Lightnews — Scholar-powered news

Ben Recht

@beenwrekt.bsky.social

7.3K followers 220 following 1.3K posts

Blog: https://argmin.substack.com/ Webpage: https://people.eecs.berkeley.edu/~brecht/

argmin.substack.com

Posts Media Videos Starter Packs

Ben Recht @beenwrekt.bsky.social · 2h

Amazingly, multiple NFL games have ended with the beautiful score 11-10.

All Games in Pro Football History with a 11 to 10 score | Pro-Football-Reference.com

All Games in Pro Football History with a 11 to 10 score

www.pro-football-reference.com

Ben Recht @beenwrekt.bsky.social · 1d

oh totally. The formatting that gives an air of authority.

It's funny how many word documents are uploaded on the bioRxiv. docx makes it clear that you shouldn't take what you are reading too seriously.

1 1

Ben Recht @beenwrekt.bsky.social · 1d

Maybe it would be better if we all gave up on arXiv and just wrote blogs. Not even sure if I'm joking.

1 1

Ben Recht @beenwrekt.bsky.social · 1d

I know I'm mostly alone on this island, but I'm not sure the norms are of much help. That said, I'm open to new evidence if things are really getting out of hand.

1 1

Ben Recht @beenwrekt.bsky.social · 1d

yes, but most published papers may as well have been blog posts.

Ben Recht @beenwrekt.bsky.social · 1d

NGL: the duet with Hayley Williams hits me hard.

Ben Recht @beenwrekt.bsky.social · 1d

The new Jay Som record is so good. jaysom.bandcamp.com/album/belong

Belong, by Jay Som

11 track album

jaysom.bandcamp.com

2 8

Ben Recht @beenwrekt.bsky.social · 1d

It's been all downhill since Andy Rooney died.

Ben Recht @beenwrekt.bsky.social · 3d

but there's (3) predictions of bad practice based on the iid model seldom pan out.

this is a lesson that I'd like social scientists to pay more attention to.

1 1

Ben Recht @beenwrekt.bsky.social · 3d

OK, this is conflating two things (my fault).

(1) the iid assumption is a stupid crutch we use to apply the law of large numbers willy nilly
(2) in machine learning, all we know is that more data is better.

(1) applies to social science, (2) doesn't.

1 1

Ben Recht @beenwrekt.bsky.social · 3d

I don't think I understand. Who is collecting small data from a non-adaptive system?

1 2

Ben Recht @beenwrekt.bsky.social · 3d

Are they real problems, though? The evidence is far less clear than the replication police want you to believe.

1 3

Ben Recht @beenwrekt.bsky.social · 3d

Almost a decade ago, I coauthored a paper asking us to rethink our theory of generalization in machine learning. Today, I’m fine putting the theory back on the shelf.

Reshelving generalization

You don't need a theorem to argue more data is better than less data

www.argmin.net

1 4 19

Ben Recht @beenwrekt.bsky.social · 3d

Astute. How do we break that cycle?

2 1

Ben Recht @beenwrekt.bsky.social · 3d

Would be happy to discuss. Though I find myself confused more often than not every time I try to read a paper from either camp. They are all convinced that explanation lies in the models rather than in the training data.

1 7

Ben Recht @beenwrekt.bsky.social · 3d

Absolutely. Yes.

Ben Recht @beenwrekt.bsky.social · 3d

LOL, agree.

Ben Recht @beenwrekt.bsky.social · 3d

There's been a bifurcation into two camps, one in the FACCT orbit and one in the Less Wrong orbit, and they want different things.

Which camp were you interested in?

3 6

Ben Recht @beenwrekt.bsky.social · 4d

I should give Brave a spin. I was using Arc, but the company pivoted to AI and abandoned their browser.

Ben Recht @beenwrekt.bsky.social · 4d

Keeping it real.

1 2

Ben Recht @beenwrekt.bsky.social · 4d

Kevin successfully shamed me into closing all of my tabs. And for that I am thankful.

1 10

Ben Recht @beenwrekt.bsky.social · 4d

Yes. What is the diagnosis if I create dysfunctional to-do lists in multiple modalities?

I'm aware I need help.

2 3

Ben Recht @beenwrekt.bsky.social · 4d

I'd die. Currently rocking 32.

Also, I am aware this is real Sicko shit...

Ben Recht @beenwrekt.bsky.social · 4d

I'm always haunted by the splash image for your Boston Review essay.

A conversation. Or, Operation Head. By Niki Sublime. https://www.flickr.com/photos/nikisublime/221259238/

1 1

Ben Recht @beenwrekt.bsky.social · 4d

I couldn't figure out how to elegantly work a citation into the main text, so I went with an oblique reference in the title. I'm going to add a couple of pointers at the end to your essays.

1 1