Austin, Texas, the "live music capital For more information about AAAI is pleased to announce the continued Conferences/conferences.html. Expository Writing Award will be presented members. AAAI is delighted to announce the 31-August 3 in Austin, Texas. This The conference will be held July collocation of SARA-2000 with AAAIaward joins the two special awards 31-August 3, 2000, at the Austin Convention 2000. The Symposium on Abstraction, established last year, the AAAI Classic Center and Hyatt Regency Reformulation, and Approximation Paper Award and the AAAI Distinguished Austin in Austin, Texas. AAAI-2000 will be held July 26-29, just outside Austin in Lago Vista on Lake Travis, Service Award. For more information about The AAAI Effective Expository Writing the Innovative Applications of SARA-2000, please visit sara2000.unl. Award honors the author(s) of a Artificial Intelligence, the Mobile edu/ high-quality, effective piece of writing, Robot Competition and Exhibition, AAAI also welcomes SARA-2000 as accessible to the general public or the Intelligent Systems Demonstrations, our first affiliate conference. For more to a broad AI audience (not just a subarea), the Robot Building Laboratory, information about the AAAI Affiliates written within the last two and the Doctoral Consortium. New Program, please write to Carol Hamilton years. The contribution should be for 2000 will be a technical paper at hamilton@aaai.org. Nominated papers must be Uncertainty: Operations Research AAAI is pleased to announce the continuation in English and must have been published Meets AI (Again)"; Justine Cassell, of its Student Abstract and in a publicly accessible place "Why Do We Need a Body Anyway?"; Poster Program, the SIGART/AAAI (for example, periodical, hard copy, or Carla Gomes, "Structure, Duality, and Doctoral Consortium, and the AAAI online journal but not only as a web Randomization: Common Themes in Scholarship and Volunteer Programs. The author(s) AI and OR"; James Hendler, "Missed Students interested in attending the of the award-winning paper(s) will Perceptions: AI versus the Funding National Conference on Artificial receive a $2500 prize (shared if more Agencies"; Geoff Hinton, "Modeling Intelligence in Austin, July 31-August than one author) as well as lodging High-Dimensional Data Distributions 3, 2000, should consult the AAAI web and travel to the National Conference by Combining Simple Experts"; Rich site for further information about all on Artificial Intelligence.

Austin, Texas, the "live music capital For more information about AAAI is pleased to announce the continued Conferences/conferences.html. Expository Writing Award will be presented members. AAAI is delighted to announce the 31-August 3 in Austin, Texas. This The conference will be held July collocation of SARA-2000 with AAAIaward joins the two special awards 31-August 3, 2000, at the Austin Convention 2000. The Symposium on Abstraction, established last year, the AAAI Classic Center and Hyatt Regency Reformulation, and Approximation Paper Award and the AAAI Distinguished Austin in Austin, Texas. AAAI-2000 will be held July 26-29, just outside Austin in Lago Vista on Lake Travis, Service Award. For more information about The AAAI Effective Expository Writing the Innovative Applications of SARA-2000, please visit sara2000.unl. Award honors the author(s) of a Artificial Intelligence, the Mobile edu/ high-quality, effective piece of writing, Robot Competition and Exhibition, AAAI also welcomes SARA-2000 as accessible to the general public or the Intelligent Systems Demonstrations, our first affiliate conference. For more to a broad AI audience (not just a subarea), the Robot Building Laboratory, information about the AAAI Affiliates written within the last two and the Doctoral Consortium. New Program, please write to Carol Hamilton years. The contribution should be for 2000 will be a technical paper at hamilton@aaai.org. Nominated papers must be Uncertainty: Operations Research AAAI is pleased to announce the continuation in English and must have been published Meets AI (Again)"; Justine Cassell, of its Student Abstract and in a publicly accessible place "Why Do We Need a Body Anyway?"; Poster Program, the SIGART/AAAI (for example, periodical, hard copy, or Carla Gomes, "Structure, Duality, and Doctoral Consortium, and the AAAI online journal but not only as a web Randomization: Common Themes in Scholarship and Volunteer Programs. The author(s) AI and OR"; James Hendler, "Missed Students interested in attending the of the award-winning paper(s) will Perceptions: AI versus the Funding National Conference on Artificial receive a $2500 prize (shared if more Agencies"; Geoff Hinton, "Modeling Intelligence in Austin, July 31-August than one author) as well as lodging High-Dimensional Data Distributions 3, 2000, should consult the AAAI web and travel to the National Conference by Combining Simple Experts"; Rich site for further information about all on Artificial Intelligence.

