List-Decoding Capacity Implies Capacity on the q-ary Symmetric Channel

Francisco Pernice,Oscar Sprumont,Mary Wootters

2024-10-26

Abstract:It is known that the Shannon capacity of the q-ary symmetric channel (qSC) is the same as the list-decoding capacity of an adversarial channel, raising the question of whether there is a formal (and black-box) connection between the two. We show that there is: Any linear code $C\subseteq \mathbb{F}_q^n$ that has minimum distance $d_{\min}=\omega(q^3)$ and achieves list-decoding capacity also achieves capacity on the qSC.

Information Theory

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **Does there exist a formal (and black - box) connection that links the Shannon capacity of the q - ary symmetric channel (qSC) to the list - decoding capacity of the adversarial channel?**

Specifically, the author explores the following issues:

1. **Known background**: The Shannon capacity of the q - ary symmetric channel (qSC) is the same as the list - decoding capacity of the adversarial channel. This raises the question of whether a formal connection can be established between the two.
2. **Research motivation**: In the list - decoding literature, list - decoding is usually regarded as a bridge connecting the Shannon model and the Hamming model, so that the Shannon capacity of the channel can be achieved even in the worst - case (Hamming model). But can this connection be formalized?

### Main contributions of the paper

The main contribution of the paper lies in proving that such a formal connection exists, at least in one direction. Specifically, the author proves:

- **Theorem 1**: For any list - decoding code \(C_n\) with a sufficient minimum distance \(d_{\text{min}}(C_n)=\omega\left(\frac{q^3}{(1 - p)^2}\right)\), if it reaches the list - decoding capacity on the adversarial channel introducing \(p\)-proportion of disruptions, then it also reaches the capacity on qSC.

- **Theorem 2**: For a linear, \((p, L)\)-list - decoding code \(C\subseteq F_q^n\), if its minimum distance \(d_{\text{min}}(C)\geq\frac{q^3}{(1 - p)^2}\ln^4(nL)\), then \(C\) can be used for reliable communication on qSC \(p'\), where \(p' = p-\frac{7}{\ln n}\).

### Technical details

To prove these results, the author uses some key technical tools, including but not limited to:

- **Maximum - likelihood decoder**: A maximum - likelihood decoder \(D^*\) is defined, and its success probability is analyzed.
- **Expectation function**: An expectation function \(f(z)\) is defined, and how its expected value changes with the noise parameter \(p\) is shown.
- **Threshold phenomenon**: Using the inequalities of Margulis and Russo, it is proved that when \(p\) is close to a certain critical value, the probability of successful decoding will rapidly jump from close to 1 to close to 0.

Through these technical means, the author successfully establishes a formal connection between the list - decoding capacity and the qSC capacity, especially in the case of a larger field size \(q\).

### Conclusion

In general, through rigorous mathematical proofs, this paper reveals the internal connection between the list - decoding capacity and the qSC capacity, and provides new perspectives and tools for future coding theory research.