Hierarchical IoT Data Processing with Function Approximators
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data processing systems in IoT environments face challenges in real-time data collection and processing, particularly with the increasing amount of data and demand for immediate system control, which existing AI configurations like DNNs, reservoir computers, and annealing machines struggle to meet efficiently.
Innovation Solution
An information processing system comprising a parent device and multiple child devices, each equipped with communication interfaces for wireless or wired communication, allowing dynamic adaptation of data processing capacity and content through cooperation among function approximators and annealing machines, enabling flexible and scalable data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing AI configurations (DNN, reservoir computer, annealing machine) are used for data processing, then processing capability is provided, but the system cannot efficiently meet the increasing data amount and real-time processing demand
Solution Approach 1:
The patent segments the data processing system into a parent device and multiple child devices, where child devices perform parallel processing of different data portions. This segmentation enables the system to handle increasing data amounts by adding more child devices while maintaining real-time processing capability through distributed computation.
Solution Approach 2:
The patent transitions from a single-device processing architecture to a multi-device hierarchical architecture, adding the dimension of spatial distribution. The parent device coordinates while child devices process in parallel, creating a three-dimensional processing structure (parent-child relationships, multiple child devices, and data flow paths) that simultaneously increases capacity and reduces processing time.
2Quantity of substance
If processing capacity is increased to handle more data, then data processing amount is improved, but system complexity increases
Solution Approach 1:
The system segments processing functions between a parent device (coordination, data distribution) and child devices (parallel computation). This segmentation allows scaling of processing amount by simply adding child devices without proportionally increasing overall system complexity, as each child device operates independently under parent coordination.
Solution Approach 2:
Child devices are designed with universal functionality to process different types of data and perform various computational tasks. This multi-functionality allows the system to increase processing capacity by adding identical child devices rather than designing specialized components for each function, thereby controlling system complexity while scaling processing capability.
3Speed
If real-time control is implemented across the entire system, then system responsiveness is improved, but communication and coordination overhead increases
Solution Approach 1:
The patent segments control authority between the parent device (overall coordination) and child devices (local execution). Child devices can execute processing tasks independently based on distributed data, reducing the need for constant parent-child communication overhead while maintaining real-time responsiveness through localized decision-making capability.
Solution Approach 2:
The parent device performs preliminary actions by distributing data and processing instructions to child devices before actual computation begins. This preliminary distribution enables child devices to work independently and in parallel, reducing subsequent communication overhead while maintaining real-time control responsiveness through pre-coordinated task allocation.
Data Source
AI summary
According to one embodiment, provided is an information processing system including a parent device and a plurality of child devices. The child device constitutes at least a portion of at least one device selected from a function approximator and an annealing machine, each of the parent devices and the plurality of child devices include a communication interface, and the communication interface is at least one selected from a wireless communication interface and a wired communication interface including an analog circuit. Data to be processed by the child device is transmitted from the parent device to at least one of the plurality of child devices, and an output of at least one node of the child device is transmitted to at least one of the parent device and the other child devices.