December 2024
At the start of the month, before the visa system was updated, I made an appointment for a face-to-face visa interview in Shanghai for next year. The visa fee was over a thousand yuan, so expensive! And if you don't pass, you have to pay it all over again when you reschedule, isn't that just daylight robbery?! The interview requires submitting an invitation letter and proof of funding from the school, but the SP conference hasn't yet opened applications for participation, and the email I sent vanished without a trace.
Don't waste time on meaningless, endless tasks. In all things, seek moderation.
20241223-20241229 Week 94
- Took on a software system competition task, working hard to push it forward
20241216-20241222 Week 93
- Put together a working demo, though there are still many areas that can be improved
- Need to look into the approaches of related work
20241209-20241215 Week 92
- Proctored the C programming lab exam
- Pushed forward code progress
20241202-20241208 Week 91
- Organized the PPT for C programming review
- Advanced my study of LLVM and updated technical notes
November 2024
At the beginning of the month I gathered the kids' various opinions on C programming. The most common complaint was that the boss's roll-call during class was nerve-wracking. As for the workload, most felt it was moderate, and there were some good suggestions, like setting up timed but ungraded mock tests, providing more practice problems, using in-class material as practice questions, and so on. These can serve as further optimization for future work. The third stage exam scores were quite unsatisfactory, mainly because of weak programming ability. Even though I doubled their exam time, fewer than 30 students got everything correct in the end. I heard that other teachers' quizzes were also quite difficult, so there's still room to argue.
Extra self-improvement is still very important. Every day I should organize and learn one or two new knowledge points and post them to tech-blog.
Looking back at this month, it feels like a lot of time was spent dealing with miscellaneous chores, and not much progress was made on my own research topic. When a research topic is finished and you look back, you realize it might have been something you could have wrapped up in two or three months. Why does it take so long in practice? Maybe this is the reason.
20241125-20241201 Week 90
- Pushed forward my own research topic
- Starting this week, I summarize and learn a new tech knowledge point every day
20241118-20241124 Week 89
- Prepared the social practice PPT for the defense presentation
- Updated the sec website with new 2025 data
- Optimized the theme style of the tech website
20241111-20241117 Week 88
- Studied Linux clang compilation
- Sent the lab printer in for repair
- Summarized an LLM survey
20241104-20241110 Week 87
- Bought biscuits for the C programming kids
- Manually organized and analyzed FN issues from the experiment
- Organized PCC
20241028-20241103 Week 86
- Participated in the party branch's team-building event in Liangzhu
- Organized the lab PPT
October 2024 Summary
At the beginning of the month, I bought some calligraphy practice books, so I could write a few characters whenever I felt restless. Mid-month, I subscribed to a year's worth of Sanlian Lifeweek. I had heard about it long ago, and this month I finally subscribed on impulse. At the end of the month, a good friend who went abroad to study after undergraduate graduation came back to China. The trip back to Hangzhou might be a six-day business trip, so we could get together for a meal.
20241021-20241027 Week 85
- Hosted my undergraduate friend visiting Hangzhou
- Attended an Ant discussion meeting
- Organized the lab PPT
20241014-20241020 Week 84
- Pushed forward the LLM Survey
- Pushed forward the lab PPT
20241007-20241013 Week 83
- Looks like I'll be continuing to push the lab PPT progress this whole month
- Handled the post-summer-social-practice project closure application
- Completed the first stage exam for C programming
- Prepared for the weekend group meeting
20240930-20241006 Week 82
- Pushing forward the lab PPT progress this week
September 2024 Summary
At the beginning of the month I participated in Shanghai Jiao Tong University's justhunt, where truly one problem equals the workload of a Tier-1 paper. From this month on, I became my boss's TA for C programming! I had wanted to be a TA three or four years ago, hh. This semester I'm still a part-time counselor for the school as in past years. Right at the start of the semester, my schedule was packed back to back. Finally finished justhunt on 9.11! (carried by my teammates the whole way)
In mid-month, the boss assigned me to organize the lab work report content. It was indeed an arduous task! Suddenly the burden on my shoulders multiplied, and I felt a bit anxious. Sometimes, because I haven't sorted out exactly how much work there is, anxiety about the unknown is triggered. I still need to list out the things to be done and divide them by the four-quadrant matrix of urgency and importance, so I don't lose my footing.
