: The standard model of knowledge in multi-agent systems suffers from what has been called the logical omniscience problem: agents know all tautologies, and know all the logical consequences of their knowledge. For many types of analysis, this turns out not to be a problem. Knowledge is viewed as being ascribed by the system designer to the agents; agents are not assumed to compute their knowledge in any way, nor is it assumed that they can necessarily answer questions based on their knowledge. Nevertheless, in many applications that we are interested in, agents need to act on their knowledge. In such applications, an externally ascribed notion of knowledge is insufficient: clearly an agent can base his actions only on what he explicitly knows. Furthermore, an agent that has to act on his knowledge has to be able to compute this knowledge; we do need to take into account the algorithms available to the agent, as well as the "effort" required to compute knowledge. In this paper, we show...

Abstract. The framework of algorithmic knowledge assumes that agents use deterministic knowledge algorithms to compute the facts they explicitly know. We extend the framework to allow for randomized knowledge algorithms. We then characterize the information provided by a randomized knowledge algorithm when its answers have some probability of being incorrect. We formalize this information in terms of evidence; a randomized knowledge algorithm returning “Yes ” to a query about a fact ϕ provides evidence for ϕ being true. Finally, we discuss the extent to which this evidence can be used as a basis for decisions. 1.

Many computer systems have a functional requirement to release information. Such requirements are an important part of a system’s information security requirements. Current information-flow control techniques are able to reason about permitted information flows, but not required information flows. In this paper, we introduce and explore the specification and enforcement of required information release in a language-based setting. We define semantic security conditions that express both what information a program is required to release, and how an observer is able to learn this information. We also consider the relationship between permitted and required information release, and define bounded release, which provides upper- and lowerbounds on the information a program releases. We show that both required information release and bounded release can be enforced using a security-type system. 1.

Vol. 1 (3:1) 2005, pp. 1–26 www.lmcs-online.org

We investigate the complexity of various combinatorial theorems about linear and partial orders, from the points of view of computability theory and reverse mathematics. We focus in particular on the principles ADS (Ascending or Descending Sequence), which states that every infinite linear order has either an infinite descending sequence or an infinite ascending sequence, and CAC (Chain-AntiChain), which states that every infinite partial order has either an infinite chain or an infinite antichain. It is well-known that Ramsey’s Theorem for pairs (RT2 2) splits into a stable version (SRT22) and a cohesive principle (COH). We show that the same is true of ADS and CAC, and that in their cases these versions are strictly weaker (which is not known to be the case for RT 2 2 and SRT2 2). We also analyze the relationships between these principles and other systems and principles previously studied by reverse mathematics, such as WKL0, DNR, and BΣ2, showing for instance that WKL0 is incomparable with all of the systems we study; and prove computability-theoretic and conservation results for them. Among these results are a strengthening of the fact, proved by Cholak, Jockusch, and Slaman,

The framework of algorithmic knowledge assumes that agents use deterministic knowledge algorithms to compute the facts they explicitly know. We extend the framework to allow for randomized knowledge algorithms. We then characterize the information provided by a ran-domized knowledge algorithm when its answers have some probability of being incorrect. We formalize this information in terms of evidence; a randomized knowledge algorithm returning "Yes " to a query about a fact ~o provides evidence for qo being true. Finally, we discuss the extent to which this evidence can be used as a basis for decisions. 1

Any reasoning agent in a system has a model of the system which represents the partial view available to that agent. We suggest the following notion of knowledge: agent i knows that ff holds if and only if ff is true in the sub-model visible to i. This corresponds to considering knowledge as an agent's ability to answer questions about the subject, and to provide evidence on which the knowledge is based. Simply stated, on facing a query, the agent runs a model checking algorithm on M V and ff, where M V is the agent's view-based model of the system (or evidence available to that agent). We argue that this is one way of getting around the logical omniscience problem, at least when the reasoners are components of distributed computing systems. We also prove a completeness theorem for the logic. 1 Motivation 1.1 Basic logic of knowledge The notion of knowledge has attracted philosophers for centuries. During the last decade, computation theorists have studied this notion extensively wit...

We study the consequences of thruthful communication among agents in a game-theoretic setting. To this end we consider strategic games in the presence of an interaction structure, which specifies groups of players who can communicate their preferences with each other, assuming that initially each player only knows his own preferences. We focus on the outcome of iterated elimination of strictly dominated strategies (IESDS) that can be obtained in any intermediate state of the communication process. The main result of the paper, Theorem 4.2, provides the epistemic characteri-zation of such “intermediate ” IESDS outcomes under the assumption of common knowledge of rationality. To describe the knowledge of the players we adapt the general framework of Apt et al. [3] to reason about preferences. Finally, we de-scribe a distributed program that allows the players to compute the outcome of IESDS in any intermediate state. An initial, short version of this paper appeared as [28]. 1