Tutorial Title: Computational auditory scene analysis

Tutorial Abstract: The acoustic environment is typically composed of multiple simultaneous events, and a remarkable achievement of the auditory system is its ability to disentangle the acoustic mixture and group the acoustic energy that originates from the same event or source. This process of auditory grouping and segregation is referred to as auditory scene analysis (ASA). Decades of psychoacoustic research has uncovered a number of principles responsible for ASA, which motivated an emerging field of study called computational auditory scene analysis (CASA). CASA aims at sound source separation based on ASA cues, including pitch, location, amplitude/frequency modulation, and onset/offset. Recent developments in CASA show promising results in segregating speech from a variety of intrusions, where traditional signal-processing methods have difficulty. This tutorial will review recent CASA models and systems after a summary of related psychoacoustic results. An outline of the topics to be covered is the following:
      * Monaural source separation, with emphasis on speech segregation
      * Binaural source separation, with emphasis on speech segregation
      * Potential applications to robust speech and speaker recognition, hearing aids design, and human computer interaction
      * Discussion on the CASA perspective on sound source separation, comparison with related approaches (such as speech enhancement, adaptive beamforming and independent component analysis), and summary of existing challenges and directions for future progress.

Tutorial Title: Agent Communication

Tutorial Abstract: In this tutorial I will discuss the importance of agent communication in multi-agent systems. First I will explain why agent communication differs from most other forms of information exchange between programs. The main part of the tutorial will discuss how the communication should be integrated with the rest of the agent's actions. We will see some implementation of an agent communication language and discuss the different aspects of it. Finally, we discuss some general issues like verification of communication protocols and semantics of agent communication and issues related to multi-party communication.

Tutorial Title: Principles of InformationTheoretic Learning

Tutorial Abstract: The principles of how to train linear and nonlinear systems with entropy and divergence will be covered from an engineering perspective. Several algorithms will be explained in detail such that students can quickly learn how to integrate the new training methodology with their own tool boxes.

Tutorial Title: Statistical Relational Learning

Tutorial Abstract: Statistical relational learning (SRL) is a new and exciting area of AI that combines statistical learning and inductive logic programming. This tutorial will cover the necessary background in these two areas, and introduce some of the main SRL problems, approaches, and applications.