Turn abandoned patents
into your own filing
into your own filing
This invention describes a system for improving a device's speech recognition. It works by comparing the accuracy of the device's current speech recognition result with a result from another device. If the other device's result is more accurate, the system updates the device's speech recognition capabilities using that better result; otherwise, it keeps its current system.
Why it matters: Filed before the widespread adoption of highly efficient, on-device machine learning models and federated learning techniques. Today, the ability to dynamically compare and update speech recognition systems across devices is more practical and powerful due to advancements in distributed AI and model optimization.
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
