Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Attention Is Off By One

evanmiller.org

Post reply on HN
Print view View original

151–160 of 347 posts

Re: Attention Is Off By One

#151

Post by PoignardAzur »

cratermoon wrote: post #117
r3trohack3r wrote: post #55

Earlier quoted context omitted.

There was this sociologist who had written a paper for us all to read ahead of time. I started to read the damn thing, and my eyes were coming out: I couldn’t make head nor tail of it! I figured it was because I hadn’t read any of the books on the list. I had this uneasy feeling of “I’m not adequate,” until finally I said to myself “I’m gonna stop, and read one sentence slowly so I can figure out what the hell it mea…

I believe Feynman understood that he was oversimplifying, and I believe he was able to do because his reason for reading the paper was not the same as the reason another sociologist might have. Thus a sentence like, "The individual member of the social community often receives his information via visual, symbolic channels", does, to a non-expert, mean "people read", but to another sociologist of a researcher in relat…

Well, maybe, but you can rationalize arbitrary amounts of pointless jargon that way.

Besides, in the example Faynman gives the simple sentence is actually shorter. Maybe that shorter sentence loses some information that the jargon carried, but Occam's razor suggests the writer was just trying to sound smarter.

Re: Attention Is Off By One

#152

Post by chessgecko »

zackangelo wrote: post #136
chessgecko wrote: post #101

Earlier quoted context omitted.

I guess yeah I was mostly responding to Now it’s possible that softmax should be replaced wholesale, but it’s worked pretty well for the most part, except for this one wee little bug that prevents attention heads from saying nothing. So I propose a very small tweak on which I am willing to stake all future Internet claims to being correct. The tweak is so small, yet so obvious, and it’s been sitting here under everyo…

He’s not trying or claiming to improve attention. He’s trying to reduce outliers to improve the ability to quantize the parameters.

He refers all over the blog post to an "error" in attention. specifically says

The problem with using softmax is that it forces each attention head to make an annotation, even if it has no information to add to the output vector. Using softmax to choose among discrete alternatives is great; using it for optional annotation (i.e. as input into addition) is, like, not cool, man.

I'm saying it uses the current position to do this, that if it was a significant error I would expect it to improve the training loss. I sort of interpreted the blog post as being a bit more positive on the idea than just being about improving the quantization

Re: Attention Is Off By One

#153

Post by PoignardAzur »

WithinReason wrote: post #111
janalsncm wrote: post #78

Earlier quoted context omitted.

> I believe the outlier problem that this solves only appears for very large models. Any reason to believe this? The author never mentioned it, and I can’t think of any other a priori reason why it should be true.

See figure 1: https://arxiv.org/pdf/2208.07339.pdf Outliers appear at model size 6.7B and are not present at 2.7B

6.7B isn't "needs a datacenter" scale.

Re: Attention Is Off By One

#154

Post by karaterobot »

doliveira wrote: post #30
mlsu wrote: post #8

I don't really understand the subject matter enough, so I apologize in advance for the meta-comment... The author mentions that he would maybe have written this as a scientific paper: > I tried writing a serious-looking research paper about the bug and my proposed fix, but I lost a series of pitched battles against Pytorch and biblatex, so I figured I’d just write a blog post instead. (History is written by the winne…

Nah, scientific papers are supposed to be precise and technical. This reads like those quite frequent suggestions here of switching all equations in papers to plain English or code: it honestly comes from a place of ignorance, and I say that as basically a layman myself. What should be encouraged is for academics to blog about their research as well. It would even help when recruiting and onboarding new members. Righ…

The writing quality of academic papers is very poor, whatever its intended characteristics are, and we deserve better.

I'm skeptical that the only way for them to be precise and technical is to make them impenetrable. I think there is a culture of academic writing (many different cultures, really) that has adopted a voice and writing style which became a parody of itself over time.

Here's a trivial example: You frequently see papers use the passive voice, something a middle school English teacher would mark with a red pen. 500 participants were asked, vs. we asked 500 participants. In what sense is the former more precise and technical? It's not. It does not convey any additional meaning. People use it to sound objective and distant, even when they really aren't.

Realistically, academic writers usually don't even think about it as much as that. They're just copying the tone of other papers, because there is a culture and it enforces certain behaviors on its members irrespective of the value.

Re: Attention Is Off By One

#155

Post by dekhn »

coldtea wrote: post #148
dekhn wrote: post #132

Earlier quoted context omitted.

I hadn't seen that Feynman quote before, but I discovered then when reading Donna Harraway's books (Cyborg Manifesto, Modest_Witness@Second_Millennium.FemaleMan©Meets_OncoMouse, Primate Visions). The criticism was """Haraway's work has been criticized for being "methodologically vague"[39] and using noticeably opaque language that is "sometimes concealing in an apparently deliberate way""""

> Haraway's work has been criticized for being "methodologically vague"[39] and using noticeably opaque language that is "sometimes concealing in an apparently deliberate way So you're saying that "Her work is basically handwaving and bullshitting".

