2017/05/31 by Susanne Still · 41 citations
Mathematics · Neuroscience · Physics and Astronomy · #Advanced Thermodynamics and Statistical Mechanics #Artificial intelligence #Computer science #Connection (principal bundle) #Dissipation #Efficient energy use #Inference #Limit (mathematics) #Mathematical analysis #Mathematics #Neural dynamics and brain function #Observable #Physics #Quantum mechanics #Representation (politics) #Statistical Mechanics and Entropy #Statistical physics #Thermodynamics #Upper and lower bounds #cond-mat.stat-mech
paper · pdf · doi:10.1103/physrevlett.124.050601
published in Physical Review Letters 124(5), 050601 (American Physical Society) · Accepted for publication in Physical Review Letters
arxiv created 2019/10/03 · openalex publication_date 2020/02/06 · arxiv updated 2020/02/12 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
This Letter exposes a tight connection between the thermodynamic efficiency of information processing and predictive inference. A generalized lower bound on dissipation is derived for partially observable information engines which are allowed to use temperature differences. It is shown that the retention of irrelevant information limits efficiency. A data representation method is derived from optimizing a fundamental physical limit to information processing: minimizing the lower bound on dissipation leads to a compression method that maximally retains relevant, predictive, information. In that sense, predictive inference emerges as the strategy that least precludes energy efficiency.