4/21/24. Building AI Models with Tensorflow
Frank L -
In the past few weeks, I’ve helped run some new and improved GAN models that are much more consistent with producing good results. Overall, the GAN project that I’ve been working with is gradually coming to a close. In the meantime, I have also started using TensorFlow to start building some really basic machine learning models. And by basic, I mean extremely basic.
One of the first models that I learned how to create was a linear regression model, which essentially does the equivalent of plotting data points, drawing a line of best fit, and using that line of best fit to make predictions. What makes this model really neat is that I can train it on multiple variables at a time. I could use training data that has 10 different variables and ask it analyze the relationship between all of them. Another model I learned how to create was a classification model that can predict what group an object belongs to based on its attributes. Both of these models are relatively simple, but are incredibly useful. I’m pretty sure that advertising companies use AI built on these tools to match ads to people who are most likely going to be interested in specific products or services.
Below, I’ve put one of the results I got from a linear regression model I built to predict survival probabilities for passengers that were on the Titanic. The number at the bottom shows the predicted survival probability and the number above that represents whether they survived or not. 1 means they survived and 0 means they didn’t.
Now you might be wondering why I chose this dataset specifically. The answer is that there isn’t a reason, I just used the exact same data from the tutorial that I watched.
Comments:
All viewpoints are welcome but profane, threatening, disrespectful, or harassing comments will not be tolerated and are subject to moderation up to, and including, full deletion.