In this experiment we want our AI to find a curve that best fits the experimental data collected during a chemistry experiment.

A student investigated the influence of temperature on the reaction rate of hydrogen peroxide and potassium sodium tartrate with the catalyst cobalt chloride. The student measured reaction times three times each for 30, 40, 50, 60, and 70 degrees celsius and plotted the data in a chart.

• The x-axis represents the temperature in degrees Celsius and the y-axis time in seconds
• The black points represent measured data from the experiment.

Click on the 'Start Learning' Button to start the learning process. 10 red dots will appear, which represent the function that the AI believes fits best to the data points. Observe how the fit function will be way off target in the beginning and how it will get closer and closer to the black experimental data points as learning progresses.

Once learning is complete, drag the temperature slider slowly to the right. This creates data points for temperature values from 0 to 100 degrees celsius. The AI will suggest corresponding reaction times.

By dragging the slider you create all sorts temperature values, most of which the AI has never seen before. The AI will then suggest a reaction time, which could be experimentally verified in the lab. Further experimental data for lower temperatures is required to better train the AI.