Stochastic meta-descent (SMD) is a new technique for online adaptation oflocal learning rates in arbitrary twice-differentiable systems. Like matrix momentum it uses full second-order information while retaining O(n) computational complexity by exploiting the efficient computation of Hessian-vector products. Here we apply SMD to independent component analysis, and employ the resulting algorithmfor the blind separation of time-varying mixtures. By matching individual learning rates to the rate of change in each source signal's mixture coefficients, our technique is capable of simultaneously trackingsources that move at very different, a priori unknown speeds. 1 Introduction Independent component analysis (ICA) methods are typically run in batch mode in order to keep the stochasticity of the empirical gradient low. Often this is combined with a global learning rate annealing scheme that negotiates the tradeoff between fast convergence and good asymptotic performance.

Stochastic meta-descent (SMD) is a new technique for online adaptation of local learning rates in arbitrary twice-differentiable systems. Like matrix momentum it uses full second-order information while retaining O(n) computational complexity by exploiting the efficient computation of Hessian-vector products. Here we apply SMD to independent component analysis, and employ the resulting algorithm for the blind separation of time-varying mixtures. By matching individual learning rates to the rate of change in each source signal's mixture coefficients, our technique is capable of simultaneously tracking sources that move at very different, a priori unknown speeds.

In hyperspectral imagery one pixel typically consists of a mixture of the reflectance spectra of several materials, where the mixture coefficients correspond to the abundances of the constituting materials. We assume linear combinations of reflectance spectra with some additive normal sensor noise and derive a probabilistic MAP framework for analyzing hyperspectral data. As the material reflectance characteristics are not know a priori, we face the problem of unsupervised linear unmixing.

The project pursued in this paper is to develop from first information-geometric principles a general method for learning the similarity between text documents. Each individual document is modeled as a memoryless information source. Based on a latent class decomposition of the term-document matrix, a lowdimensional (curved) multinomial subfamily is learned. From this model a canonical similarity function - known as the Fisher kernel - is derived. Our approach can be applied for unsupervised and supervised learning problems alike.

In hyperspectral imagery one pixel typically consists of a mixture of the reflectance spectra of several materials, where the mixture coefficients correspond to the abundances of the constituting materials. We assume linear combinations of reflectance spectra with some additive normal sensor noise and derive a probabilistic MAP framework for analyzing hyperspectral data. As the material reflectance characteristics are not know a priori, we face the problem of unsupervised linear unmixing.

I was asked to give a visionary talk about the future applications of Artificial Intelligence technology; but I should warn you that I'm actually not very good as a visionary. Most of my predictions about what will happen in the industry don't come true even though they ought to. So I'm not going to tell you what the future holds; what I will do is to point out some of the technological trends that are at work. The outline of the talk is as follows: I'll start off by looking at the previous IAAI conferences and reflect on what we've learned from them. Then I'll look at what's changing in the hardware base that sets the context for all the computer applications we do. I think that will lead to interesting new viewpoints. Next I'll sketch what applications might arise from this new viewpoint. Finally, I'll discuss how the development of practical applications ought to interact with the scientific enterprise of trying to understand intelligence, in particular, human intelligence.

The Fifth International Conference on Artificial Intelligence Planning and Scheduling (AIPS 2000) was held on 14-17 April 2000 at Breckenridge, Colorado; it was colocated with the Seventh International Conference on Principles of Knowledge Representation and Reasoning (KR2000). This conference brought together researchers working in all aspects of problems in planning, scheduling, planning and learning, and plan execution for dealing with complex problems.

Each award winner and received a B.S. in electrical received a certificate and a check engineering from the Technion Haifa for $2500.

The Eighth Annual Mobile Robot Competition and Exhibition was held as part of the Sixteenth National Conference on Artificial Intelligence in Orlando, Florida, 18 to 22 July. The goals of these robot events are to foster the sharing of research and technology, allow research groups to showcase their achievements, encourage students to enter robotics and AI fields at both the undergraduate and graduate level, and increase awareness of the field. The 1999 events included two robot contests; a new, long-term robot challenge; an exhibition; and a National Botball Championship for high school teams sponsored by the KISS Institute. Each of these events is described in detail in this article.

In 1998, the international planning community was invited to take part in the first planning competition, hosted by the Artificial Intelligence Planning Systems Conference, to provide a new impetus for empirical evaluation and direct comparison of automatic domain-independent planning systems. This article describes the systems that competed in the event, examines the results, and considers some of the implications for the future of the field.