Application of spin glass theory to some complex systems

After outlining the essential results of spin glasstheory, the archetype of disordered systems in statistical physics, I'll show how these general concepts apply to systems as diverse as random proteins and heteropolymers, neural networks and learning, and certain Combinatorics optimization problems (the travelling salesman problem, marriage problems, optimal transport, etc.). These systems can be represented as disordered systems, and can then be studied using the method of replicas and the mean field.