HML: Lesson - Teaching a Machine

An illustration representing a robot machine learning from the data surrounding it.Teaching a Machine

Think About It...

As we continue exploring machine learning, ask yourself a few questions:

  • Have you ever taught another person - a skill or academic lesson?
  • What did you learn through the process of teaching?

What Does it Mean to Teach a Machine

When teaching a machine there are various things to consider. In this lesson, we will assess the importance of data amounts, feedback, and patience.

Amount and Quality of Data

Having a large amount of data is important for teaching a machine because it helps it learn more accurately. As stated earlier it is essential to consider where this data comes from.

Sometimes the data might not represent everyone equally. For example, if we’re teaching a machine to recognize faces, but most of the pictures are of one type of person, it might not be able to recognize other types of faces as well. So, we need to make sure our data is diverse and represents everyone fairly.

Feedback

Giving feedback to the machine helps it improve, but we also need to be careful about the kind of feedback we give. It is imperative that the feedback is helpful and doesn’t reinforce any biases. For example, if the machine consistently misidentifies certain objects based on stereotypes, giving it feedback that reinforces those stereotypes wouldn’t be fair or accurate. Feedback should be used to help machines learn in a fair and unbiased way.

Patience

Patience is important for machines to learn, for teachers of those machines, and the creators of training sets and models. Time should be taken to train the machine properly and ensure it’s learning in an ethical way. Rushing the process could lead to mistakes or biases that could harm people. Instead, we need to be patient and thorough in our teaching and learning process.

Notes:

  • Training sets are a collection of examples used to train machine learning models.
  • Machine learning models are computer programs that learn from examples in the training set to recognize patters in data and/or make predictions.
  • Test sets are used to assess how well the model performs.

Conclusion

Overall, when teaching a machine, we need to consider not only how much data it has, the feedback it receives, and the time it takes to learn but also the ethical implications of our teaching methods. We want to create machines that are not only intelligent but also fair and just.

With machine learning there are so many things to date that we could explore, but we simply do not have the time to cover everything in this lesson such as neural networks and deep learning. The point of the lesson was to investigate how machine learning operates, formulate a perspective on AI’s societal consequences, and address societal impacts of AI. You should now be able to evidence an awareness and understanding of some of the basic ideas and concepts of machine learning as well as AI ethics . It is our hope that information presented in this lesson ignites a desire to learn more, dig deeper, and create an environment in which machine learning makes a positive impact for all.

[CC BY-NC-SA 4.0 Links to an external site.] UNLESS OTHERWISE NOTED | IMAGES: LICENSED AND USED ACCORDING TO TERMS OF SUBSCRIPTION - INTENDED ONLY FOR USE WITHIN LESSON.
graphicwithart/Shutterstock.com. Image used under license from Shutterstock.com and may not be repurposed.