Evgenia Christoforou

Applying the dynamics of evolution to achieve reliability in master–worker computing

We consider Internet‐based master–worker task computations, such as SETI@home, where a master process sends tasks, across the Internet, to worker processes; workers execute and report back some result. However, these workers are not trustworthy, and it might be at their best interest to report incorrect results. In such master–worker computations, the behavior and the best interest of the workers might change over time. We model such computations using evolutionary dynamics, and we study the conditions under which the master can reliably obtain task results. In particular, we develop and analyze an algorithmic mechanism based on reinforcement learning to provide workers with the necessary incentives to eventually become truthful. Our analysis identifies the conditions under which truthful behavior can be ensured and bounds the expected convergence time to that behavior. The analysis is complemented with illustrative simulations. Copyright

Achieving reliability in master-worker computing via evolutionary dynamics

This work considers Internet-based task computations in which a master process assigns tasks, over the Internet, to rational workers and collect their responses. The objective is for the master to obtain the correct task outcomes. For this purpose we formulate and study the dynamics of evolution of Internet-based master-worker computations through reinforcement learning.

A Mechanism for Fair Distribution of Resources without Payments

We design a mechanism for Fair and Efficient Distribution of Resources (FEDoR) in the presence of strategic agents. We consider a multiple-instances, Bayesian setting, where in each round the preference of an agent over the set of resources is a private information. We assume that in each of r rounds n agents are competing for k non-identical indivisible goods, (n > k). In each round the strategic agents declare how much they value receiving any of the goods in the specific round. The agent declaring the highest valuation receives the good with the highest value, the agent with the second highest valuation receives the second highest valued good, etc. Hence we assume a decision function that assigns goods to agents based on their valuations. The novelty of the mechanism is that no payment scheme is required to achieve truthfulness in a setting with rational/strategic agents. The FEDoR mechanism takes advantage of the repeated nature of the framework, and through a statistical test is able to punish the misreporting agents and be fair, truthful, and socially efficient. FEDoR is fair in the sense that, in expectation over the course of the rounds, all agents will receive the same good the same amount of times. FEDoR is an eligible candidate for applications that require fair distribution of resources over time. For example, equal share of bandwidth for nodes through the same point of access. But further on, FEDoR can be applied in less trivial settings like sponsored search, where payment is necessary and can be given in the form of a flat participation fee. FEDoR can be a good candidate in a setting like that to solve the problem of starvation of publicity slots for some advertisers that have a difficult time determining their true valuations. To this extent we perform a comparison with traditional mechanisms applied to sponsored search, presenting the advantage of FEDoR.

Internet Computing: Using Reputation to Select Workers from a Pool

The assignment and execution of tasks over the Internet is an inexpensive solution in contrast with supercomputers. We consider an Internet-based Master-Worker task computing approach, such as SETI@home. A master process sends tasks, across the Internet, to worker processors. Workers execute, and report back a result. Unfortunately, the disadvantage of this approach is the unreliable nature of the worker processes. Through different studies, workers have been categorized as either malicious (always report an incorrect result), altruistic (always report a correct result), or rational (report whatever result maximizes their benefit). We develop a reputation-based mechanism that guarantees that, eventually, the master will always be receiving the correct task result. We model the behavior of the rational workers through reinforcement learning, and we present three different reputation types to choose, for each computational round, the most reputable from a pool of workers. As workers are not always available, we enhance our reputation scheme to select the most responsive workers. We prove sufficient conditions for eventual correctness under the different reputation types. Our analysis is complemented by simulations exploring various scenarios. Our simulation results expose interesting trade-offs among the different reputation types, workers availability, and cost.

Brief announcement: achieving reliability in master-worker computing via evolutionary dynamics

This work considers Internet-based task computations in which a master process assigns tasks, over the Internet, to rational workers and collect their responses. The objective is for the master to obtain the correct task outcomes. For this purpose we formulate and study the dynamics of evolution of Internet-based master-worker computations through reinforcement learning.