Mufan Li's Avatar

Mufan Li

@mufan-li.bsky.social

Assistant Professor @UWaterloo Statistics and @VectorInst | Prev: Postdoc @Princeton PhD @UofTStatSci https://mufan-li.github.io/

581 Followers  |  149 Following  |  10 Posts  |  Joined: 21.11.2024  |  1.7363

Latest posts by mufan-li.bsky.social on Bluesky

Thank you!

28.11.2024 02:10 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

I think the award has its fair place, since it’s hard to predict the value and impact of a paper 10 years later. I just think we shouldn’t devalue a paper based on current popularity alone, given that’s more about sociology rather than the work itself.

28.11.2024 00:26 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

I think this rise and fall of popularity of methods says more about the community than the method itself. Honestly who knows if there’s just one clever trick needed to make GANs better than diffusion models.

We really don’t need to be so obsessed about SOTA.

27.11.2024 23:16 β€” πŸ‘ 9    πŸ” 0    πŸ’¬ 2    πŸ“Œ 0

May I be added as well?

27.11.2024 22:51 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

People should write more blog posts. I have been pleasantly surprised over and over again by how much people found them beneficial, when I just wanted to share some cool math.

24.11.2024 22:52 β€” πŸ‘ 14    πŸ” 0    πŸ’¬ 0    πŸ“Œ 1

I don’t think β€œsurprising” is a meaningful quality to evaluate. Most results are trivial once understood properly, and surprise mostly depends existing intuition being inadequate.

E.g. scaling laws are basically non-parametric rates, and you deviate if you are not minimax optimal.

24.11.2024 17:05 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Nice to know there are other metal heads here :)

24.11.2024 15:40 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Preview
Sky Follower Bridge - Chrome Web Store Instantly find and follow the same users from your Twitter follows on Bluesky.

Starter packs are helpful as well as the twitter import tool chromewebstore.google.com/detail/sky-f...

23.11.2024 20:36 β€” πŸ‘ 7    πŸ” 3    πŸ’¬ 0    πŸ“Œ 0

Perhaps if the users with higher connectivity (influencers) post more on B than X, then the majority of users who are just consuming the content will prefer B instead.

The influencers also do not have to be following an optimal strategy.

23.11.2024 23:10 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Honestly my AC batch is decent. Far better than the quality of reviews I’ve gotten for my own submissions.

I would prefer more reviewers to have a stronger opinion though, but usually each paper got 1-2 of those.

23.11.2024 20:29 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Everyone Should Learn Optimal Transport, Part 1 In my opinion, optimal transport (OT) is a seriously underrated topic. I think part of the reason is the way OT is often introduced: as an optimization problem or a metric on probability distributions...

Since I'm moving over, I would like to reshare my most recent blog post on why you should learn Optimal Transport. Part 2 is coming, most likely during the winter break :)

mufan-li.github.io/OT1/

22.11.2024 19:58 β€” πŸ‘ 36    πŸ” 6    πŸ’¬ 2    πŸ“Œ 0
Preview
YOU NEED MORE LUX | Meaningness Seriously bright indoor lights for seasonal affective disorder, or just the winter blahs

Great writeup:
meaningness.com/sad-light-le...

22.11.2024 01:35 β€” πŸ‘ 6    πŸ” 1    πŸ’¬ 2    πŸ“Œ 0

@mufan-li is following 20 prominent accounts