Yes, but also, wrapping the handwaving and bullshitting in a layer of obfuscation:

"Michel Foucault’s biopolitics is a faccid premonition of cyborg politics, a very open feld. By the late twentieth century, our time, a mythic time, we are all chimeras, theorized and fabricated hybrids of machine and organism—in short, cyborgs. The cyborg is our ontology; it gives us our politics. The cyborg is a condensed image of both imagination and material reality, the two joined centers structuring any possibility of historical transformation. In the traditions of “Western” science and politics—the tradition of racist, male-dominant capitalism; the tradition of progress; the tradition of the appropriation of nature as resource for the productions of culture; the tradition of reproduction of the self from the refections of the other—the relation between organism and machine has been a border war"

(donna was woke before woke was a thing)

Re: Attention Is Off By One

#156

Post by refulgentis »

mellosouls wrote: post #131

Unless he gives a good reason why he has not demonstrated his claim (eg. "This effect only presents at a scale beyond my resource"), the thesis seems severely weakened by the lack of effort to prove it in a toy version. He just says he doesn't want to spend any more time on it, which is unlikely to convince or motivate anybody else that he has discovered something important.

It got tons of people really excited.

I don't know what to say past that, but it's worth reflecting on.

Re: Attention Is Off By One

#157

Post by WithinReason »

PoignardAzur wrote: post #153
WithinReason wrote: post #111

Earlier quoted context omitted.

See figure 1: https://arxiv.org/pdf/2208.07339.pdf Outliers appear at model size 6.7B and are not present at 2.7B

6.7B isn't "needs a datacenter" scale.

It's in the million dollar range. XLnet which is a 1.3B model cost $245,000 to train for example.

Re: Attention Is Off By One

#158

Post by madrox »

jrochkind1 wrote: post #61

While not about AI or the algorithm mentioned, on the subject of little errors that you can't convince anyone are errors.... In 2011, I wanted to copy the reddit ranking algorithm in a project of my own, so I went to source code to look at it... the algorithm in the source code I found wasn't doing anything at all sensible with negative-sum voted posts. I thought I discovered the error, some terms swapped in the simp…

When I was an intern at Yahoo working on OAuth back in 2008 (2007? It was long ago and I'm old) I had the pleasure of implementing an internal tool for generating OAuth 1.0 URLs, which meant encoding a lot of things in query parameters. My tool did not generate URLs which were compatible with Yahoo's implementation (certain parameters effectively should be encoded twice, which my tool did). The implementing engineer insisted my tool was wrong, cited my status as a lowly intern, and even pulled out the OAuth spec and bent over backwards to say how his implementation was correct and I'm clearly reading it wrong. It literally took bringing in Eran Hammer-Lahav to weigh in on the topic to say I was correct, at which point the engineer agreed that of course that was correct. I got zero acknowledgment or apology for the days of ad hominem attacks against me.

I did learn an important lesson that more senior people are not always right, and as someone who's usually more senior than my colleagues now I try to remember it daily.

Re: Attention Is Off By One

#159

Post by smallnamespace »

acchow wrote: post #139
Vervious wrote: post #103

Earlier quoted context omitted.

Sure. But it's an ELI5 "in addition to formalism", not "in lieu of formalism". In theory conferences like STOC or FOCS, the first section of the paper often comprises such an overview. Certainly some papers are better written than others. But sometimes a blog post cannot replace a paper, unless it also goes into the depth and detail that formalism requires. (Then it becomes a 30 page blog post, where most people don'…

The complaint about research papers is that almost all of them omit the ELI5 and provide only the formalism. You can have both and weave them together into a digestible narrative. I see Physics textbooks sometimes written this way.

Papers are mostly read by other researchers, where the added background is actively bad because it obscures the real meat of the paper to the main audience.

If you just wanted a digestible intro then you would usually buy a textbook.

I think the argument that every research paper ought to be a mashup of a textbook + the actual research to be a bit silly from a “people should specialize at what they’re good at” standpoint.

Put in another context, I also don’t want every recipe to reintroduce what it means to “fry” or “braise” or “marinate”. We have Google for that.

Re: Attention Is Off By One

#160

Post by caddemon »

pessimizer wrote: post #114
mlsu wrote: post #8

I don't really understand the subject matter enough, so I apologize in advance for the meta-comment... The author mentions that he would maybe have written this as a scientific paper: > I tried writing a serious-looking research paper about the bug and my proposed fix, but I lost a series of pitched battles against Pytorch and biblatex, so I figured I’d just write a blog post instead. (History is written by the winne…

> The tone is self-effacing, it does not have an "ego" the way scientific papers tend to have. I can't imagine judging scientific papers based on whether the author might be looking down on me, or thinks he knows better than me. > if we were "allowed" to cite research that reads like this Maybe you're looking down on yourself? You can cite anything you want to cite.

Well if you yourself are trying to publish in a scientific venue you can't always cite exactly what you want to cite. Though it's probably uncommon for a peer reviewer to ask for a specific citation to be removed, the review process absolutely does affect the references list, and expectations about this process affect it doubly so.
Post reply on HN