Measuring the tendency of CNNs to Learn Surface Statistical Regularities · HackerPapers discussions

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — mentioned by gwern

Thu, 24 Sep 2020 15:16:35 GMT

Yes, because your validation set performance is imperfectly correlated with performance on more samples (because it's of finite size), so it will tend to overestimate within-distribution. There's also the problem that you probably want to deploy it on data which was not exactly collected the same way. So you have issues of both internal and external validity. In practice, at least for CNN classifiers on ImageNet etc, it seems to be a pretty minor issue overall. Some links: https://arxiv.org/abs/1711.11561 https://arxiv.org/abs/1806.00451 https://arxiv.org/abs/1902.10811…

Paper and discussion history. A mention is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — mentioned by stochastic_monk

Tue, 03 Jul 2018 13:24:37 GMT

In light of [0], there is both theoretical and experimental reason to believe that CNNs are mostly picking up fourier decomposition. [0] https://arxiv.org/pdf/1711.11561.pdf

Paper and discussion history. A mention is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — mentioned by nabla9

Thu, 21 Jun 2018 15:36:56 GMT

The AI winter happens when outside hype and investment money overshoots what can be delivered in the timescale needed for good ROI. Disappointments are followed with periods of underinvestment. Neural Networks are already decades old invention. Underlying all this new boom is the same basic architecture, alternating layers of affine transformations and nonlinear transformations trained with backpropagation and gradient descent. What made the field suddenly explode was GPGPU's and bunch of tweaks that helped to solve vanishing/exploding gradients problems (started with RNN training layers and…

Paper and discussion history. A mention is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — mentioned by mlthoughts2018

Tue, 19 Jun 2018 21:52:55 GMT

This seems like it could dramatically worsen overfitting-like effects for algorithms like CNNs for image processing, where surface statistics of the available data set seem to be more responsible for the learned model than any type of “semantic” understanding. If you prespecify what data augmentation you would do, like preregistering the details of a clinical trial, you’ll be less susceptible to a spurious result from this. It seems like especially things like color distribution manipulation would have a potentially very adverse effect that counters any gains from clamping the supervised…

Paper and discussion history. A mention is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — mentioned by denzil_correa

Sun, 10 Jun 2018 18:17:12 GMT

One of the cited papers "Measuring the tendency of CNNs to Learn Surface Statistical Regularities" is also a great insight into this phenomena [0]. > Our main finding is that CNNs exhibit a tendency to latch onto the Fourier image statistics of the training dataset, sometimes exhibiting up to a 28% generalization gap across the various test sets. Moreover, we observe that significantly increasing the depth of a network has a very marginal impact on closing the aforementioned generalization gap. Thus we provide quantitative evidence supporting the hypothesis that deep CNNs tend to learn…

Paper and discussion history. A mention is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — mentioned by mlthoughts2018

Sun, 03 Jun 2018 02:13:13 GMT

A better title for the article might have been, “Why current deep learning is not general enough for AGI or some other families of structured problems.” Viewed this way, it’s not a criticism of pragmatically using deep learning or experimenting with deep learning for some narrow tasks. Rather, it expresses what aspects of current deep learning make it unsuitable for general transfer learning, hierarchical or causal inference, or many Bayesian techniques requiring greater use of priors. Other comments have pointed out that there are plausible rebuttals in deep reinforcement learning and…

Paper and discussion history. A mention is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — mentioned by mlthoughts2018

Fri, 01 Jun 2018 01:42:50 GMT

I don’t work in the “high-tech HR” space, but do have experience with embedding vector methods, joint-embeddings, and trying to solve generic information retrieval tasks, where the underlying data is not text or images, by using embedding methods. Even in a domain like reverse image search, where this approach has been studied to death, there are big concerns about how much of the problem is solved because a vector similarity approach can approximate the real manifold distance of some underlying true structure vs. just having huge models overfit to a particular class of natural image…

Paper and discussion history. A mention is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — mentioned by tc

Mon, 22 Jan 2018 16:36:05 GMT

As one example of how this can be useful, Jo and Bengio recently used Fourier filtering to measure the susceptibility of neural networks to adversarial examples. By changing the statistics of the images in a principled manner, they confirmed that even networks that generalize well are learning mostly surface-level statistics. E.g. an image of a car is more likely to also have asphalt and building colors than greenery. The networks overweight these kinds of features, and that turns out to be good enough to get high scores on many data sets. Using Fourier filtering, they were able to alter the…

Paper and discussion history. A mention is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — mentioned by eggie5

Wed, 17 Jan 2018 20:11:26 GMT

I would say training a CNN from scratch or even fine-tuning one takes a lot of domain knowledge and best practices which often are not standardised yet. Besides, we don't even know why they generalise in the first place! See: https://arxiv.org/abs/1611.03530, https://arxiv.org/abs/1711.11561

Paper and discussion history. A mention is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — HN submission

Wed, 17 Jan 2018 03:02:46 GMT

Measuring the Tendency of CNNs to Learn Surface Statistical Regularities

66 points; 36 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Measuring the tendency of CNNs to Learn Surface Statistical Regularities — HN submission

Tue, 16 Jan 2018 06:17:37 GMT

CNNs learn surface statistical regularities, not high-level abstract concepts

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.