This paper introduces a representat ion of ev iden t i a l re la t ionsh ips which permits updating of be l i e f in two simultaneous modes: causal ( i . e . top-down) and diagnost ic ( i . e . bottom-up). I t extends the h ie ra rch i ca l t ree representat ion by a l lowing mu l t i p l e causes to a given mani fes ta t ion . We develop an updating scheme that obeys the axioms of p r o b a b ...

A technique is introduced for extracting and reconstructing a wide class of building types from a registered range image and optical image. An attentional focus stage, followed by model indexing, allows top-down robust surface ttting to reconstruct the 3D nature of the buildings in the data. Because of the effectiveness of model selection, top-down processing of noisy range data still succeeds ...

We address the problem of the automated verification of temporal properties of infinite state reactive systems. We present some improvements of a verification method based on the specialization of constraint logic programs (CLP). First, we reformulate the verification method as a two-phase procedure: (1) in the first phase a CLP specification of an infinite state system is specialized with resp...

Bottom-up cortical representations of visual conspicuity interact with top-down internal cognitive models of the external world to control eye movements, EMs, and the closely linked attention-shift mechanisms; to thus achieve visual recognition. Conspicuity operators implemented with image processing algorithms, IPAs, can discriminate human Regions-of-Interest, hROIs, the loci of eye fixations,...

Model-driven development is the process of creating models of a software system and transforming them into source code. Since the stepwise transformations can be done automatically or by hand, the notations of the models should be both precise and understandable. This is especially important if the software system is developed by a large, international team where the persons who model differ fr...

Efficient unsupervised training and inference in deep generative models remains a challenging problem. One basic approach, called Helmholtz machine or Variational Autoencoder, involves training a top-down directed generative model together with a bottom-up auxiliary model used for approximate inference. Recent results indicate that better generative models can be obtained with better approximat...

Context-dependent nature of biological phenomena are well documented in every branch of biology. While there have been few previous attempts to (implicitly) model various facets of biological context-dependence, a formal and general mathematical construct to model the wide spectrum of context-dependence, eludes the students of biology. An objective and rigorous model, from both 'bottom-up' as w...

This paper presents an agent-based methodological approach to design distributed service-oriented systems which can adapt their behaviour according to changes in the environment and in the user needs, even taking the initiative to make suggestions and proactive choices. The highly dynamic, regulated, complex nature of the distributed, interconnected services is tackled through a methodological ...

Attention and memory are very closely related and their aim is to simplify the acquired data into an intelligent structured data set. Two main points are discussed in this paper. The first one is the presentation of a novel visual attention model for still images which includes both a bottom-up and a top-down approach. The bottom-up model is based on structures rarity within the image during th...

We prove that the randomized decision tree complexity of the recursive majority-of-three is Ω(2.55), where d is the depth of the recursion. The proof is by a bottom up induction, which is same in spirit as the one in the proof of Saks and Wigderson in their 1986 paper on the complexity of evaluating game trees. Previous work includes an Ω (