“Green codes with short wires at the decoder: Fundamental limits and constructions”

This talk intends to bring out the following ideas: Understanding decoding power is important for design of relevant codes at short distances. Just as we have models for channel to understand transmit power, we need models of decoding implementation to understand decoding power. Node model: total power diverges to infinity as P_e → 0. Optimal transmit power bounded away from Shannon limit. Wire model: design codes for just the desired error-probability in order to avoid paying unnecessarily large wiring costs. Code designs can help reduce wire lengths by operating closer to the tradeoff between performance and wire-lengths.