UGB Chapter 36 Defense
Chapter 36 Defense
In May, the air at Huaxia University was filled with a restless atmosphere.
Graduation season brings a flurry of ceremonies—thesis defenses, graduation photos, farewell dinners, and job fairs. Students in graduation gowns are seen everywhere in the corridors taking selfies with selfie sticks, the library lobby is plastered with recruitment posters from various companies, and the corners of the cafeteria echo with the sounds of drunken laughter and sobbing.
Zuo Cheng had no time to experience these things.
Since the signing of the contract on April 10th, the Sky Dome project has been running for 35 days. Chen Hao's signal receiving module is 80% complete, Fang Ze's embedded platform framework is built, and Zhang Lei has written more than 200 automated test cases. Zuo Cheng's own channel estimation module's core algorithm has been running successfully on the PC and is being ported to the embedded platform.
The progress slightly exceeded expectations. The three-level memory architecture design saved a lot of debugging time in actual development—Fang Ze said that this architecture "is like the experience summarized by someone who has fallen into a pit." Zuo Cheng smiled and didn't say anything.
But the thesis defense can't be delayed any longer.
On the evening of May 10th, Yu Ying sent Zuo Cheng the third revised version of her defense presentation slides.
"Look, the main change is in the data presentation of the fourth section. The tables have been replaced with bar charts for comparison, which makes it more visually impactful."
Zuo Cheng opened the file and flipped through it.
Yu Ying's style of creating PPTs differs from Shen Yue's. Shen Yue pursues a balance between visual tension and information density, while Yu Ying pursues a seamless logical connection—the transition between each slide and the next is meticulously designed, allowing the judges' attention to slide down her train of thought like a slide to the end.
"The formula derivation on page seven could be simplified further," Zuo Cheng said. "Two professors on the defense committee aren't from the communications field; they'll get distracted by overly long derivations."
"I also want to simplify things, but this derivation step is a crucial turning point in the entire plan. Skipping it would make the conclusion seem too abrupt to the review committee." Yu Ying thought for a moment, "Let's compromise—keep the derivation, but I'll add a simple analogy explanation next to it. Those who don't understand can look at the analogy, and those who do can look at the formula."
"OK."
Yu Ying spent a few minutes revising it on the computer and sent it back. Zuo Cheng glanced at the analogy—"This is equivalent to first using a weather forecast to draw a rough route map, and then using real-time navigation to correct every turn"—concise, accurate, and without any technical jargon.
"You know better than I do how to talk to non-professionals," Zuo Cheng said.
"Because I've helped Professor Lin revise quite a few undergraduate theses," Yu Ying laughed. "I've seen all sorts of outrageous statements; I've developed an immunity to them."
The two sat side by side in the library, each in front of a laptop, revising a PowerPoint presentation until 11 p.m. The library was about to close, and the librarian coughed twice in the distance.
Yu Ying closed her laptop and stretched.
"That's almost done. Go back and read it through again. If there are any problems, revise them tomorrow."
"Okay." Zuo Cheng closed his laptop. "Thank you."
"Don't mention it." Yu Ying slung her backpack over her shoulder, took a couple of steps, then stopped, turned around, and said, "Zuo Cheng."
"Um?"
"I attended the defense as an observer."
You don't need to—
"I want to go," Yu Ying interrupted him, her tone soft but firm. "It's your first formal defense, and I'd like to sit down there."
Zuo Cheng looked at her profile. The lights in the library corridor softened her features, and the ends of her ponytail draped over her shoulders, slightly curled.
"Okay," he said.
May 15th, 2 PM, Conference Room on the Sixth Floor of the School of Information Science and Technology, Huaxia University.
The defense committee consisted of five professors, chaired by Professor Qian Zhengping, the academic leader in signal processing, with the other four from the fields of communication engineering, electronic information, computer science, and automation, respectively. Lin Zhiyuan attended as the supervisor but did not participate in the scoring.
There were more than a dozen people in the audience—most of them were junior graduate students who came to observe and learn, and a few undergraduates who were scheduled to defend their dissertations later today.
