Keynote Presentation: Symmetries in Software

Presented as part of the 2017 HCSS conference.

ABSTRACT:

A key goal of software abstraction is to provide a simplified interface that is designed to be invariant to the underlying implementation to reduce the complexity of compositional reasoning. As abstraction provides a conserved quantity that is invariant to underlying implementations, it meets the requirement of a symmetry in the physics sense of the word, i.e., a property (the interface) remains conserved under an action (re-implementation).  With this connection made, we can ask the question:  Are there other types of symmetries that can provide other (better?) invariants?   As there is no universal abstraction that is efficient for all compositions, we require an ability to adaptively strengthen the symmetry based on the specific composition that is desired.   In this sense, the adaption of a symmetry through run-time modifications is an application of control theory, which can be described as a mechanism to robustly and efficiently generate dynamic symmetries (abstractions) in the execution of code.  This talk will introduce the concepts of symmetries and relate them to software abstractions and control theory.  Specific examples will be used to illustrate the key concepts.

Evan Fortunato is one of the founding members Apogee Research where he leads the research and development of technology driven solutions to address operational needs in the Defense and Intelligence Communities.  His areas of expertise include information theory, ISR (intelligence, surveillance and reconnaissance) systems, offensive and defensive cyber capabilities and electronic warfare techniques.  Prior to his role at Apogee Research, he served as a program manager in DARPA's Information Innovation Office running a portfolio of cyber programs.  Evan received his B.S. in Physics and his Ph.D. in Nuclear Engineering from the Massachusetts Institute of Technology.