Chapter 17: A Pause in the Journey

Behind him, Qiu Yan and the others watched Zhou Yun shake his head, feeling a bit helpless. After three months of working together, they had grown familiar with Zhou Yun and had a clear understanding of his abilities.

If they were to sum him up in three words, Zhou Yun was like a wish-granting machine.

For example, whenever they encountered problems in their research and asked Zhou Yun for help, their conversations typically went like this:

“Zhou Yun, something’s off with the experimental results on this.”

“Hmm, let me take a look.”

By the next day, Zhou Yun would have a solution ready.

It was almost unbelievable.

Often, they wondered why someone like Zhou Yun would choose to pursue graduate studies at their school and join their group. Had the principal’s ancestors struck gold?

In their eyes, a student like him could easily have secured guaranteed admission at prestigious places like Yanda or Shuimu.

Yet, if Zhou Yun were to leave now, they would be a thousand times reluctant to see him go.

It was now 11:30 a.m., the market had closed for the morning session.

Watching the screen freeze, Zhou Yun stopped overthinking. No matter how many ideas were swirling in his mind, none were achievable at this moment. It was better to take things step by step.

He took a deep breath and stood up, prompting the others to instinctively step back.

“Thank you, senior brothers and sisters, for your hard work during this period.” The workload had indeed been heavy, mainly in data collection. Besides Zhou Yun, the lab had only four people.

They had gathered all the data from this gold ETF since its inception, along with data for all the underlying stocks. Just thinking about the sheer volume was overwhelming.

“No trouble at all, it’s nothing.”

“Yeah, this is less work than writing a paper.”

Zhou Yun shook his head. “Regardless, without your help, the progress wouldn’t have been this fast.”

“Hey, we’re your seniors; no need to be so polite. Since we’re almost done, how about we all go out for a meal? Everyone has to come this time!” Qiu Yan glanced around.

“Alright.”

“No problem.” Yu Tianrui and Wang Jiayue didn’t refuse this time. Both were introverted; the last time they’d met Zhou Yun was their first encounter, and dining together was still a bit daunting for them.

But now, having grown more familiar, sharing a meal was no big deal.

“Then it’s settled. I’ll book a place. Let’s get ready and head out soon.”

“What place?” Deng Yonghua appeared at the lab door, smiling at everyone.

“Well...” Qiu Yan briefly explained the situation to Deng Yonghua.

Hearing that Zhou Yun had made new research progress, his smile broadened. “Go ahead and eat. Remember to get receipts; the group will reimburse you. After the meal, Zhou Yun, come see me—I have something to discuss.”

“Thank you, Professor!”

Deng Yonghua said no more and left immediately.

After the meal, Zhou Yun went to Deng Yonghua’s office.

“Zhou Yun, is your stock model research nearly finished?”

“Yes, I just wrapped it up this morning, so I wanted to celebrate with a meal.”

“Congratulations. Where are you planning to submit the paper? NeurIPS again? But their deadline’s passed this year. If you want to submit, it’ll have to be for 2025.”

“I’m aiming for ICLR 2025. The deadline is three months away. The model isn’t quite ideal yet, so I want to see if I can optimize it further. But even as it stands, I think it should be fine for ICLR.”

Zhou Yun’s confidence was noticeably higher than before.

He wasn’t boasting. Although his model couldn’t compete with those from large quantitative firms, within the realm of time series prediction, it was, without exaggeration, unparalleled.

Achieving this level was largely thanks to mathematics.

Had Zhou Yun not reinforced his math foundation, many of the architectural and formula optimizations would have eluded him.

Seeing Zhou Yun’s confident demeanor, Deng Yonghua was very pleased. This was the right spirit!

“Alright, since you’re confident, I won’t say much. I’ll review it again before submission. With last time’s experience, I don’t expect many revisions. Next time, you might not even need me to look at it.”

The smile on Deng Yonghua’s face never faded; having such an outstanding student made it hard to suppress his happiness.

“Professor, did you want to see me about something?”

Deng Yonghua slapped his forehead. “Almost forgot the main point. Last month, when you mentioned a new project, I began processing it. But since you haven’t officially enrolled yet, the procedures were a bit complicated. The school does offer support for research conducted by students like you.”

He then pulled out a form. “This is the school’s research funding support—top-tier, twenty thousand yuan. Fill it out using the paper for your NeurIPS submission.

“There are no other restrictions. It’s just to encourage student research. Don’t worry, just fill it in. If your paper gets accepted, the school will reward you about fifty thousand more, totaling around seventy thousand.”

Zhou Yun glanced at the form and, as Deng Yonghua said, took a pen and began filling it out.

While Zhou Yun filled the form, Deng Yonghua sighed. “Our school’s funding is still too little. For top-tier results like NeurIPS papers, other schools would offer at least a hundred thousand yuan in rewards, not to mention other research funds and scholarships.”

The disparity between institutions was already apparent in monetary awards.

Zhou Yun was prepared for this. “It’s alright, Professor. If the school’s research funds are limited, there’s always national funding to apply for.”

Deng Yonghua was momentarily surprised, then smiled. “I underestimated you. Don’t worry, if you really have a good project, I’ll help you with the application. As long as your topic holds up, approval will be no problem!”

It seemed that a NeurIPS paper had significantly boosted Zhou Yun’s confidence.

“Professor, I’ve finished filling it out.”

Deng Yonghua took the form, examining it. “Looks good. Once your paper is accepted, you’ll receive this money. Let’s see—should be soon, right? It’s already July. Usually, feedback arrives around this time. Keep an eye on your email these days.”

“I have been.”

...

United States, Massachusetts Institute of Technology.

In his office, He Kaiming, about to finish his work for the day, glanced at OpenReview—the NeurIPS review system.

He noticed that the reviews for the earlier paper titled “AgileEdge: Adaptive Co-Optimization for Pervasive Low-Latency Edge AI” had been completed.

The reviewers had given high scores, averaging 7.8 out of 10.

Moreover, all four reviewers shared a common comment: “The proposed method shows clear superiority over baselines in both theoretical guarantees and empirical results.”