Errata

Machine Learning Design Patterns

Errata for Machine Learning Design Patterns

Submit your own errata for this product.

The errata list is a list of errors and their corrections that were found after the product was released. If the error was corrected in a later version or reprint the date of the correction will be displayed in the column titled "Date Corrected".

The following errata were submitted by our customers and approved as valid errors by the author or editor.

Color key: Serious technical mistake Minor technical mistake Language or formatting error Typo Question Note Update

Version Location Description Submitted By Date submitted Date corrected
Other Digital Version
Ch 1
Subheading: Reproducibility

Under the heading "reproducibility", the 3rd paragraph starts with: "In order to address this problem of **repeatability** ", it should ideally be reproducibility instead of repeatability.

Note from the Author or Editor:
yes, please change "repeatability" to "reproducibility"

 Gargi Balasubramaniam  Nov 23, 2020 
Printed
Page 19
Figure 2-1, formula

In the formula in the figure it states f(x1,x1)=((1 && x2)... where it should be f(x1,x2)=((x1 && x2)...

 Jürgen Plasser  Dec 09, 2020 
Printed
Page 39
All

Pages 39-50 (inc) missing in printed edition

Note from the Author or Editor:
This was a one-off problem, and a replacement was sent to the customer.

 Michael Shearer  Nov 07, 2020  Nov 18, 2020
Printed
Page 54
Figure 2-15

On the left table in Figure 2-15, the label for AD should be - and the label for BD should be +. The left table in Figure 2-15 is not compatible with the conceptual plot on the right.

Note from the Author or Editor:
This is indeed a mistake. As the suggestions notes, in Figure 2-15, in the table on the left, the label for AD should be - and the label for BD should be +.

 Anonymous  Jun 08, 2021 
Printed
Page 92
Figure 3-7

It would be more correct to change Figure 3-6 to a diagram like the one at the top of Sigmoid function's page of Wikipedia (https://en.wikipedia.org/wiki/Sigmoid_function) ?
Since the domain of Sigmoid function is usually [-∞, ∞] and the range is [0, 1].

Note from the Author or Editor:
Yes, the domain of sigmoid is -inf to +inf and the range is from (0,1) exclusive. Also, the y-intercept of the sigmoid function is 0.5. The current figure is a bit misleading with respect to these properties and it would be more clear to as figure similar to the one in Wikipedia as suggested by the reviewer.

Ryoko  May 06, 2021 
Printed
Page 115
1st paragraph, 2nd bullet point

Mixed Input Representation is listed as a solution to a described problem (2nd bullet point. However, I could not find a description of what is Mixed Input Representation in the book.

Note from the Author or Editor:
Change "Mixed Input Representation" to "Multimodal Input"

 Anonymous  Mar 17, 2021 
Printed
Page 133
Table 3-6, 3-7

Table 3-6 and Table 3-7 show the distances between specific data points and the centroid of the four clusters, so should they be following?

Figure 3-6:
CENTROID_ID DISTANCE
--------------------------------
4 0.29998627812137374
1 1.2370167418282159
2 1.376651161584178
3 1.6853517159990536

Figure 3-7:
CENTROID_ID DISTANCE
------------------------------
3 3.061985789261998
4 3.3124603501734966
2 4.330205096751425
1 4.658614918595627

Note from the Author or Editor:
The way the table is currently printed is a mistake. The table should consist of three columns with the column titles CENTROID_ID, NEAREST_CENTROIDS_DISTANCE.CENTROID_ID and DISTANCE.
The table should be as follows:
CENTROID_ID | NEAREST_CENTROIDS_DISTANCE.CENTROID_ID
| NEAREST_CENTROIDS_DISTANCE.DISTANCE
----------------------------------------------------------------
4 | 4 | 0.29998627812137374
| 1 | 1.2370167418282159
| 2 | 1.376651161584178
| 3 | 1.6853517159990536


CENTROID_ID | NEAREST_CENTROIDS_DISTANCE.CENTROID_ID
| NEAREST_CENTROIDS_DISTANCE.DISTANCE
----------------------------------------------------------------
3 | 3 | 3.061985789261998
| 4 | 3.3124603501734966
| 2 | 4.330205096751425
| 1 | 4.658614918595627

 Ryoko  May 06, 2021 
Printed
Page 134
4th paragraph, last line

examples from the *minority* class.
should be
examples from the *majority* class.

Note from the Author or Editor:
This suggestion is correct. The sentence "Additionally, as mentioned by SMOTE, the approach of generating synthetic examples from the minority class is often combined with removing a random sample of examples from the minority class." should instead read: "Additionally, as mentioned by SMOTE, the approach of generating synthetic examples from the minority class is often combined with removing a random sample of examples from the majority class."

 Ryoko  May 06, 2021 
Page 163
footnote

> Jia Deng et al.,"ImageNet: A Large-Scale Hierarchical Image Database," (https://oreil.ly/Wio_D) IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR) (2009): 248.255.
The link is not found.
Should it be replaced with an active link?
https://ieeexplore.ieee.org/document/5206848

 Ryoko  Sep 03, 2021 
Page 286
Bullet items

> - An instance of tfx.orchestration.pipeline (https://oreil.ly/62kf3) where we define our pipeline and the components it includes.
> - An instance of kubeflow_dag_runner (https://oreil.ly/62kf3) from the tfx (https://oreil.ly/62kf3) library. We'll use this to create and run our pipeline.

All three URLs are the same.
If these shortened URLs are developed, I get "https://www.nhc.noaa.gov/aboutsshws.php", but it does not seem to be relevant in this context.
Each of these three URLs should be as follows?
1st URL
->
https://github.com/tensorflow/tfx/blob/master/tfx/orchestration/pipeline.py

2nd URL
->
https://github.com/tensorflow/tfx

3rd URL
->
https://github.com/tensorflow/tfx/blob/master/tfx/orchestration/kubeflow/kubeflow_dag_runner.py

 Ryoko  Sep 03, 2021 
Printed, PDF, ePub,
Page 366
Figure 8-2

Step 8 in the ML Lifecycle diagram should be "Plan for deployment" not "Plan for development"

 Michael Munn
 		 Jan 28, 2021