Abstract:In this paper, we study the lossless analog compression for i.i.d. nonsingular signals via the polarization-based framework. We prove that for nonsingular source, the error probability of maximum a posteriori (MAP) estimation polarizes under the Hadamard transform, which extends the polarization phenomenon to analog domain. Building on this insight, we propose partial Hadamard compression and develop the corresponding analog successive cancellation (SC) decoder. The proposed scheme consists of deterministic measurement matrices and non-iterative reconstruction algorithm, providing benefits in both space and computational complexity. Using the polarization of error probability, we prove that our approach achieves the information-theoretical limit for lossless analog compression developed by Wu and Verdu.
What problem does this paper attempt to address?
### Problems Addressed by the Paper
This paper aims to study the lossless analog compression of non-singular signals through a polarization framework. Specifically, the authors demonstrate that for non-singular source signals, the error probability of the Maximum A Posteriori (MAP) estimation undergoes polarization under the Hadamard transform. Based on this finding, they propose a partial Hadamard compression method and develop a corresponding analog successive cancellation (SC) decoder. This scheme leverages a deterministic measurement matrix and a non-iterative reconstruction algorithm, offering advantages in terms of spatial and computational complexity. Through the polarization phenomenon of error probability, the authors prove that their method achieves the information-theoretic limits of lossless analog compression as proposed by Wu and Verdú.
### Main Contributions
1. **Polarization of Error Probability**: The authors demonstrate that under the Hadamard transform, the error probability of the MAP estimation for non-singular source signals polarizes, extending the polarization phenomenon to the analog domain.
2. **Partial Hadamard Compression and Analog SC Decoder**: Based on the polarization phenomenon of error probability, the authors propose a partial Hadamard compression method and develop a corresponding analog SC decoder, proving that this method can achieve the information-theoretic limits of lossless analog compression.
3. **New Technical Methods**: To analyze the polarization phenomenon in the real number domain, the authors introduce the concept of weighted discrete entropy and develop new technical methods to address the speed of polarization.
### Background and Related Work
- **Lossless Analog Compression**: Wu and Verdú established the fundamental limits of lossless analog compression, stating that for a non-singular source signal \(X\), its information dimension \(d(X)\) determines the minimum number of measurements required for compression.
- **Existing Methods**: Most existing methods rely on random projections and high-complexity decoding algorithms, whereas the method proposed in this paper uses a deterministic measurement matrix and a non-iterative decoding algorithm, offering higher efficiency.
### Methods and Techniques
- **Hadamard Transform**: The Hadamard transform is a linear transform that can convert a signal into a set of new variables whose conditional distributions are either completely determined or completely unpredictable.
- **Weighted Discrete Entropy**: To quantify the contribution of the discrete component of a non-singular distribution to uncertainty, the authors introduce the concept of weighted discrete entropy.
- **Polarization Process**: By analyzing the evolution process of conditional distributions, the authors demonstrate the polarization phenomenon of error probability and provide the convergence speed of polarization.
### Conclusion
By introducing the partial Hadamard compression method and the analog SC decoder, this paper successfully achieves lossless analog compression and reaches the information-theoretic limits. This method not only has significant theoretical implications but also demonstrates high efficiency and feasibility in practical applications.