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

Machine Learning Confronts the Elephant in the Room

quantamagazine.org

Post reply on HN
Print view View original

91–100 of 116 posts

Re: Machine Learning Confronts the Elephant in the Room

#91

Post by CodesInChaos »

ivan_gammel wrote: post #37
sp332 wrote: post #35

Earlier quoted context omitted.

Yes, check out "attention" mechanisms. The network can use wide convolutions for the image as a whole, effectively mixing data kind of like a blur function, and also use finer convolutions for a small piece of the image that it chooses to focus on.

Does the problem of elephants still exist with such approach? It doesn't seem right that an unexpected feature in a corner of the image would have a dramatic impact on the recognition of the features elsewhere.

I don't see any evidence of "dramatic" impact. The objects visible in both images only show minor changes in the confidence score . The objects missing in one image are close to the visualization threshold (50% confidence) in the other. And for the "chair turning into a couch", we don't see the likelihood of the chair, because it's hidden behind the box for the couch.

So the dramatic changes seen in the image seem to be artifacts of the visualization and not of the network's performance.

Re: Machine Learning Confronts the Elephant in the Room

#93

Post by salty_biscuits »

sweezyjeezy wrote: post #47
salty_biscuits wrote: post #41

Earlier quoted context omitted.

I just spent 15 minutes putting a picture of a bear with a top hat riding Abraham Lincoln with laser eyes into all the online demos I could find. They weren't unintelligible responses, just a bit crap. They said "weapon", "animal", "person", etc. I'm not sure how much global context they really use, in my experience it has been latching onto a weird bit of local texture. More work to do...

Try this one : https://i.ytimg.com/vi/I11Vlcpz3jc/hqdefault.jpg I tried it on a couple of online demos, neither spotted that there was a car in the picture.

That is interesting. This is what I find most annoying about neural nets though, they are impossible to debug. Can you work out if they are failing because of global context (car in water) or local texture (weird swirling patterns and odd colors)?

Re: Machine Learning Confronts the Elephant in the Room

#94

Post by ivan_gammel »

ThePhysicist wrote: post #61
ivan_gammel wrote: post #43

Earlier quoted context omitted.

How close are those convolutional filters to the physical sight of animals? Were they designed specifically to mimic nature or they are artificial concept with similar purpose?

Actually the artificial neural networks used for vision tasks are not so different from our visual cortex, there's an interesting Nature paper on this: https://www.nature.com/articles/s42003-018-0110-y Of course the neurons in our brain work in a very different way, but at an architecture level the strategies of natural and artificial neural networks seem to match.

Thank you for the link! Indeed, that's very interesting reading.

Re: Machine Learning Confronts the Elephant in the Room

#95

Post by PavlikPaja »

CodesInChaos wrote: post #89
PavlikPaja wrote: post #81

Earlier quoted context omitted.

How well do neural networks train with no normalization at all, compared with softmax?

You need to perform some kind of normalization, since probability must be between 0 and 1 (and being wrong on a confident prediction gives huge penalties using the popular maximum likelyhood loss functions). But you can use component wise normalization (sigmoid) instead of combined normalization (softmax). These correspond to the assumption that the classes are independent (component wise sigmoid) or mutually exclusi…

"probability must be between 0 and 1" - why? (I get it's used in mathematics, but I see no reason why a NN would have to output probability that way.)

"and being wrong on a confident prediction gives huge penalties using the popular maximum likelyhood loss functions" - It should.

Re: Machine Learning Confronts the Elephant in the Room

#96

Post by QML »

weregiraffe wrote: post #31
hnhg wrote: post #21

Earlier quoted context omitted.

Because there's always another type of 'elephant' to be thrown in. The networks are just not inherently robust against unexpected features (I'll call them outliers for want of a better term), and this is what needs to be fixed, and not increasing the variety and frequency of outliers.

>Because there's always another type of 'elephant' to be thrown in. Well, you could randomly generate and render 3D objects, and paste them into photos. This will give you a huge space of 'elephants'.

That would not seem computationally efficient if for every class (i.e. object) you were to create a 3D modeling and paste it into a training set

Re: Machine Learning Confronts the Elephant in the Room

#97

Post by QML »

magicalhippo wrote: post #48
hnhg wrote: post #21

Earlier quoted context omitted.

Because there's always another type of 'elephant' to be thrown in. The networks are just not inherently robust against unexpected features (I'll call them outliers for want of a better term), and this is what needs to be fixed, and not increasing the variety and frequency of outliers.

Couldn't you train multiple nets with somewhat different training sets, and for each classification see how many agrees? If they all agree you have a high confidence, if they mostly disagree you have a low confidence.

That’s actually the concept of ensemble learning! [1] The question is, how would that lead to predicting a randomly placed elephant as an elephant with high confidence?

[1] https://en.m.wikipedia.org/wiki/Ensemble_averaging_(machine_...

Re: Machine Learning Confronts the Elephant in the Room

#98

Post by QML »

sorokod wrote: post #6

Is this not a whole new attack surface?

Adversarially generated inputs have been known for some time [1], so I wouldn’t call it a new class of attacks unless 1. there are other and older ways of attacking neural networks 2. new as in the last couple of years.

[1] https://blog.openai.com/adversarial-example-research/

Re: Machine Learning Confronts the Elephant in the Room

#99

Post by p1esk »

dukoid wrote: post #54

Wouldn't humans who haven't seen Elephants before or who are not trained to recognize the general category of mammals / animals be prone to make similar errors?

In this case the object detector has been trained to recognize elephants. It's just never seen elephants in such a setting (living room).

It would be similar to a situation where a human, who sees an improbable object for a given environment, reevaluates his initial perception/understanding of that environment. I can't come up with a good example of such a situation, but you get the idea.

Re: Machine Learning Confronts the Elephant in the Room

#100

Post by p1esk »

QML wrote: post #98
sorokod wrote: post #6

Is this not a whole new attack surface?

Adversarially generated inputs have been known for some time [1], so I wouldn’t call it a new class of attacks unless 1. there are other and older ways of attacking neural networks 2. new as in the last couple of years. [1] https://blog.openai.com/adversarial-example-research/

Yes, this can be used for generating adversarial inputs, and yes, this is a new attack surface. From the paper:

The images generated here could be viewed as a variant of adversarial examples, in which small image perturbations (imperceptible to humans) cause a large shift in the network’s output. The images we generate are of a somewhat opposite flavor: while we do not limit the magnitude of the difference between the original and modified image, the detectors are sometimes “blind” to the inserted object. In addition, our examples are not “targeted” in the sense that no optimization process is required to generate them; they seem prevalent enough so that a simple scan of transplanting translated versions of one object in the other can give rise to multiple wrong interpretations.

Post reply on HN