Approximate and Probabilistic Computing : Design , Coding , Verification

Computing has entered the era of approximation, in which hardware and software generate and reason about estimates. Navigation applications turn maps and location estimates from hardware GPS sensors into driving directions; speech recognition turns an analog signal into a likely sentence; search turns queries into information; network protocols deliver unreliable messages; and recent advances promise that approximate hardware and software will trade result quality for energy efficiency. Millions of people already use software which computes with and reasons about approximate/probabilistic data daily. These complex systems require sophisticated algorithms to deliver accurate answers quickly, at scale, and with energy efficiency, and approximation is often the only way to meet these competing goals. Despite their ubiquity, economic significance, and societal impact, building such applications is difficult and requires expertise across the system stack, in addition to statistics and applicationspecific domain knowledge. Non-expert developers need tools and expertise to help them design, code, and verify these complex systems. The aim of this seminar was to bring together academic and industrial researchers from the areas of probabilistic model checking, quantitative software analysis, probabilistic programming, and approximate computing to share their recent progress, identify challenges in computing with estimates, and foster collaboration with the goal of helping non-expert developers design, code, and verify modern approximate and probabilistic systems. Seminar November 29–4, 2015 – http://www.dagstuhl.de/15491 1998 ACM Subject Classification D.2.4 Software/Program Verification, D.3.3 Language Constructs and Features, F.3.2 Semantics of Programming Languages, F.1.2 Modes of Computation