Twitxy Saved

Kevin Yang

Verified account This account is verified. Learn more

@yang3kc

Kaicheng Yang, PhD | Assistant professor @BingCompSci | AI & Society, + | views mine
Binghamton, NY kaichengyang.me Joined May 2019
1.5K Following 1.6K Followers 3.5K Posts

Latest posts 5

Pinned post

Kevin Yang Verified account @yang3kc
This semester, I have been teaching an AI and Society class. As the name suggests, the goal is to explore the transformative impact of AI, particularly LLMs and AI agents, on society. This is a seminar class; see the reading list link below.

1/3
5 replies 26 reposts 169 likes 10K views View on X
Kevin Yang Verified account @yang3kc
This. Also, I think it's about positive feedback. PhD journey is long and painful, so students need positive feedback (as early as possible) to keep going.

Positive feedback can come in other forms, such as attending conferences. One of my first few PhD papers was a short workshop paper. The work was not very comprehensive and had a relatively short lifespan. But it gave me the chance to attend a conference, share what I was doing, and see what others were doing. I also met many people at the conference who became my good friends.

I think advisors should try to create opportunities for their students to get such positive feedback. Students should also actively seek these opportunities whenever possible.
Yue Zhao Verified account @yzhao010 6h
为什么「延迟兑现」在 #学术界 如此珍贵?

因为大部分人没有等待的资本。比如你 12 月 15 号要申请 phd,你需要论文,你需要在项目里做出一些 progress 来让 PI 给你写出强推。而精品需要时间,你只能先做一些过得去的东西(注意此处的限定是一个正经 work 而不是纯 slop),拿到 offer 再徐徐图之。

对于已经有了 3 篇、5 篇、10 篇(所谓)top 论文的人来说,在旁边说别水,其实就是跟富家子弟说“人生有很多容错”是一个道理,何必如此着急。前一阵子的慢科研倡议为什么被大家诟病?因为这就跟发达国家要求发展中国家承担更多节能减排任务一样。

当我逐渐到了都能理解的年纪和阶段的时候,我不会一刀切的对大家说别水 paper,因为很多人解决的生存问题。但我的倡议更多只是,吃饱了以后别养成快餐的习惯,做点至少感兴趣的东西,不那么功利的东西。

换句话说,当大家有了一些论文、impact 以后,有了容错垫之后,如何选择才会决定你的高度。之前,我支持任何人先吃饱(但绝不许给大家喂屎)。
View quoted post on X
Kevin Yang Verified account @yang3kc
Reviewing some papers with strong AI-gen vibes. Calling them slop might be too harsh, but boy, they're hard to read. Sometimes I wonder if the authors have even tried to read their manuscript becuase it's obviously not readable for humans.

The best part is that I need to finish these reviews in time so my coauthors can receive theirs. 🤷
Kevin Yang Verified account @yang3kc
Last week, my wife and I welcomed our first child into this world. We did our best to prepare for his arrival, yet we still felt unprepared. For days, we barely slept and scrambled to take care of the baby. At the same time, we felt so blessed to have him in our lives.

I also want to thank my wife. She has been going through a lot during the pregnancy, delivery, and postpartum period, even though we fortunately didn't encounter any major issues.

Becoming a father and caring for a newborn has also been a profound experience for me. In the days before we went to the hospital, I was still thinking about how to burn all my AI credits and maximize my productivity. But caring for a baby is exactly the opposite. There is no efficiency, no automation, and no scalability, but it takes tons of patience, love, and care. There are no goals or objectives, either. What matters most are the days, hours, and minutes that I spend with the baby.

As AI becomes more and more capable and sets out to replace humans in all tasks, raising a child seems like the most human thing we can do. And I'm glad that I have this opportunity.
7 replies 0 reposts 35 likes 774 views View on X