The AAAI-99 Workshop Program (a part of the sixteenth national conference on artificial intelligence) was held in Orlando, Florida. The program included 16 workshops covering a wide range of topics in AI. Each workshop was limited to approximately 25 to 50 participants. Participation was by invitation from the workshop organizers. The workshops were Agent-Based Systems in the Business Context, Agents' Conflicts, Artificial Intelligence for Distributed Information Networking, Artificial Intelligence for Electronic Commerce, Computation with Neural Systems Workshop, Configuration, Data Mining with Evolutionary Algorithms: Research Directions (Jointly sponsored by GECCO-99), Environmental Decision Support Systems and Artificial Intelligence, Exploring Synergies of Knowledge Management and Case-Based Reasoning, Intelligent Information Systems, Intelligent Software Engineering, Machine Learning for Information Extraction, Mixed-Initiative Intelligence, Negotiation: Settling Conflicts and Identifying Opportunities, Ontology Management, and Reasoning in Context for AI Applications.

A major problem in machine learning is that of inductive bias: how to choose a learner's hypothesis space so that it is large enough to contain a solution to the problem being learnt, yet small enough to ensure reliable generalization from reasonably-sized training sets. Typically such bias is supplied by hand through the skill and insights of experts. In this paper a model for automatically learning bias is investigated. The central assumption of the model is that the learner is embedded within an environment of related learning tasks. Within such an environment the learner can sample from multiple tasks, and hence it can search for a hypothesis space that contains good solutions to many of the problems in the environment. Under certain restrictions on the set of all hypothesis spaces available to the learner, we show that a hypothesis space that performs well on a sufficiently large number of training tasks will also perform well when learning novel tasks in the same environment. Explicit bounds are also derived demonstrating that learning multiple tasks within an environment of related tasks can potentially give much better generalization than learning a single task.

We introduce a semi-supervised support vector machine (S3yM) method. Given a training set of labeled data and a working set of unlabeled data, S3YM constructs a support vector machine using both the training and working sets. We use S3 YM to solve the transduction problem using overall risk minimization (ORM) posed by Yapnik. The transduction problem is to estimate the value of a classification function at the given points in the working set. This contrasts with the standard inductive learning problem of estimating the classification function at all possible values and then using the fixed function to deduce the classes of the working set data.

Calculation of Q(t) and R(t) using (4, 5, 7, 9) to execute the path average and the average over sets is relatively straightforward, albeit tedious. We find that -"Yt(l -"Yt)

We introduce a semi-supervised support vector machine (S3yM) method. Given a training set of labeled data and a working set of unlabeled data, S3YM constructs a support vector machine using both the training and working sets. We use S3 YM to solve the transduction problem using overall risk minimization (ORM) posed by Yapnik. The transduction problem is to estimate the value of a classification function at the given points in the working set. This contrasts with the standard inductive learning problem of estimating the classification function at all possible values and then using the fixed function to deduce the classes of the working set data.

Calculation of Q(t) and R(t) using (4, 5, 7, 9) to execute the path average and the average over sets is relatively straightforward, albeit tedious. We find that -"Yt(l -"Yt)

Calculation of Q(t) and R(t) using (4, 5, 7, 9) to execute the path average and the average over sets is relatively straightforward, albeit tedious. We find that -"Yt(l -"Yt)

We describe a unifying method for proving relative loss bounds for online linearthreshold classification algorithms, such as the Perceptron and the Winnow algorithms. For classification problems the discrete loss is used, i.e., the total number of prediction mistakes. We introduce a continuous lossfunction, called the "linear hinge loss", that can be employed to derive the updates of the algorithms. We first prove bounds w.r.t. the linear hinge loss and then convert them to the discrete loss. We introduce anotion of "average margin" of a set of examples . We show how relative loss bounds based on the linear hinge loss can be converted to relative loss bounds i.t.o. the discrete loss using the average margin.