Yu Ying sat in the last row of the audience seats, by the window.
Zuo Cheng stood in front of the projection screen, glanced at the audience, and his gaze lingered on Yu Ying for less than a second.
"Hello everyone, I am Zuo Cheng, a 2013 undergraduate student majoring in Information Engineering. My graduation thesis is titled 'Research on a Two-Layer Channel Prediction Architecture for Low-Earth Orbit Satellite Communication,' and my advisor is Professor Lin Zhiyuan."
He spoke at a moderate pace during the twenty-minute report.
Every slide of the PPT was meticulously crafted by Yu Ying, with seamless logical connections. Instead of reading it aloud, he explained each technical point clearly in his own words—elaborating where necessary and concisely mentioning where appropriate.
During the simulation demonstration, he brought up the real-time simulation program that Fang Ze had helped build—the same one used during the Tianqiong review, but it had been trimmed according to the paper's content, retaining only the core two-layer architecture demonstration.
The image of two curves almost completely overlapping on the screen made Professor Qian Zhengping raise an eyebrow.
Q&A session.
Qian Zhengping was the first to speak: "In your two-layer architecture, how is the synchronization mechanism designed between the deterministic prediction at the bottom layer and the stochastic compensation at the top layer? What happens if there is a timing discrepancy between the two layers?"
Zuo Cheng had been asked a similar question by the chief engineer during the Sky Dome review. He answered fluently, explaining the design principles of the synchronization mechanism and the three-level degradation scheme for exception handling, with corresponding simulation verification data for each level.
The computer science professor asked a second question: "In your algorithm complexity analysis, what is the worst-case computation latency? Does it meet real-time requirements?"
"In the worst-case scenario, the single-frame processing latency will not exceed two milliseconds," Zuo Cheng replied. "This data was measured on the ARM Cortex-A72 platform—" He paused, realizing he almost revealed the embedded platform information for the Sky Dome project, "—it was measured on an embedded development board with equivalent performance, meeting the millisecond-level response requirements for antenna beam management."
The professor nodded.
A professor specializing in automation asked an interdisciplinary question: "What are the similarities and differences between your framework of 'layering deterministic and stochastic factors' and the 'feedforward plus feedback' architecture in control theory?"
Zuo Cheng's eyes lit up.
"Essentially, they are expressions of the same idea in different fields," he said. "In control theory, feedforward is predictive control based on a system model, and feedback is real-time correction based on error signals—this corresponds one-to-one with my two-layer architecture. The difference is that the 'controlled object' in channel prediction is a time-varying stochastic process, which is much more complex than the dynamic characteristics of traditional control systems. Therefore, the upper-layer 'feedback' cannot use simple PID; it requires adaptive algorithms to track changes in statistical characteristics."
This answer exceeded the automation professor's expectations. He flipped through the paper, glanced at Zuo Cheng, and wrote something on the grading sheet.
The last question came from a professor specializing in communications, asking about the differences between the paper's innovative points and existing literature. This was a standard question, and Zuo Cheng finished his explanation in three minutes.
The dissertation committee deliberated for less than ten minutes.
Professor Qian Zhengping announced the results: "After unanimous evaluation by the defense committee, Zuo Cheng's graduation thesis is rated as excellent, and the defense is passed. The committee also recommends—"
He paused for a moment and glanced at Lin Zhiyuan.
"—The research depth and innovation of this thesis have reached the level of a master's thesis, and it is recommended that the college give it priority support when recommending it as an outstanding graduation thesis at the provincial level."
Someone in the audience took a soft breath.
Undergraduate theses being rated as "at the level of master's theses" is not unheard of in the history of the School of Information Science and Technology at Huaxia University, but it is extremely rare.
Zuo Cheng stood on the stage and bowed slightly.
Thank you, teachers.
As he stepped down from the podium, he caught a glimpse of Yu Ying clapping in the audience. Her clapping wasn't particularly forceful, but her smile was genuine.
After the defense, the corridor was crowded and noisy. Several lower-year students surrounded him, asking questions, and Lin Zhiyuan was also pulled aside to talk to other professors. Zuo Cheng finally managed to escape and saw someone he didn't expect to see in the stairwell.
