Data Workshop

This week I was learning how to train my own car. Thanks to participation in the five-day challenges from the Crown of Machine Learning series, organized by Data Workshop, I could practically see how the process of learning an autonomous car works. Of course, in the simplified version, the scheme of operation is the same, regardless of the degree of complexity, only the input data is different. Curious about what it looks like?

Data is the most important thing. The more data you collect at the beginning, the better results you will get at the end. What data do we need? First of all, photos of the road, the position of the gas pedal, and the steering wheel angle. We obtain initial data from our own car journeys along the route. The principle is simple: the more diverse the route, the more diverse the data, and the greater the chance that the car will cope on a different route. You can also drive the same route from beginning to end and from end to beginning, which also makes a huge difference. For example, a model trained on data from the route only from beginning to end, on another route was not able to drive further than three corners. Mainly because the corners were sharper than in the first route, and the model did not reduce the speed in front of them, so it fell out of the route. The second model, which was created on the basis of the route from beginning to end and from end to beginning, coped better with the second route. It crashed only on a tree, probably because there were no trees on the first route.

With the input data, you generate another portion of data based on it and finally train the model. Python requires only basic code writing skills and a good knowledge of photo editing libraries. In principle, nothing complicated, but you need solid equipment for calculations.

I didn’t do anything else this week. Just me and the autonomous cars for a total of six and a half hours of interesting experience and general knowledge of how it works. Next week,” “What’s the Frontend” launches. I can’t wait. Especially for this occasion, I bought a second monitor to see the effects of work immediately on it. I hope to quickly catch the basics, because the list of websites to be done is constantly growing, and the implementation is missing.

Share the Post:
Cover of book AI in Project Management

What's inside

Explore the transformative power of AI in project management with Krzysztof Nyrek’s “AI in Project Management.” This essential guide offers in-depth knowledge on automating routine tasks, optimizing resource management, and utilizing AI for risk estimation and decision-making. Learn how to implement AI tools for improved project efficiency, cost reduction, and enhanced quality management. Ideal for project managers and IT professionals, this book provides actionable insights and practical examples to harness AI technology effectively. Enhance your project outcomes and stay ahead in the digital era with this indispensable resource.

Embrace the future of project management with AI-driven solutions that save time, reduce costs, and boost project success.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

X