Universal Semantic Communication II: A Theory of Goal-Oriented Communication

We continue the investigation of the task of meaningful communication among intelligent entities (players, agents) without any prior common language. Our generic thesis is that such communication is feasible provided the goals of the communicating players are verifiable and compatible. In a previous work we gave supporting evidence for one specific goal — where one of the players wished to solve a hard computational problem and communicated with the other in the hope of finding a solution. In this work we initiate a “generic” study of the goals of communication. We present two definitions: one of a “generic” meta-goal, which captures the (potentially unrealizable) wishes of communicating agents, and the other being a “generic” syntactic goal, which captures effects that can observed by an agent. We then show, under some technical conditions, that those metagoals that have corresponding syntactic versions are also universally achievable, i.e., achievable when the two communicators do not (necessarily) share a common language. We also show how our formalism captures a variety of commonplace examples of goals of communication, including simple control-oriented goals that aim to effect a remote physical action by communication, as well as more subtle intellectual goals where the communicator’s intent is mostly to gain knowledge. Our examples from the latter class include a variety of settings where meta-goals differ significantly from syntactic goals.

[1]  A. M. Turing,et al.  Computing Machinery and Intelligence , 1950, The Philosophy of Artificial Intelligence.

[2]  Richard E. Ladner,et al.  Probabilistic Game Automata , 1988, J. Comput. Syst. Sci..

[3]  Lance Fortnow,et al.  Low-Depth Witnesses are Easy to Find , 2007, Twenty-Second Annual IEEE Conference on Computational Complexity (CCC'07).

[4]  Mihir Bellare,et al.  The Complexity of Decision Versus Search , 1991, SIAM J. Comput..

[5]  Nikhil R. Devanur,et al.  A computational theory of awareness and decision making , 2009, TARK '09.

[6]  Anne Condon Space-bounded probabilistic game automata , 1991, JACM.

[7]  N. Block Psychologism and Behaviorism , 1981 .

[8]  Peter Norvig,et al.  Artificial Intelligence: A Modern Approach , 1995 .

[9]  Willard Van Orman Quine,et al.  Word and Object , 1960 .

[10]  Lance Fortnow,et al.  Computational depth: Concept and applications , 2006, Theor. Comput. Sci..

[11]  STUART M. SHIEBER The Turing Test as Interactive Proof , 2007 .

[12]  Dr. Marcus Hutter,et al.  Universal artificial intelligence , 2004 .

[13]  Carsten Lund,et al.  Algebraic methods for interactive proof systems , 1992, JACM.

[14]  Adi Shamir,et al.  IP = PSPACE , 1992, JACM.

[15]  Richard E. Ladner,et al.  Probabilistic Game Automata , 1986, J. Comput. Syst. Sci..

[16]  Brendan A. Juba,et al.  Universal semantic communication I , 2008, Electron. Colloquium Comput. Complex..