Core Viewpoint - The article discusses a novel approach to machine learning deployment at the edge, introducing a disaggregated computing model that utilizes radio frequency (RF) for computation, thereby addressing bandwidth and privacy issues associated with traditional cloud-based models [5][6][11]. Group 1: Traditional Approaches - Traditional machine learning inference methods involve either uploading model inputs to the cloud for processing or broadcasting models to edge devices, both of which have significant drawbacks such as bandwidth consumption and privacy concerns [5][6][7]. Group 2: Disaggregated Computing Model - The proposed disaggregated computing model broadcasts the machine learning model via RF, allowing edge devices to modulate inputs onto RF signals, with all computations performed in the RF domain using frequency mixers [8][11][14]. - This model eliminates the need for local storage of models on edge devices, reducing storage overhead and energy consumption [11][30]. Group 3: Experimental Validation - Experiments were conducted using a software-defined radio testbed, demonstrating the feasibility of broadcasting models to multiple edge devices, achieving a maximum vector inner product of 32,768 points with energy consumption at the femtojoule level, significantly lower than traditional digital computations [17][23][27]. Group 4: Performance Metrics - In tests on the MNIST dataset, a single-layer model achieved 95.7% accuracy with an energy consumption of 6.03 fJ/MAC, while a three-layer model maintained 98.1% accuracy using traditional methods [27]. - For the AudioMNIST dataset, the proposed method achieved 97.2% accuracy with energy consumption reduced to 2.8 fJ/MAC [28]. Group 5: Innovations and Implications - Key innovations include the ability to broadcast models wirelessly for simultaneous inference across multiple edge devices, and the use of existing RF components to perform computations without additional hardware modifications [29][30][31]. - The approach allows for scalable neural network computations, supporting modern deep learning model requirements without the need for specialized AI chips [31].

Core Viewpoint - The company, Aiguang Technology (300123.SZ), is involved in satellite production line construction through its subsidiary Chengdu Aiguang, although the current order amount is relatively small [1] Group 1: Company Involvement - Chengdu Aiguang is one of the suppliers for the satellite production line construction [1] - The company has relevant products in the satellite field that can be supplied, including low-noise amplifiers, mixers, power dividers, filters, and signal detection circuits [1] - Additionally, composite functional microwave module components, such as TR components, can also be supplied [1]

Core Viewpoint - The company Aiguang Technology is actively involved in the development and sales of self-researched MMIC chips, which are utilized in various applications, including satellite communication [1] Group 1: Product Development and Sales - The MMIC chips developed by the company not only meet internal product demands but are also sold externally to a wide range of customers [1] - The company offers various supporting products for satellite internet and communication applications, including low-noise amplifiers, mixers, power dividers, filters, and signal detection circuits [1] Group 2: Market Applications and Future Plans - The company's products have already been applied in certain satellite constellation construction projects, such as the GW constellation plan [1] - The company plans to continue aligning with the trends in satellite communication and low Earth orbit constellation development, actively addressing customer needs and expanding its market share in related fields [1]

Core Insights - The company, Yaguang Technology, has indicated its capabilities in the satellite internet and communications sectors, specifically through its microwave channel products [2] Group 1: Product Offerings - The company offers a range of microwave channel products that can support applications in satellite internet and communications, including low-noise amplifiers, mixers, power dividers, filters, and signal detection circuits [2] - Additionally, the company provides composite functional microwave module components, such as TR components, which can also be utilized in these sectors [2]