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

Shirt Without Stripes

github.com

Post reply on HN
Print view View original

261–270 of 639 posts

Re: Shirt Without Stripes

#261

Post by tines »

bartread wrote: post #221
Alex3917 wrote: post #212

Earlier quoted context omitted.

At least this is relatively innocuous. Until recently if you did a Google Image Search for "person" or "people", it only showed white men.

I couldn't quite believe your comment when I read it so I did a Google image search for "person" and the results weren't a lot better than you'd suggested. Mostly white men, a few white women, a very few black women, a handful of Asians, and multiple instances of Terry Crews. The net result of that Google search, combined with the "Shirt Without Stripes" repo, leaves me even more unimpressed with the capabilities of…

I think the skewing of results lessening your impressed-ness is the wrong takeaway. If anything, the AI is a more perfect mirror of the society it learned from than you expected. Perhaps the right way to look at it is that we are capable of producing things that we don't understand, that are more sophisticated than we realize.

Re: Shirt Without Stripes

#262

Post by _kst_ »

Alex3917 wrote: post #212
rgovostes wrote: post #23

The point that the author is making, in a very understated way, is that all three companies have PR websites that breathlessly describe their advanced AI capabilities, yet they cannot understand a very simple query that young children can.

At least this is relatively innocuous. Until recently if you did a Google Image Search for "person" or "people", it only showed white men.

A former coworker had the last name "Person". They once received a letter (snail mail) addressed to "TheirFirstName Human".

I never figured out what kind of mistake could have led to that.

Re: Shirt Without Stripes

#263

Post by dhimes »

animalCrax0rz wrote: post #259
albertzeyer wrote: post #243

Why should it not be possible to solve this with statistical methods? The model just needs to be able to understand the important meaning of "no" in here, in the context of the whole sentence. I would guess that most modern NNs from the NLP area (Transformer or LSTM) would be able to correctly differentiate the meaning. The problem is, I think there is no fancy NN (yet) behind Google search, and the other web searche…

"there is no evidence of cancer" and "there is evidence of no cancer" are two different statements with different meaning, so it's more complex a task than just understanding the importance of "no" in a sentence. It's involves semantic analysis of the sentence. The paper I linked to below describes a technique they call "deep parsing." Check it out for more context.

While it was an interesting point, it doesn't seem completely applicable here. Presumably the classifications include context. "A shirt with no stripes" should be distinguishable from "a stripes with no shirt" in the this context.

Re: Shirt Without Stripes

#264

Post by lolc »

shawnz wrote: post #233
cheschire wrote: post #171

Earlier quoted context omitted.

That is an inclusive search though. What if the user actually wanted to simply exclude stripes but wanted to see polka dots and paisley?

I think that is probably not a common enough use case to optimize for. Additionally it would be easy for a user in that position to just search "shirt" and ignore the occasional striped one, or to search "polka dot" and "paisley" seperately

You're arguing the search is "good enough". Because we can adapt to the machines. But the company who will not force us to do so will get our business. The company who creates the best digital butler will win. They know this. They try hard. And they still fail at simple stuff when judged by humans.

Re: Shirt Without Stripes

#265

Post by Lammy »

6gvONxR4sf7o wrote: post #246
Lammy wrote: post #231

Earlier quoted context omitted.

I think it's disingenuous to say "the web is basically google's adversary" when Google AdWords is the reason so many pages fight for top ranking.

Well sure, but I was hoping my narrow meaning would be clear: the search team operates in an adversarial environment.

The Search team isn't some helpless independent group adrift on currents outside of their control.

Re: Shirt Without Stripes

#266

Post by smoyer »

DenisM wrote: post #197

This problem is known as "attribution" - you have a "no" or "without" in the sentence, but you don't know where it belongs. One could (and one does) argue that the problem cannot be solved with statistical methods (ML), especially not in any domain where accuracy is required, such as medical recored analysis: "no evidence of cancer" and "evidence of no cancer" are very different things. Zooming out, the language fiel…

While this is indeed an example of the attribution problem, I'd argue that this particular query will never be solved. I don't search for a "shirt without stripes", I search for a "solid shirt, or a " hawai'ian shirt".

I'd be curious to see how many sentences with attribution problems actually have other structural issues. If I want to write clearly and without ambiguity, I rewrite sentences that have these problems. Why wouldn't I do the same for search queries?

Re: Shirt Without Stripes

#267

Post by dhimes »

_kst_ wrote: post #262
Alex3917 wrote: post #212

Earlier quoted context omitted.

At least this is relatively innocuous. Until recently if you did a Google Image Search for "person" or "people", it only showed white men.

A former coworker had the last name "Person". They once received a letter (snail mail) addressed to "TheirFirstName Human". I never figured out what kind of mistake could have led to that.

Maybe a veterinarian's customer database? They would have to distinguish pet names from humans, but keep a record of all.

Re: Shirt Without Stripes

#268

Post by mcbits »

bartread wrote: post #221
Alex3917 wrote: post #212

Earlier quoted context omitted.

At least this is relatively innocuous. Until recently if you did a Google Image Search for "person" or "people", it only showed white men.

I couldn't quite believe your comment when I read it so I did a Google image search for "person" and the results weren't a lot better than you'd suggested. Mostly white men, a few white women, a very few black women, a handful of Asians, and multiple instances of Terry Crews. The net result of that Google search, combined with the "Shirt Without Stripes" repo, leaves me even more unimpressed with the capabilities of…

Bing results, "person" shows stick figure drawings, Pearson Education logos, Person of the Year, people named Person, etc.

"Person without stripes" shows several zebras, tigers, a horse painted like a zebra, and a bunch of people with stripes.

Re: Shirt Without Stripes

#269

Post by jcburns »

chrisseaton wrote: post #201
exhilaration wrote: post #181

Earlier quoted context omitted.

I disagree, "shirt without stripes" is an unusual word choice, not one that our ML models would be optimized for. Try "solid color shirt" and you'll see how much better the results are - at least on Google.

> "shirt without stripes" is an unusual word choice Lol what? These are words a toddler would understand.

nothing unusual about those words in that sequence. unambiguous.

Re: Shirt Without Stripes

#270

Post by sonar_un »

seiferteric wrote: post #137

I have noticed in the past few years google results have become noticeable worse for similar reasons. Google used to _surprise_ me with how good it was able to understand what I was really looking for even when I put in vague terms. I remember being shocked on several occasions when putting in half remembered sentences, lyrics, expressions from something I had heard years ago and it being the first! result. I almost…

I have also found that search results are getting frustratingly worse. Often even when I put in explicit search terms and quotes, and filter out words that I don't want, Google will return results that don't adhere to what I am looking for or just return no results at all. I remember when I would search for something and find much more relevant information. Now the first 5 or 6 search results are ad-sponsored and aren't relevant, but I have to go to the 3rd or 4th page to find something that matches. I also often have to search for things that were posted in the last year or less because the older postings are increasingly irrelevant.
Post reply on HN