Back to results
US 12,762,350 B1Granted
Neural network based gaming assistant
AssigneeNVIDIA Corporation
InventorsPrakash Gurumurthy, Divyansh Jain, Yan Breek
Life of this patent
Estimating the term…
Abstract
Apparatuses, systems, and techniques to generate gaming strategies using neural networks. In at least one embodiment, one or more neural networks are used to generate one or more gaming strategy outputs based, at least in part, on one or more text inputs describing one or more multimedia gaming contents.
Claims20
A processor comprising:
- one or more circuits to:
- generate, by a first neural network encoder, from a reference gameplay video, a reference gameplay frame embedding, and from a user gameplay video, a user gameplay frame embedding;
- generate, by a second neural network encoder, from a first set of control commands associated with the reference gameplay video, a reference control embedding, and from a second set of control commands associated with the user gameplay video, a user control embedding;
- transform, by a text generation neural network, a combination of at least the user gameplay frame embedding, the user control embedding, the reference gameplay frame embedding, and the reference control embedding into one or more text inputs describing one or more multimedia gaming contents;
- and
- generate, by a neural network language model, one or more gaming strategy outputs based, at least in part, on the one or more text inputs describing the one or more multimedia gaming contents.
3 more elements
A system comprising:
- a memory storing a plurality of processor-executed instructions to operate one or more neural networks;
- and
- one or more processors executing the plurality of processor-executed instructions to:
- generate, by a first neural network encoder, from a reference gameplay video, a reference gameplay frame embedding, and from a user gameplay video, a user gameplay frame embedding;
- generate, by a second neural network encoder, from a first set of control commands associated with the reference gameplay video, a reference control embedding, and from a second set of control commands associated with the user gameplay video, a user control embedding;
- transform, by a text generation neural network, a combination of at least the user gameplay frame embedding, the user control embedding, the reference gameplay frame embedding, and the reference control embedding into one or more text inputs describing one or more multimedia gaming contents;
- and
- generate, by a neural network language model, one or more gaming strategy outputs based, at least in part, on the one or more text inputs describing the one or more multimedia gaming contents.
5 more elements
A machine-readable medium having stored thereon a set of instructions, which if performed by one or more processors, cause the one or more processors to at least:
- generate, by one or more neural network encoders, from a reference gameplay video, a reference gameplay frame embedding, and from a user gameplay video, a user gameplay frame embedding;
- generate, by the one or more neural network encoders, from a first set of control commands associated with the reference gameplay video, a reference control embedding, and from a second set of control commands associated with the user gameplay video, a user control embedding;
- transform, by a text generation neural network, a combination of at least the user gameplay frame embedding, the user control embedding, the reference gameplay frame embedding, and the reference control embedding into one or more text inputs describing one or more multimedia gaming contents;
- and
- generate, by a neural network language model, one or more gaming strategy outputs based, at least in part, on the one or more text inputs describing the one or more multimedia gaming contents.
2 more elements
Dependencies are inferred from the claim text.