Systems and methods for modeling probability distributions

This invention describes a way to build a sophisticated artificial intelligence model called a Deep Boltzmann Machine. It does this by training a simpler model, a Restricted Boltzmann Machine, using a special combination of two learning techniques: one that tries to make the model's outputs look very much like the original data, and another that pits the model against an 'adversary' to make its outputs even more realistic. The initial model is specifically designed to handle many different kinds of data at once through a composite input layer.

Why it matters: Filed before the explosion of generative AI, this method's hybrid approach to training complex generative models using adversarial techniques might offer new stability or efficiency for diverse data types. The focus on Deep Boltzmann Machines and mixed data types aligns with current challenges in building robust generative AI.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedSoftware logic, AI/ML expertise, MCMC

3 AI reinvention directions for this patent

See all three

AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.

Reinvent this with AI

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

This one

+10