信息论：2025–2026 年度综述

An annual survey of the cs.IT arXiv category, covering research published from September 2025 – August 2026. Built on a close-reading set of 132 papers and organised into 37 chapters, it maps the year by the field's research sub-directions.

Book framework (share of the year's output):

• Deep Learning for Wireless Communications (15.8%) — Integrates deep reinforcement learning and generative models with federated learning to optimize precoding, resource allocation, and channel modeling in complex wireless networks.
• Algebraic Coding Theory and Decoding (14.2%) — Develops new constructions and decoding algorithms for linear, cyclic, and Reed-Solomon codes, focusing on error correction capabilities and algebraic structures.
• Statistical Inference and Complexity Bounds (8.7%) — Establishes optimal bounds and complexity limits for data sampling, hypothesis testing, and privacy-preserving estimation in high-dimensional statistical models.
• Joint Source-Channel Coding and Estimation (8.1%) — Explores joint source-channel coding strategies and advanced channel estimation techniques, including deep learning-based methods for MIMO and RIS systems.
• Reconfigurable and Movable Antenna Systems (8.0%) — Investigates optimization techniques for fluid, movable, and reconfigurable antenna systems to enhance coverage, beamforming, and energy efficiency in wireless networks.
• Information-Theoretic Foundations and Geometry (6.9%) — Applies information geometry and fundamental information inequalities to analyze theoretical limits, complexity measures, and thermodynamic aspects of information systems.
• Rate-Distortion and Semantic Communication (6.3%) — Analyzes rate-distortion trade-offs and semantic communication frameworks, focusing on compression efficiency and distortion mitigation in finite-blocklength regimes.
• Quantum Coding and Channel Capacity (5.7%) — Examines quantum LDPC codes, entanglement-assisted communication, and capacity bounds for quantum channels, including decoding algorithms for quantum error correction.
• Massive MIMO and Cell-Free Systems (5.0%) — Optimizes beamforming, channel acquisition, and interference management in massive MIMO and cell-free architectures for next-generation wireless networks.
• Entropy Methods and Random Access (3.8%) — Utilizes maximum entropy principles and information measures to analyze random access protocols, multiplexing strategies, and distributional properties of random variables.
• Machine Learning for Intelligent Wireless Networks (17.6%) — Uses deep reinforcement learning, generative models, and federated learning to optimise precoding, resource allocation, and channel modelling in wireless networks.

Each sub-direction is covered across three chapters — how it framed its questions, the strength of its key findings, and its open debates and frontiers — so the reader gets a traceable picture of what this part of the field actually produced over the year.