Multi-node neural network constructed from pre-trained small networks

This invention describes a way to build a very large artificial intelligence (AI) network by combining many smaller, already-trained AI networks. It works by first making each small network bigger, but mostly empty, and then layering these bigger, empty networks on top of each other in a two-step process to form the final large network. The method also allows for adapting the large network for new tasks by swapping out some of its smaller component networks.

Why it matters: Filed just as large language models became widely accessible, this invention's focus on combining and adapting pre-trained networks, especially through sparse representations, directly addresses current challenges in making massive AI models more efficient and modular for diverse applications.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedHeavy software system, AI expertise

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