Cleaning Up Incorrectly Labelled Data (C3W2L02)

The data for your supervised learning problem comprises input X and output labels Y. What if you're going through your data and you find that some of these output labels Y are incorrect, that you have data which is incorrectly labeled, Is it worth your while to go in to fix up some of these labels? Let's take a look. In the cat classification problem, y equals 1 for cats and 0 for non cats. So let's say you're looking through some data and that's a cat. That's not a cat. That's a cat. That's a cat. That's not a cat. That's a cat. No wait, that's actually not a cat. So this is an example with an incorrect label. So I've used the term mislabeled examples to refer to if your learning algorithm.

Outputs the wrong value of y.

But I'm going to say incorrectly labeled examples when it to refer to if in the data set you have in the training set or the dev set or the test set, the label for Y, whatever human label assigned to this piece of data is actually incorrect and that's actually a dog. So that Y really should have been zero. But maybe the labeler got that one wrong. So if you find that your data has some incorrectly labeled examples, what should you do? Well, first let's consider the training set. It turns out that deep learning algorithms are quite robust to random errors in the training set. So so long as your errors or your incorrectly labeled examples, so long as those errors are not too far from random, you know, maybe sometimes the laborer just wasn't paying attention or they accidentally randomly hit the wrong key on the keyboard. If the errors are reasonably random, then it's probably okay to just leave the.

Errors as they are and not spend.

Too much time fixing them. Um, there's certainly no harm to going into your training set and re examining.

The labels and fixing them.

Sometimes that is worth doing, but your algorithm might be okay, okay, even if you don't, so long as the total data set size is big enough and the actual percentage of errors is, you know, maybe not too high. So I see a lot of machine learning algorithms that trained even when we know that there are few X mistakes in the training set labels.

And usually it works okay.

There is one caveat to this, which is that deep learning algorithms are robust random errors. They are less robust to systematic errors. So for example, if your labelist consistently labels white dogs as cats, then that is a problem because your classifier will.

Learn to classify all white colored dogs as cats.

But random errors or, you know, near random errors are usually not too bad.

For most deep learning algorithms.

Now, this discussion has focused on what to do about incorrectly labeled examples in your training set. How about incorrectly labeled examples in your dev set or test set? If you're worried about the impact of incorrectly labeled examples on your dev set or test set, what I recommend you do is during error analysis, to add one extra column so that you can also count up the number of examples.

Where the label yes was incorrect.

So, for example, maybe when you count up the impact on 100 mislabeled dev set examples, and so you're going to find 100 examples where your classifier's output.

Disagrees with the label in your dev set.

And sometimes for a few of those examples, your classifier disagrees with the label.

Because the label was wrong rather than.

Because your classifier was wrong. So maybe in this example you find that the labeler missed a cat in.

The background, so put a check mark.

There to signify that example 98 had an incorrect label. And maybe for this one, the picture is actually a picture of a drawing of a cat rather than a real cat.

And maybe you want the label to have labeled at y equals 0 rather than y 1, and so you put.

Another check mark there. And just as you count up the percent of errors due to other categories, like we saw in the previous video, you'd also count up the fraction of.

Percentage of errors due to incorrect labels.

Where the Y value in your dev.

Set was wrong and that accounted for.

Y, your learning algorithm made a prediction that differed from what the label on your data says. So the question now is, is it worthwhile going in to try to fix up, you know, this 6% of incorrectly labeled examples? My advice is, if it makes a significant difference to your ability to evaluate algorithms on your dev set, then go ahead and spend the time to fix the incorrect labels. But if it doesn't make a significant.

Difference to your ability to use the.

Dev set to evaluate classifiers, then it might not be the best use of your time. Let me show you an example that.

Illustrates what I mean by this.

So, three numbers I recommend you look at to try to decide if it's worth going in and reducing. The number of mislabeled examples are the following. Recommend you look at the overall dev set error. And so in the example we had from the previous video, we said that maybe your system has 90% overall accuracy, so 10% error. Then you should look at the number of errors or the percentage of errors that are due to incorrect labels. So it looks like in this case, 6% of the errors are due to incorrect labels. So 6% of 10% is 0.6%. And then you should look at errors.

Due to all other causes.

So if you made 10% error on your dev set, and if 0.6% of those are because the labels are wrong, then the remainder 9.4% of them are due to other causes, such as misrecornizing.

Dogs, great cats, and blurry images.

So in this case, I would say there's 9.4% worth of error that you could focus on fixing. Whereas, you know the errors due to incorrect labels is a relatively small fraction.

Of the overall set of errors.

So by all means, go in and fix these incorrect labels if you want, but it's maybe not the most important thing to do right now. Now, let's take another example. Suppose you've made a lot more progress.

On your learning problem.

So instead of 10% error, let's say.

