Dynamic Denoising For Amp Applied To Sparse Regression Inner Codes With Outer Codes
summary
Concise claim: embedding LDPC structure into the AMP denoiser via belief propagation yields measurable finite-length gains versus separate decoding, with state-evolution approximations (Berthier–Montanari–Nguyen) still describing the effective Gaussian noise. Implementation uses dynamic denoiser integrates LDPC via BP and a Hadamard subsampled sensing matrix for fast transforms. Multiuser operation runs as parallel per-user AMP with additive Onsager terms and supports user-centric cell‑free MRC combining; empirical curves show a consistent leftward SNR shift (red curve) at practical block lengths.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 0:22Motivation and outline
- 1:51LDPC code structure
- 5:59Message passing basics
- 8:31Denoising with belief propagation
- 12:45Single-user performance
- 15:57Extending to multiple users
- 19:34Joint multi-user decoding
- 21:45Cell-free architecture
- 24:59Empirical results
- 27:29Conclusion and future work
- 28:57Q&A
related talks
28mCommunication Avoiding Algorithms for Linear Algebra and Beyond
01h 07mIndustry Presentations
11mExploiting Combined Locality for Wide-Stripe Erasure Coding in Distributed Storage
15mDistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language...
48mPersonalized Collaborative Learning with Affinity-Based Variance Reduction
01h 02mConditional Hardness for Massively Parallel Computing (MPC) Via Distributed Lower Bounds
17mParaRC: Embracing Sub-Packetization for Repair Parallelization in MSR-Coded Storage
44mTutorial on List Decodable Coding Schemes and Interactive Communication in Networks (Part 2 of 2)
