This volume collects a subset of the high-quality papers that were accepted for presentationatthe20thInternationalSymposiumonComputerandInformation Sciences (ISCIS). ISCIS 2005 was organized by the Department of Computer Engineering of Bo? gazi, ci University and was held in Istanbul, Turkey during 26-28October,2005. Following on the success of previous ISCIS symposia, ISCIS 2005 attracted a total of 491 papers. The submitted papers went through a rigorous subm- sion period with the help of the Program Committee and external reviewers. Keeping up with the tradition of previous editions of ISCIS, the selected papers covera wide spectrum of topicsin computer and information sciences.The main topics of ISCIS 2005 can be listed as: computer networks, sensor and satellite networks, securityand cryptography, performanceevaluation, databasesystems, algorithms, multiagentsystems, e-commerce, machinelearning, naturallanguage processing and information retrieval, image and speech processing, and theory of computing. In addition to the paper presentations, the symposium enjoyed four invited talks. The honorary chair of ISCIS 2005, Erol Gelenbe from Imperial College, London, gave a talk on a probability model of natural or arti?cial infection. The second talk was by Onn Shehory of IBM Haifa Research Labs, on online auctions. JacquesStern from EcoleNormaleSup erieure, France, focused his talk on alternatives to RSA. The last invited speaker, Yanghee Choi from Seoul National University, Korea, delivered a talk on location and address issues in ubiquitous comput

The second edition of the workshop on Declarative Agent Languages and Te- nologies (DALT 2004) was held July 2004 in New York City, and was a great success. We saw a signi?cant increase in both the number of submitted papers and workshop attendees from the ?rst meeting, held July 2003 in Melbourne. Nearly 40 research groups worldwide were motivated to contribute to this event by submitting their most recent research achievements, covering a wide variety of the topics listed in the call for papers. More than 30 top researchers agreed to join the Program Committee, which then collectively faced the hard task of selecting the one-day event program. The fact that research in multi-agent systems is no longer only a novel and promising research horizon at dawn is, in our opinion, the main reason behind DALT's (still short) success story. On the one hand, agent theories and app- cations are mature enough to model complex domains and scenarios, and to successfully address a wide range of multifaceted problems, thus creating the urge to make the best use of this expressive and versatile paradigm, and also pro't from all the important results achieved so far. On the other hand, bui- ing multi-agent systems still calls for models and technologies that could ensure system predictability, accommodate ?exibility, heterogeneity and openness, and enable system veri?cation.