You brought the errors down to 2%.

But still 0.6% of your overall errors are due to incorrect labels. So now if you were to examine a set of mislabeled dev set images, so set that comes from this 2% of dev set data you're mislabeling, then a very large fraction of them 0.6 divided by 2%, so that's actually 30% rather than 6% of your labels of your incorrect examples are actually due to incorrectly labeled examples. And so errors due to other causes are now 1.4%. When such a high fraction of your mistakes, at least as measured on your dev set, are due to incorrect labels, then it maybe seems much more worthwhile to fix up the incorrect labels in your dev set.

And if you remember the goal of the dev set, the main purpose of the dev set is you want to.

Really use it to help you select.

Between two classifiers, A and B.

So if you're trying out two class.

S, A and B, and one has 2.1% error and the other has 1.9% error on your dev set, but you don't trust your dev set anymore to be correctly telling you whether this classifier is actually better than this, because 0.6% of these mistakes are due to incorrect labels, then there's a good reason to go in and fix the incorrect labels in your dev set.

Because in this example on the right.

It'S just having a very large impact on the overall assessments of the errors of the algorithm. Whereas an example on the left, you know the percentage impact it's having on Your algorithm is still smaller.

Now, if you decide to go into.

Your dev set and manually re examine the labels and try to fix up some of the labels, here are a few additional guidelines or principles to consider. First, I would encourage you to apply whatever process you apply to both your depth and test sets at the same time.

We've talked previously about why you want.

Your depth and test sets to come from the same distribution. The dev set is sort of telling you where to aim to target, and when you hit it, you want that to generalize to the test set so your team really works more efficiently if the dev and test sets come from the same distribution.

So if you're going in to fix.

Up your dev set, I would apply the same process to the test set to make sure that they continue to come from the same distribution. So if you hire someone to examine the labels more carefully, do that for both your dev and test sets.

Second, I would urge you to consider.

Examining examples your algorithm got right as well as ones it got wrong. It is easy to look at the examples your algorithm got wrong and just see if any of those need to be fixed.

You but it's possible that there's some.

Examples that your algorithm got right that should also be fixed. And if you only fix ones that your algorithms got wrong, you end up with a more biased estimate of the error of your algorithm. It kind of gives your algorithm a little bit of an unfair advantage if you just try to double check what it got wrong, but you don't also double check what it got right because it might have gotten something right that it was just lucky on. And fixing the label would cause it to go from being right to be wrong. On that example, the second bullet isn't always easy to do, so it's not always done.

The reason it's not always done is.

Because if your classifier is very accurate, then it's getting a lot fewer things wrong than right. So if your classifier has, you know.

98% accuracy, then it's getting 2% of.

Things wrong and 98% of things right. So it's much easier to examine and validate the labels on 2% of the data, and it takes much longer to validate the labels on 98% of the data. So this isn't always done, but it's just something to consider.

Finally, if you go into your DEP and test data to correct some of the labels there, you may or may.

Not decide to go and apply the same process for the training set.

Remember, we said at the start of this video that it's actually less important.

To correct the labels in your training set.

And it's quite possible you decide to.

Just correct the labels in your dev.

And test set, which are also often smaller than the training set, and you might not invest all the extra effort needed to correct the labels in the.

Much larger training set.

This is actually okay. And we'll talk later this week about some processes for handling when your training.

Data is different in distribution than your depth and test data.

Learning algorithms are quite robust to that is super important that your dev and test sets come from the same distribution, but if your training set comes from a slightly different distribution, often that's a.

Pretty reasonable thing to do.

And we'll talk more about how to.

Handle this later this week.

So I'd like to wrap up with just a couple pieces of advice. First, deep learning researchers sometimes like to say things like, oh, I just fed the data to the algorithm and I trained it and it worked. And you know, there is a lot of truth to that. In the deep learning era, there is more of feeding data to an algorithm and just training it and doing less hand engineering, using less human insight. But I think that in building practical systems, often there's also more manual error analysis and more human insight that goes into these systems. And sometimes deep learning researchers like to acknowledge. Um, second is that somehow I've seen some engineers and researchers be reluctant to manually look at examples. Maybe it's not the most interesting thing.

To do to sit down and look.

At 100 or a couple hundred examples to counter the number of errors, but this is something that I still do myself when I'm leading a machine learning team and I want to understand what mistakes it's making. I will actually go in and look at the data myself and try to counter the fraction of errors. And I think that because these, you know, minutes or maybe small number of hours of counting data can really help you prioritize where to go next. I find this a very good use of your time and I urge you.

To consider doing it if you've built.

A machine learning system and you're trying to decide what ideas or what directions to prioritize next. So that's it for the error analysis process. In the next video, I want to share with you some thoughts on how error analysis fits into how you might go about starting out on a new machine learning project.