Inference vs Generalisations
As my learning has continued on the different types of logic and inference rules that can be used to build knowledge-based AI tools, I've been careful to also consider the limitations of such tools in real-life scenarios. For example, how many of you have heard the general 'rule' that exercise helps reduce your blood pressure? Probably most, if not all of you, right?
The reality though is quite different. Part of the research I did looking at data from thousands of heart patients while at uni proved in fact that exercise actually helps 'correct' blood pressure, rather than just a blanket-rule of reducing it. So while it may be true that those with high blood pressure can reduce this with the help of exercise, people with a lower blood pressure (e.g. those with 'weak hearts' due to illness) can actually have their blood pressure raised towards normal levels through a properly designed exercise program.
In science, there's also another important issue to consider. Unlike universally accepted laws (e.g. thermodynamics and energy conservation in chemistry), scientific opinion can change more often over time. This is especially true in organic and dynamic systems susceptible to change on many levels, such as human physiology and biochemistry. And so even though we currently 'know' that finding A has led to conclusion B, if finding A is subsequently changed with more evidence, conclusion B may still be true. And vice versa. Consider the evidence from the same research relating to the effects of medications on outcomes. So although logic can be straightforward, science is more annoying.
Welcome to the pain that is my brain rn 🙂
Stay awesome,
EMH
HOW MOOVPAD IS BEING BUILT
For the overview of how MOOVPAD apps are being developed, the reasoning behind particular decisions during development, policies, and more in relation to all the technical things, please see the link to the left.
This will be an ongoing work in progress, and will always be linked to the bottom of each upcoming Blog post.