Holographic Transformers for Complex-Valued Signal Processing: Integrating Phase Interference into Self-Attention
This addresses the challenge of coherent signal modeling for applications like radar imaging and wireless communications, representing a novel method for a known bottleneck rather than an incremental improvement.
The paper tackles the problem of deep learning models overlooking phase interference effects when processing complex-valued signals by introducing the Holographic Transformer, which integrates wave interference principles into self-attention. Results show strong performance on PolSAR image classification and wireless channel prediction, achieving high classification accuracy and F1 scores, low regression error, and increased robustness to phase perturbations.
Complex-valued signals encode both amplitude and phase, yet most deep models treat attention as real-valued correlation, overlooking interference effects. We introduce the Holographic Transformer, a physics-inspired architecture that incorporates wave interference principles into self-attention. Holographic attention modulates interactions by relative phase and coherently superimposes values, ensuring consistency between amplitude and phase. A dual-headed decoder simultaneously reconstructs the input and predicts task outputs, preventing phase collapse when losses prioritize magnitude over phase. We demonstrate that holographic attention implements a discrete interference operator and maintains phase consistency under linear mixing. Experiments on PolSAR image classification and wireless channel prediction show strong performance, achieving high classification accuracy and F1 scores, low regression error, and increased robustness to phase perturbations. These results highlight that enforcing physical consistency in attention leads to generalizable improvements in complex-valued learning and provides a unified, physics-based framework for coherent signal modeling. The code is available at https://github.com/EonHao/Holographic-Transformers.