20240923-20240929 Week 81
- Pushed forward the lab results PPT and the survey progress this week
20240916-20240922 Week 80
- Handled C programming TA matters this week
- Took on a new responsibility, organizing the lab results PPT
20240909-20240915 Week 79
- Oakland paper accepted! A bit of luck. Now I have to prepare the camera-ready as soon as possible
- As a C programming TA, I organized and updated six or seven PPTs. Fortunately there's material from past years to reference, so it's still manageable
- Time to keep pushing forward my own work
20240902-20240908 Week 78
- Officially back in school! This year I'm enrolled as a PhD student, attending all kinds of opening ceremonies for the school and the college
- Participated in the justhunt competition, and was carried all the way by my teammates
August 2024 Summary
I've found that sometimes it's actually better to take notes and record thoughts on paper (compared to typing), because the slower pace leaves time for the mind to actually think things through. Plus you're not bound by the form of text — you can easily draw diagrams. And you can also put many previously unused notebooks to good use.
In the latter part of the month, I went hard on rebuttals — it was truly torturous. Although the comments this time mostly had a clear line of response and the experiments to add weren't many, the back-and-forth was indeed mentally exhausting. Sometimes I had to think for half a day to construct a barely passable answer. Add to that my habit of reading carelessly, answering off-topic, over-explaining, stretching arguments, lacking polish, and being a self-detonating tank — all of which made the rebuttal even more agonizing. The short version was at least bearable. Note that Overleaf can also use Markdown, and rebuttals should make reasonable use of second-level headings, bold text, bullets, and so on. When answering reviewer questions individually, start each paragraph with one or two keywords summarizing what the paragraph is about.
At the end of the month I used Navicat to analyze datasets — it's really useful! You can directly view a remote server's database locally. I bought primeblocks and built a lab server management website.
20240826-20240901 Week 77
- Rewrote a lab server status query website. I really like its current style
20240819-20240825 Week 76
- Doing the rebuttal! Although the comments aren't very favorable, it's likely going to be rejected...
20240812-20240818 Week 75
- Started learning LLVM
- Participated in the CCBC15 puzzle competition. It was my first time encountering puzzle hints. Although I didn't finish the contest, I feel like I've fallen into a new pit.
July 2024 Summary
Although I said I'd diet to lose weight, I can't eat too little at each meal, can I? At the start of the month I tried to limit myself to one meat dish, bringing the cost of lunch and dinner down to around 12 or 13 yuan, whereas before it was basically 18 or 19. But if I eat too little at night, I get really hungry, and I lose the energy and willpower to keep working. Or worse, I order a midnight snack — not only another hefty expense but also wrecking the diet plan. Now that I think about it, what's the root cause of that "can't keep working" mood and feeling at night? Is there a way to avoid it...
If you're lying in bed, you should just go to sleep — don't keep replying to messages. While it doesn't cause insomnia, it does make you feel mentally less sharp the next day. Don't stay up too late at night, either, so you can get up earlier the next day.
In mid-month I went to Wuxi to start social practice — here you have to wash your own plates and bowls hhh. I'd wake up at seven in the morning, eat breakfast at the hotel, and basically be at work by eight. The everyday work was about the same — continuing to push forward my own previous research topic.
At the end of the month, I made it into the second round. Hope my luck is a little better...
20240701-20240811 Week 69 - Week 74
- Wuxi summer practice. Details omitted
June 2024 Summary
At the start of the month I revised the paper and submitted again. Hope I don't run into the same reviewers... There really wasn't time to redo the experiments, so I'll just have to wait until after submission to slowly think about whether I can revise more.
In the dorm I discovered a — cock! roach! ?! Then in the blink of an eye it was gone!! That very night I bought roach poison. After it arrived and I unwrapped the packaging, the small box held sugar pellets that gave off an extremely rich brown sugar milk tea fragrance. Forget the roaches, even I wanted to lick a few. But I don't know if it's because of the toxin evaporating — I woke up at five-something in the morning, and once I got up my nose kept bleeding.
Painstakingly revised the paper and submitted to Oakland. Let's see if I can make it into the second round.
In mid-month I suddenly noticed that comments had been enabled for my WeChat public account. After looking into it, I found out it's a recent grayscale test. Yay! And LLMs are exactly all the rage right now, so from now on I can use AI to draw the cover images for my posts. I can also put learning AIGC on the agenda.
By the end of the month, I realized I hadn't actually put the public-account-related matters on the agenda orz. I bought a small white timer for a few yuan, took apart the casing, removed the buzzer's vibrating diaphragm — now the buttons won't make noise! The way LCD screens are controlled is really magical. I saw it for the first time — the whole thing is just a piece of glass, then on the side it connects to the circuit board through a layer of conductive foam, and then it can display. When I have time I should look up the principles.
20240624-20240630 Week 68
- Inherited Brother Chong's machine! Thanks Brother Chong!
- For recent work, the main goals are pushing forward my own topic and writing a survey
20240617-20240623 Week 67
- On top of automated code analysis, doing manual analysis, collecting datasets, and analyzing data characteristics
- Quickly pushed forward the completion of a patent
- Started preparing for a math exam, cramming half a semester of material in three days...
