Approximate minimum path-metric conjecture for SCL decoding with list size at least 8

About 3 years old · traced to

Let L=2mL=2^m with m≥3m\ge 3, and let Lk−m{{\mathsf L}_{k-m}} denote the log-likelihood ratio at bit position k−mk-m. For each vk−m+1k∈{0,1}m{\bf v}_{k-m+1}^k\in\{0,1\}^m, let Pvk−m=1, vk−m+1kP_{v_{k-m}=1,\,{\bf v}_{k-m+1}^k} be the path metric of the path beginning with vk−m=1v_{k-m}=1 and continuing with vk−m+1k{\bf v}_{k-m+1}^k, and let P1P_1 be the corresponding aggregate path metric. Approximate minimum path-metric conjecture. When Lk−m≥0{{\mathsf L}_{k-m}}\ge 0, we assume

min⁡vk−m+1k∈{0,1}m{Pvk−m=1, vk−m+1k}≈P1L.\min_{{\bf v}_{k-m+1}^k\in\{0,1\}^{m}}\left\{P_{v_{k-m}=1,\,{\bf v}_{k-m+1}^k}\right\}\approx\frac{P_1}{L}.

This extends the preceding approximation to successive cancellation list decoding with L=2m≥8L=2^m\ge 8 and is used to estimate decoding performance. The supplied text gives no evidence that the conjecture has been proved or disproved.

References

Primary source

Jinnan Piao, Dong Li, Xueting Yu, Zhibo Li, Ming Yang, Jindi Liu and Peng Zeng, “Performance Analysis for Polar Codes under Successive Cancellation List Decoding with Fixed List Size”, arXiv:2306.17496 (2023).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.