Get startedGet started for free

Concluding remarks

1. Concluding remarks

Congratulations for completing this challenging and broad course. Let us review the incredible progress you have made.

2. Concluding remarks

You started by refreshing your knowledge of the standard supervised learning workflow, including feature engineering, model fitting and model selection, all with an eye to the ever-present danger of overfitting. You then took a step back to better understand the process by which a problem gets translated to a supervised learning task. You considered data fusion as a way to produce training examples out of disparate data sources, and understood how to make the best of noisy labels and heuristic expert knowledge. You gained a solid hands-on understanding of the critical importance of loss functions, and making sure that your choice of loss reflects the real-world costs of false positives and false negatives.

3. Concluding remarks

You then stepped outside the comfort zone of supervised learning and grappled with anomaly and novelty detection. A lack of labels will stop most data scientists - but not you! You now know from first-hand experience that it is possible to get fantastic accuracy using detectors that are trained without any actual labels, as long as you pick your distance metric carefully, even with entirely unstructured data. This course has given you machine learning superpowers that will help you stand out, and exposed you to a number of fascinating real-world problems and real-world data ranging from cybersecurity to retail banking and healthcare.

4. Congratulations!

In your next job interview, try mentioning that you now know how to extract value from data even with noisy labels, or without any labels at all; or explaining that you are familiar with agile model development - and see how this battle-tested knowledge will help you get ahead of other candidates. Best of luck, and congratulations again on this achievement!

Create Your Free Account

or

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.