this system drew more than 26, 000 factors relevant for making roughly 800 of these mixtures. With these data, the team trained a separate AI model to predict the crystalline state of MOFs based on these conditions. And finally, to make the data more user-friendly, they built a chatbot to answer questions about it. The team notes that, unlike previous AI-based efforts, this one does not require expertise in coding. What’s more, scientists can shift its focus simply by adjusting the narrative language in the prompts. This new system, which they call the “ChatGPT Chemistry Assistant,” could also be useful in other fields of chemistry.

【推荐2】In the far north, well above the Arctic Circle in Alaska, ground squirrels (地松鼠) are like little balls within a deep hole in the ground. If you look at one, you might think it is dead. The squirrel is as cold as ice. Its body temperature is –2℃. Its heart beats only once every 15 seconds. Its breathing stops for minutes at a time.

It’s not exanimate, of course — just hibernating (冬眠). But spring is on its way to Alaska. As the days are becoming longer and the ground becomes warm, the Arctic ground squirrels will be warm, too. At first, the increase will be almost too small to notice. “You see them begin to breathe a little more quickly — see their heart rate speed up,” says Brian Barnes, a zoologist of the University of Alaska. “As they get up to 10 to 20℃, you see them shivering (颤抖) quite clearly,” he notes, “just as we shiver.” This shivering is a type of way to create heat. “Once their body temperatures rise above 30℃,” he says, “they wake up, clean themselves, and move.” A squirrel that looked dead a few hours before is now very much alive.

Arctic ground squirrels are among the world’s coolest hibernators. Chilling out (放松) for months at a time lets them grow in this extremely cold place, where food is short. By studying how ground squirrels hibernate, scientists hope to answer some big questions. Among them: How do these animals go from warm to cold and back again? And might people ever do the same? The ability to chill out could help humans who suffer from brain injuries and heart problems survive.

【推荐2】I’d like to talk about my first teaching experience. It was in the fall of 2010—I had to teach integral calculus (积分学). I had taught before but it was always 2-3 students at a time. I had no experience of classroom teaching. So I had butterflies in my stomach.

Anyway, the appointed hour came and I had no choice but to go on. I introduced myself, asked each one of them to do a brief introduction and got down to business. Fortunately, it was a worksheet session, so I didn’t have to do much. The students were well prepared—most of them had done AP calculus. The first day was a success.

As the quarter went by, I found the work more and more easygoing. So I was lax and stopped preparing the homework problems beforehand. How wrong I was!

One day we were doing surfaces of revolution. I used to do them in a more different way than it was taught in the text. I tried to do the first problem but it wasn’t just a piece of cake—I had to step back and think for five minutes before the solution came to me. To the credit of my students, no one showed any sign of impatience in the meantime. I was feeling doubly uncomfortable because it was a day of observation by the school leaders.

The next class I tried to give some additional resources on advanced materials, especially to students who would stay after the class or come to my office hours.

My evaluations were mixed. Two major complaints were about my accent and my handwriting on the board.

I learned a lot about teaching after this course. In my view, teaching is like a performing art. No amount of reading or attending workshops will prepare you for the challenge. You only get better with practice.