A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Network information theory extends Shannon's point-to-point results to multi-user communication networks. When multiple senders and/or receivers share a channel, new phenomena arise that have no single-user analog: cooperation, interference, distributed compression, and capacity regions (sets of achievable rate tuples rather than single numbers). The capacity of the multiple access channel, broadcast channel, and interference channel are central problems. While some multi-user capacity regions are fully characterized (multiple access channel, degraded broadcast channel), others remain open after 50+ years (general interference channel, relay channel), making network information theory one of the most active areas of information-theoretic research.
Shannon's original theory considers one sender and one receiver. Real communication systems involve many users: cell phones sharing a base station, devices on a Wi-Fi network, satellites communicating with ground stations. Network information theory asks: what are the fundamental limits when multiple communication sessions share the same physical medium?
The simplest multi-user channels illustrate the key ideas. The multiple access channel (MAC) has K senders transmitting independent messages to one receiver. The capacity region is the set of rate tuples (R_1, ..., R_K) such that for every subset S of users, sum_{i in S} R_i <= I(X_S; Y | X_{Sc}). The receiver can decode all messages using successive interference cancellation: decode one user, subtract their signal, decode the next. The broadcast channel (BC) has one sender transmitting different messages to K receivers. The sender uses superposition coding: layering messages at different power levels so that stronger receivers can decode more layers. The interference channel has K sender-receiver pairs sharing the same medium, where each receiver wants only its own message but hears everyone's signal.
The mathematical challenge is that multi-user problems rarely decompose into independent single-user problems. Interference creates coupling between users. Cooperation (through relays or user coordination) can increase capacity beyond what independent links achieve. Distributed source coding (Slepian-Wolf) shows that correlated sources can be compressed to their joint entropy even when the encoders do not communicate. These phenomena — interference, cooperation, and correlation — are fundamentally multi-user and have no single-user analog.
The state of knowledge is uneven. The MAC capacity region is fully known. The degraded broadcast channel capacity is known (superposition coding). The Gaussian interference channel capacity is known in certain regimes (strong interference, very weak interference) but not in general. The relay channel capacity remains open despite being posed by van der Meulen in 1971. Each unsolved problem reveals gaps in our understanding of how information flows through networks — making network information theory one of the richest and most challenging areas of the field.