US20240281620
2024-08-22
Physics
G06F40/40
A novel approach is introduced for programming new behaviors in computing systems using natural language. This method allows users to interact with software applications through spoken or written commands, enabling them to implement functions that are not pre-programmed. By utilizing natural language understanding, knowledge representation, and advanced compiler techniques, the system can learn and adapt functionalities dynamically, thereby enhancing user experience and reducing reliance on traditional programming methods.
Current natural language processing systems have limitations, primarily that they can only execute functions already coded into them. Users often face difficulties when attempting to create new functionalities, as this typically requires programming knowledge or the use of low-code/no-code tools. These existing solutions often involve a steep learning curve and are heavily dependent on the user's expertise, making it challenging for non-programmers to achieve error-free results.
The proposed method addresses these challenges by enabling users to command software applications to learn new functionalities in real-time. It involves receiving user commands during execution, fetching corresponding logic, and implementing it without requiring prior programming. Additionally, the system can debug previous actions through a natural language interface, allowing for easy examination of past behaviors and adjustments as needed.
During runtime, the software application can identify edge cases or exceptions and solicit human input to resolve them. The system updates its knowledge graph to incorporate new procedures based on user interactions, thereby reducing the need for future human intervention in similar situations. This dynamic learning process enhances the application's ability to adapt and respond to user needs effectively.
The application features a natural language interface that simplifies command input and allows for seamless integration with external knowledge sources. It can also generate a natural language trace of operations executed within the application, providing users with an understandable sequence of actions taken by the system without manual logging. This combination of features significantly improves usability and accessibility for users who may not possess programming skills.