Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a novel "out-of-air computation" (AirCPU) framework that extracts computation directly from wireless superpositions using structured coding, eliminating the need for pre-embedding computation into transmitted waveforms. AirCPU employs a multi-layer nested lattice architecture for progressive resolution of continuous-valued device data transmitted over a common digital constellation. The framework formalizes decoupled resolution, demonstrating that computation error is primarily determined by the target resolution set by the finest lattice when decoding error probability is sufficiently small, and introduces collective and successive computation mechanisms to expand the reliable operating regime in fading MACs.
Forget pre-embedded computations: AirCPU extracts computations directly from wireless signals using structured coding, unlocking new possibilities for distributed machine learning and sensor networks.
Over-the-air computation (AirComp) has traditionally been built on the principle of pre-embedding computation into transmitted waveforms or on exploiting massive antenna arrays, often requiring the wireless multiple-access channel (MAC) to operate under conditions that approximate an ideal computational medium. This paper introduces a new computation framework, termed out-of-air computation (AirCPU), which establishes a joint source-channel coding foundation in which computation is not embedded before transmission but is instead extracted from the wireless superposition by exploiting structured coding. AirCPU operates directly on continuous-valued device data, avoiding the need for a separate source quantization stage, and employs a multi-layer nested lattice architecture that enables progressive resolution by decomposing each input into hierarchically scaled components, all transmitted over a common bounded digital constellation under a fixed power constraint. We formalize the notion of decoupled resolution, showing that in operating regimes where the decoding error probability is sufficiently small, the impact of channel noise and finite constellation constraints on distortion becomes negligible, and the resulting computation error is primarily determined by the target resolution set by the finest lattice. For fading MACs, we further introduce collective and successive computation mechanisms, in addition to the proposed direct computation, which exploit multiple decoded integer-coefficient functions and side-information functions as structural representations of the wireless superposition to significantly expand the reliable operating regime; in this context, we formulate and characterize the underlying reliability conditions and integer optimization problems, and develop a structured low-complexity two-group approximation to address them.