# How to Save and Reuse Skills Learned by Artificial Intelligence Hardware > A method for capturing the internal settings of a neuromorphic AI chip after it learns a task, allowing that 'skill' to be exported and loaded onto another AI chip. - **Patent:** US 10410117 - **Original title:** Method and a system for creating dynamic neural function libraries - **Owner:** BrainChip Inc - **Granted:** 2019 - **Status:** Active - **Times cited:** 1 - **Field:** semiconductors, ai_ml, consumer_electronics ## What it does This patent describes a way to extract the specific configuration of an AI hardware device—specifically a neuromorphic chip—after it has learned a task. Instead of just saving software code, the system captures the 'control values' stored in the chip's synaptic registers, which include parameters like neurotransmitter levels, dendrite delays, and axonal delays. These values effectively act as a snapshot of the device's learned behavior. This snapshot can then be stored in a library and transferred to a second, similar AI device, allowing it to perform the same task without needing to undergo the original training process. ## What it does NOT cover - Does not cover software-based neural networks running on standard CPUs or GPUs. - Does not cover the specific algorithms used to train the initial neural network. - Does not cover the transfer of raw data or training datasets, only the resulting synaptic control values. - Does not cover cloud-based model weight sharing (e.g., standard TensorFlow or PyTorch model exports). ## The clever bit It treats the physical state of a silicon-based neural network (like synaptic delays and neurotransmitter levels) as a portable data object, effectively creating a 'file format' for hardware-level intelligence. ## Real-world examples 1. BrainChip Akida neuromorphic processor 2. Edge AI devices that require pre-trained behaviors 3. Hardware-based pattern recognition systems ## Why it matters This patent addresses a major bottleneck in neuromorphic computing: how to move learned 'intelligence' between physical hardware chips. By treating learned neural states as portable files, it enables a modular approach to AI hardware, where specialized skills can be developed on one device and deployed across a fleet of edge devices without retraining. ## Frequently asked questions ### What does How to Save and Reuse Skills Learned by Artificial Intelligence Hardware cover? A method for capturing the internal settings of a neuromorphic AI chip after it learns a task, allowing that 'skill' to be exported and loaded onto another AI chip. ### Who owns patent US 10410117? BrainChip Inc owns this patent, granted in 2019. ### When does this patent expire? This patent is expected to expire on May 13, 2035, when the invention enters the public domain. ### What is patent US 10410117 cited by? This patent has been cited by 1 later patents that build on its ideas. ### What problem does this patent solve? This patent addresses a major bottleneck in neuromorphic computing: how to move learned 'intelligence' between physical hardware chips. By treating learned neural states as portable files, it enables a modular approach to AI hardware, where specialized skills can be developed on one device and deployed across a fleet of edge devices without retraining. ### What does this patent NOT cover? Does not cover software-based neural networks running on standard CPUs or GPUs. **Full plain-English explainer:** https://patentbrief.org/patent/us/10410117/method-and-a-system-for-creating-dynamic-neural-function-libraries **Original patent:** https://patents.google.com/patent/US10410117 --- _Source: PatentBrief — https://patentbrief.org. Patent facts are from public records; the plain-English explanation is PatentBrief's._ ## Related patents Semantically similar inventions in the PatentBrief corpus: - [How to Automatically Expand Neural Networks by Adding New Nodes](https://patentbrief.org/patent/us/10832138/gpt-language-model-pre-training) — A method for growing artificial intelligence models by identifying underperforming parts of a network and adding new nodes based on the behavior of existing ones. - [How to Fix Faulty Memory Cells in AI Chips](https://patentbrief.org/patent/us/10956815/killing-asymmetric-resistive-processing-units-for-neural-network-training) — This patent describes a system that tests individual memory cells in AI chips for uneven behavior and then permanently disables the faulty ones before the chip starts learning, making AI training more efficient. - [How a Single Electronic Component Can Learn and Process AI Data](https://patentbrief.org/patent/us/10248907/resistive-processing-unit) — This patent describes a tiny electronic component called a resistive processing unit (RPU) that acts like a brain cell in an artificial intelligence network, storing and processing information directly within its changing electrical resistance. - [How a Chip Uses Memory to Speed Up AI Calculations](https://patentbrief.org/patent/us/11741188/hardware-accelerated-discretized-neural-network) — This patent describes a specialized computer chip that uses non-volatile memory and analog signals to quickly perform calculations for artificial intelligence, especially for neural networks that need to remember past information. - [How to Update AI on Small Devices with Slow Internet](https://patentbrief.org/patent/us/20250363357/systems-and-methods-for-deploying-and-updating-neural-networks-at-the-edge-of-a-) — This patent describes a method for efficiently updating artificial intelligence models on small, internet-connected devices, like smart cameras, by sending only the changes, or 'patches,' instead of the entire updated model, which saves bandwidth.