Deep Learning Interface for Cross-Device Application Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for integrating applications across different end devices are cumbersome, requiring costly and time-consuming processes, as they necessitate porting and individual solutions for each device, and lack a unified interface for functionalities like voice control and augmented reality.
Innovation Solution
A method and system that converts requests into meta-text and uses deep learning to classify and execute them across various devices, enabling a unified interface and dynamic application management without direct communication between presentation and application logic, allowing for seamless integration and output in suitable formats.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If applications are ported to each corresponding hardware and software system individually, then device-specific functionality is achieved, but development time and cost increase significantly
Solution Approach 1:
The patent implements a universal natural language interface that can process requests from multiple different terminal devices (smartphones, tablets, PCs, wearables, IoT devices) through a single unified system. The deep learning model translates various device-specific requests into a common metatext format, eliminating the need to develop separate interfaces for each device type while maintaining full adaptability to device-specific functionalities like voice control, gesture control, and augmented reality
Solution Approach 2:
The patent introduces a natural language processing system as an intermediary layer between terminal devices and application logic. This intermediary converts device-specific requests into a standardized metatext format that the deep learning model can process, then translates the model's responses back into device-appropriate formats. This mediator approach allows seamless communication across different devices without requiring individual porting of applications
2Adaptability or versatility
If custom interfaces are created for new devices, then integration capability is improved, but development effort and cost increase
Solution Approach 1:
The system provides a universal interface that handles multiple device types and functionalities through a single natural language processing model. The model can process requests from devices with different input methods (voice, text, gestures, AR) and output to various device formats, eliminating the need to create custom interfaces for each new device while maintaining full integration capability
Solution Approach 2:
The deep learning model automatically adapts to new device types and request formats through its training on diverse datasets. The system self-improves by learning from anonymized previous requests and user feedback, reducing the need for manual interface development. When new devices are introduced, the model can be retrained with minimal effort to recognize and process their specific request patterns
3Adaptability or versatility
If multiple interface solutions are implemented for different devices, then device compatibility is achieved, but system complexity increases
Solution Approach 1:
The patent merges multiple device-specific interface solutions into a single unified natural language processing system. Instead of maintaining separate interfaces for different devices, the system combines them all into one deep learning model that processes requests from any device through a common metatext format, significantly reducing system complexity while maintaining full device compatibility
Solution Approach 2:
The natural language processing system serves as a mediating layer that simplifies communication between diverse terminal devices and application logic. The intermediary converts various device-specific protocols and input methods into a standardized internal format, reducing the complexity of direct device-to-application connections while maintaining full compatibility across all device types
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
Method for forming a digital interface for processing a request between at least one terminal device and at least one application logic, wherein the method comprises: a conversion step, a classification step, an execution step, and an output step that produces an output on at least one terminal device; and wherein the classification step is performed based on a deep learning process in which a request-specific metatext and other metadata are evaluated.