4 ms·

I feel like its even worse than just resorting to bad data when good data isn't available: the field of deep learning has cultivated the perception that it's ro

by zmmmmm 5y ago

I feel like its even worse than just resorting to bad data when good data isn't available: the field of deep learning has cultivated the perception that it's robust to bad data as one of its hallmarks. That is, you can pump relatively raw data into it and it will self-select features and then self-regulate their use so therefore most of the initial steps of data cleaning, feature selection etc are not necessary, or require less expertise. This is now spilling over into general ML so that when quacks assert that their model just magically overcomes these things and people actually believe it.

nerdponx 5y ago
The irony is that there are techniques for dealing with noisy/mislabeled/bad data (e.g. gold loss correction [0], errors-in-variables models [1]), but that stuff isn't "sexy" and not enough practitioners know about it. 0: https://arxiv.org/abs/1802.05300 https://arxiv.org/abs/1802.05300 1: https://en.m.wikipedia.org/wiki/Errors-in-variables_models https://en.m.wikipedia.org/wiki/Errors-in-variables_models
- topranks 5y ago
  Is this any more than fancy outlier detection (genuinely asking, this is not my field). i.e. if the majority of data fed to a system is bad, will it work?
  - nerdponx 5y ago
    They do different things. Both are useful. Outlier detection detects data points that look different from your existing data in some way; that "lie outside" what is usual. Sometimes the assumption is that outliers are generated by an entirely different process from the rest of the data. It's important not to conflate "outliers" with legitimate data points that happened to fall at the tail ends of the data distribution. The techniques I listed attempt to actually compensate for some known or estimable level of badness in the data. For example, gold loss correction (GLC) essentially estimates the probability of misclassifying a data point, and uses that to adjust the output of the model.
- verygoodname 5y ago
  This. The thing is that some of the techniques commonly applied when training NN are often "good enough" to deal with the presence of corrupted data (e.g. using SGD to optimize a model, while applying weight decay and drop-out, adds a regularization effect that somewhat replicates the effect of assuming errors-in-variables), as long as the input data is not total trash, which deters people from applying more formalized robust approaches to it. As long as "things kind of work", it is difficult to convince other people to adopt robust methods, particularly due to the existence of a "robustness vs. efficiency" trade-off (which can make robust methods seem additionally "unsexy").

by zmmmmm 5y ago

nerdponx 5y ago

The irony is that there are techniques for dealing with noisy/mislabeled/bad data (e.g. gold loss correction [0], errors-in-variables models [1]), but that stuff isn't "sexy" and not enough practitioners know about it. 0: https://arxiv.org/abs/1802.05300 https://arxiv.org/abs/1802.05300 1: https://en.m.wikipedia.org/wiki/Errors-in-variables_models https://en.m.wikipedia.org/wiki/Errors-in-variables_models

topranks 5y ago
Is this any more than fancy outlier detection (genuinely asking, this is not my field). i.e. if the majority of data fed to a system is bad, will it work?
- nerdponx 5y ago
  They do different things. Both are useful. Outlier detection detects data points that look different from your existing data in some way; that "lie outside" what is usual. Sometimes the assumption is that outliers are generated by an entirely different process from the rest of the data. It's important not to conflate "outliers" with legitimate data points that happened to fall at the tail ends of the data distribution. The techniques I listed attempt to actually compensate for some known or estimable level of badness in the data. For example, gold loss correction (GLC) essentially estimates the probability of misclassifying a data point, and uses that to adjust the output of the model.
verygoodname 5y ago
This. The thing is that some of the techniques commonly applied when training NN are often "good enough" to deal with the presence of corrupted data (e.g. using SGD to optimize a model, while applying weight decay and drop-out, adds a regularization effect that somewhat replicates the effect of assuming errors-in-variables), as long as the input data is not total trash, which deters people from applying more formalized robust approaches to it. As long as "things kind of work", it is difficult to convince other people to adopt robust methods, particularly due to the existence of a "robustness vs. efficiency" trade-off (which can make robust methods seem additionally "unsexy").