20240610-20240616 Week 66
- Orderly progress on the new research topic, mainly implementing automated code analysis
20240603-20240609 Week 65
- This week I painstakingly revised the paper and submitted to SP.
- Slacked off Saturday, went to Lin'an district on Sunday for a stroll — gave myself two days off.
May 2024 Summary
After submitting my paper, when I look at other articles now, it feels like there aren't so many "barriers" anymore — there's a sense of parting the heavy clouds, and subconsciously I find myself actively learning from their style. So you really do have to go through these things in order to grow. I feel like the things I can do every day have suddenly been "unlocked." Fully tapping into my own initiative, my working hours in the lab now stretch from 9 a.m. to 10 p.m. After clocking out, I do some non-research personal-quality enrichment work in the dorm — pretty nice. With this arrangement, I shouldn't get distracted with other things while in the lab — I should focus on research tasks!
At the end of the month, the paper was rejected hhhh. Better hurry up and revise it for the next ddl.
20240527-20240602 Week 64
- The paper was rejected! Hurry up and revise it, transfer to the next conference.
- This week I tinkered with the lab network and optimized the printer's wired connection.
20240520-20240526 Week 63
- My own server died! The hard drive was just too crappy. Luckily my code was all backed up and the website is hosted on GitHub, otherwise it would all be gone.
- This week I was busy hurrying various PPTs along, and didn't push my own research topic forward much. Need to make haste.
20240513-20240519 Week 62
- My next research topic continues to advance. Let's start by collecting a dataset.
- Preparing my own coursework, taking over lab website maintenance.
20240506-20240512 Week 61
- Off to the next mountain and sea! Reading related work on LLM4APR.
- Wrote some Zhihu articles.
20240429-20240505 Week 60
- This week I submitted the paper and held a discussion meeting. Need to push various tasks forward urgently.
- Although there was no May Day vacation, things didn't seem to advance much. Need to put in more effort.
April 2024 Summary
At the beginning of the month I put together a draft of the paper that could be sent to the teachers for review. With less than a month left, I had to work hard to catch up. Recently I've been reading books on cognition, learning that making changes is the true meaning of reading — and indeed, my lifestyle habits have changed considerably. For example, I cancelled my alarm clock and rely on the 90-minute sleep cycle to wake up naturally. Through reading I've learned that new knowledge needs to be connected with prior experience and knowledge, and a book is read enough as long as it touches you somewhere. Now that there's a piano and a tripod in the dorm, I want to upload daily piano practice clips to the website. Just need to think about what kind of presentation form is most appropriate.
In mid-month I suddenly realized that the best moment to learn new things might be in the morning — every morning when I get up there's a strong interest in learning, but I always miss the chance to learn because I don't follow through with action. Another way to learn — could it be in the form of writing exam questions? People say that learning efficiency, from low to high, runs through listening, taking notes, and teaching, but writing test questions doesn't seem to be on that list. I wonder how much it could improve learning efficiency. Also, constantly remind yourself what's most important and what's most urgent right now, and put your golden hours of focused attention on what matters most.
In the latter part of the month I rushed paper progress like crazy, syncing with the senior brother almost every day. The workload of those few days might have exceeded the past several months hhh.
Finally, five days before the ddl, I finished the confirmed draft for the teachers to review. A huge weight lifted (for now), and I went hard on sleep to make up for the lost rest. After all, this submission has been dragging on since the end of last year, so the final submission wasn't too rushed. May Day shifted holidays are coming up, so I can take advantage of the no-class window to push forward writing the patent and project applications.
20240422-20240428 Week 59
- This week I basically completed the paper draft
20240415-20240421 Week 58
- This week I madly pushed paper progress! What I did poorly was not reading much new literature
20240408-20240414 Week 57
- This week I continued to be busy revising based on each teacher's feedback and running experiments.
- Something to watch: every day I should keep reading papers — don't get lazy about learning new knowledge just because of revising and running experiments.
20240401-20240407 Week 56
- This week I finished polishing the paper draft with GPT! Sent a version to the teachers to review first
- A rare small reunion with two undergraduate friends
March 2024 Summary
At the start of the month I was busy with code open-sourcing for the previous SyzTrust work, the gossip submission, and the work related to binbin's and Brother Hy's project. At the same time I filled out my own paper structure to be complete, but the length and content still have considerable issues to optimize. I don't plan to spend too much time on it, though. The next work, considering combining LLM with RCA, can begin gathering related materials.
In mid-month, binbin emphasized that I should still prioritize finishing my own work, and that I must produce a complete version before next week. At the same time, I have a meeting with a senior brother the week after, so I need to prepare the Motivation properly.
