How (and why) to Build an Automated Theorem Prover: De-mystifying Logical Inference

Presentation by Adam Pease at SRI, Menlo Park, CA. I discuss implementation details of writing an automated theorem prover in Java for first order logic. A subsequent video presents questions from the audience and my responses.

[socialwatcher]

Loved this presentation Adam! We can see the effort and care you put into making JavaRes and also explaining it so clearly to the rest of us. Thank you!

[WarriorStatue]

Thank you for this presentation, it has been very useful while going on my own journey of writing an ATP system based on PyRes

[Treebark]

Wonderful presentation, thank you. I'm studying operator theory for an internship, where I'm currently implementing a routine to check the truth value of a particularly involved equivalence relation. This is the first genuinely helpful video resource I've come across.

I can't help but comment at 36:40, as I've never understood partisanship towards certain programming languages. No one is out there claiming Python wins on performance haha, especially against Java. They're different tools for different use cases.

[nathantemple]

Is it possible to contribute to PyRes?

[nrrgrdn]

In your first example with an attractive blue car.
I still dont get why not just direct proof? The contradiction is just a layer around. You ended up concluding attractive(by modus ponens), only discharging your hypothetical at the very end

[luisramos]

I think the term UnitTest is only use in java, not sure about the same name in Python, however we can do test in both.

[Unique-Concepts]

Hello Prof. I want know is there any limitations of automated theorem prover?