Zhao Kai.
He leaned against the stair railing, wearing a slightly faded shirt, holding what looked like a resume in his hand. Seeing Zuo Cheng approach, he straightened up and called out.
"Zuocheng."
"Zhao Kai." Zuo Cheng stopped in his tracks. "Are you defending your thesis today too?"
"I finished answering this morning. I passed." Zhao Kai's tone was flat, showing no joy. He hesitated for a few seconds, then took out his phone from his pocket. "I have something I want to ask you."
"explain."
"I received two offers. One is from an IoT startup in Huaxia City; the compensation is good, but the company only has about twenty people, and I don't know how long it will last. The other is a research and development position at a state-owned enterprise in the provincial capital; it's stable, but the salary is average." Zhao Kai looked at Zuo Cheng. "Which one do you think I should take?"
Zuo Cheng glanced at him.
Six months ago, this person was still in the research group, working with Ma Hao to sabotage him—although Zhao Kai's involvement was far less than Ma Hao's, and he mostly just went with the flow. Later, Zhao Kai lost his eligibility for postgraduate recommendation, and his attitude changed. Before the winter break, he even asked Zuo Cheng 402 if they were recruiting.
"What do you want?" Zuo Cheng asked in return.
Zhao Kai was stunned for a moment.
"Stability is important, but I don't want it to be too boring."
"Then go to a startup," Zuo Cheng said directly. "You're twenty-two, the time when the cost of making mistakes is lowest. The stability of state-owned enterprises is an advantage now, but in ten years it will be a shackle. If a startup fails, you can find another job with your experience; you won't lose out."
Zhao Kai remained silent for a few seconds.
"Zuo Cheng, you never say anything ambiguous." He gave a wry smile. "Sometimes I really envy you; you see everything so clearly."
"It's not that I could see clearly," Zuo Cheng shook his head, "it's that I fell."
He said it very softly, and Zhao Kai didn't press him for details.
"Thank you." Zhao Kai put his phone away. "That IoT company is called Ruilian Technology; they make smart sensors. If there's an opportunity for cooperation in the future—"
"Keep in touch."
Zhao Kai nodded, turned around and left.
Zuo Cheng watched his retreating figure disappear around the corner of the stairs, thinking of Zhao Kai from his previous life—who had worked for a state-owned enterprise for ten years with no hope of promotion, and started to drift through life in his early forties. Perhaps things would be different if he chose the right path in this life.
He looked away and walked out of the School of Information Science building.
Yu Ying waited for him in the shade of the sycamore tree downstairs.
"Congratulations." She handed her a bottle of mineral water. "'Reaching the level of a master's thesis'—Professor Qian doesn't say things like that lightly."
Zuo Cheng took the water, unscrewed the cap, and took a sip.
"Your PowerPoint presentation was a huge help."
"The PPT is just a formality; the content is yours," Yu Ying said with a smile, shaking her head. "Let's go, Professor Lin said he'd treat you to dinner tonight to celebrate passing your defense."
"Just me?"
"And me too." Yu Ying's tone was natural. "Professor Lin said I could participate as a co-author of the thesis."
The two walked side by side along the sycamore-lined path. The May sunlight filtered through the leaves, casting dappled shadows on the ground.
Zuo Cheng suddenly remembered something.
"Yu Ying."
"Um?"
"You said before that we've known each other for almost eight months."
Yu Ying paused for a moment and turned her head to look at him.
"I did the math, and it's exactly eight months and three days today." Zuo Cheng looked at the road ahead, his voice calm. "Once the first milestone of the Sky Dome project is completed, there will be about twenty days left—it will be roughly eight months and twenty-three days then."
Yu Ying didn't speak, but she slowed her pace.
"On the eighth month and twenty-threeth day, there's something I want to tell you."
Yu Ying lowered her head, her ponytail covering half of her face.
"...What is it?"
You'll find out then.
Yu Ying was silent for three seconds, then quickened her pace and walked in front of him, her voice carrying a hint of barely perceptible laughter.
"Then you'd better not go back on your word."