Hi, T have very little experience with Logic based Al algorithms, and my understanding is these machines use propositional logic algorithms (which wasn’t something I have come across before, even when I was in uni).
My experience is in statistical & neural based networks and can understand the pros and cons of those networks.
I am just trying to understand what is the advantage and disadvantages of logic based algorithms specifically Tsetlin Machines? Is this something that I should learn more about? What are some good resources?
Thanks
One key difference is that they are not trained with end-to-end optimization but rather a hand crafted learning rule. This rule has strong inductive biases that work well for small datasets with pre-extracted features, like tabular data.
Their big disadvantage (and this applies to logical/symbolic approaches in general) is that they don’t work well with raw data. Even easy datasets like CIFAR10. The world is too messy for perfect logical rules; neural networks are able to capture this complexity, but simpler models struggle to.
Note that learning is a fundamentally statistical process, so Tsetlin Machines are also statistics based.