Since moving dorms, I've felt like "my connection to the world has been cut off," especially in public settings — when talking with others I'm always in a daze, not knowing what to say, not knowing what I'm doing. There's a sense of consciousness detaching from the body. I don't know what's happened. Maybe I need to stay away from social media, settle down and read some books, and gather my scattered spirit.
In the latter part of the month, the remote server was reassigned a new IP and firewall, so I can reconnect to the server again! Experiment code and so on can keep moving forward.
Near the end of the month, I finally got the paper draft done. What's left is using ChatGPT to help polish the wording, then turning around to revise the experiment code.
20240325-20240331 Week 55
- This week I continued polishing the paper
- Updated the Tuzhu website
20240318-20240324 Week 54
- This week I condensed tables, supplemented one and a half pages of content, and polished wording
20240311-20240317 Week 53
- This week I returned to the previous paper. The first version is written, but still needs major additions and revisions
20240304-20240310 Week 52
- This week I read related papers for the next work, preparing to start the survey
- I also wrote a gossip submission for the previous SyzTrust work 《SyzTrust：可信执行环境——可信？安全？》(SyzTrust: Trusted Execution Environment — Trusted? Secure?)
February 2024 Summary
At the start of February I had originally planned to rush for the 2.9 USENIX ddl, but in the end I found that the paper quality was still far short, so I didn't make it. This year's Spring Festival trip home counted as pure leisure — I didn't do any lab work at all. Instead, I learned a bit of Vue, websocket, and JS and built a "you draw I guess" web version. After returning to school on 2.18, suddenly it's late February — time to return to lab work, hoping by month's end I can reorganize the paper according to the CCS template.
At the end of the month I'm dealing with the previous SyzTrust code and PPTs, and handling my current work in spare moments. Lately I've been a bit slack on paper reading — need to pay attention and make up for it.
20240225-20240303 Week 51
- This week I handled some unfinished items from the previous work — organizing slides, open-sourcing code, submitting to public-account gossip, etc.
20240219-20240225 Week 50
- This week I returned to school for registration and check-in, switched to a new dorm, and gradually returned to research work
- For the previously unfinished paper, I plan to give a result by mid-March, and to also start the next work
20240212-20240218 Week 49
- This week I did all kinds of social activities at home, and also wrote a "you draw I guess" front-end and back-end
20240205-20240211 Week 48
- This week I went home for Spring Festival, planning to return to school on 2.18
- While I'm at home, sleep well and catch up on rest, then plan how to launch the next work
January 2024 Summary
In the first part of the month I sprinted on paper progress! Hoping to finish the first version of the paper before mid-month.
In fact, by the latter part of the month I still hadn't finished — mainly the experiment section was missing, and I needed to revise the code and rerun the experiments, and also think about how to write the prose part. Around the 20th I seemed to have caught the flu — dizzy head, weak limbs, fever of 37.9°C. After sleeping like crazy, I'd recovered from the fever by the next day. Didn't go to the hospital — just resting more should be enough. These two days I've been pushing paper progress in the dorm.
Result: on the third night my nose bled all night, and on the fourth day I had a fever again, dizzy head, headache... Truly torturous.
The remote server seemed unable to connect to GitHub, so I considered subsequently migrating the website to Gitee.
After trying it out, Gitee charges for custom domains... so unscrupulous. The fix turned out to be that GitHub's SSH git was banned — switching to HTTPS authentication did the trick.
Year-end means the lab annual meeting is coming up — going to Anji. Hoping I can finish the first draft these two days.
In the end, at the annual meeting, I still hadn't finished writing three sections of the experiments... Spent the night in the hotel rushing progress, and the next day's Tianchi tour and skiing — I had no real interest in joining either. Time to go all out on the paper. In the afternoon at the annual meeting, I checked the failed cases of system analysis one by one, and they should be enough to support writing the rest of the experiments.
20240129-20240204 Week 47
- This week I pushed paper experiments and main text progress hard
- Wednesday and Thursday I attended the lab annual meeting, with most of my free time spent in the hotel room writing the paper hh
- The paper writing really wasn't great. After getting the hang of it the first time, the methodology should fall into place.
20240122-20240128 Week 46
- Diagnosed with influenza B. This week I handled affairs in the dorm. I have to say, the dorm is much quieter than the lab.
20240115-20240121 Week 45
- This week, the first few days I went home to handle some affairs, and the latter days I continued pushing paper progress hard
- Came down with the flu and fever over the weekend — paper made almost no progress
20240108-20240114 Week 44
- This week I pushed paper progress hard, completing 50% of the length.
20240101-20240107 Week 43
- This week I worked to push paper progress according to plan, aiming to finish the first draft next week.