The paper introduces a framework that integrates analytical inference into the particle filtering scheme for human body tracking. The analytical inference is provided by body parts detection, and is used to update sub...
详细信息
The Web has emerged as a huge information repository that can be used for various knowledge based applications, an important one being question answering (QA). The paper discusses the issues involved in using Web as a...
详细信息
The Web has emerged as a huge information repository that can be used for various knowledge based applications, an important one being question answering (QA). The paper discusses the issues involved in using Web as a knowledge base for question answering involving simple factual questions. It also proposes some simple but effective methods to adapt traditional QA methods to deal with efficient these issues and lead to an accurate and efficient question answering system. The paper discusses the architecture and design of our system. The empirical results obtained on the current system are encouraging.
3D models of urban sites with geometry and facade textures are needed for many planning and visualization applications. Approximate 3D wireframe model can be derived from aerial images but detailed textures must be ob...
详细信息
3D models of urban sites with geometry and facade textures are needed for many planning and visualization applications. Approximate 3D wireframe model can be derived from aerial images but detailed textures must be obtained from ground level images. Integrating such views with the 3D models is difficult as only small parts of buildings may be visible in a single view. We describe a method that uses two or three vanishing points, and three 3D to 2D line correspondences to estimate the rotational and translational parameters of the ground level cameras. The valid set of multiple combinations of 3D to 2D line pairs is chosen by a hypotheses generation and evaluation Some experimental results are presented.
In this paper we study the performance of compressed data for classification and anomaly detection. We use networks of various complexities for our purpose, guided by the data itself rather than one uniform-complexity...
详细信息
In this paper we study the performance of compressed data for classification and anomaly detection. We use networks of various complexities for our purpose, guided by the data itself rather than one uniform-complexity network for the entire data set.
Evolutionary algorithms for optimization of dynamic problems have recently received increasing attention. Online control is a particularly interesting class of dynamic problems, because of the interactions between the...
详细信息
Evolutionary algorithms for optimization of dynamic problems have recently received increasing attention. Online control is a particularly interesting class of dynamic problems, because of the interactions between the controller and the controlled system. In this paper, we report experimental results on two aspects of the direct control strategy in relation to a crop-producing greenhouse. In the first set of experiments, we investigated how to balance the available computation time between population size and generations. The second experiments were on different control horizons, and showed the importance of this aspect for direct control. Finally, we discuss the results in the wider context of dynamic optimization.
We consider three distributed decision making tasks that arise in the design and configuration of multi-hop wireless networks: medium access scheduling, Hamiltonian cycle formation, and the partitioning of network nod...
详细信息
ISBN:
(纸本)0769514359
We consider three distributed decision making tasks that arise in the design and configuration of multi-hop wireless networks: medium access scheduling, Hamiltonian cycle formation, and the partitioning of network nodes into coordinating cliques. We first model these tasks as distributed constraint satisfaction problems (DCSPs). We show that the communication complexity of DCSPs can be related to the computational complexity of centralized constraint satisfaction problems. We then use centralized algorithms to obtain experimental results on the solvability and complexity of the three DCSPs. We show that these problems exhibit "phase transitions" in solvability and complexity as the transmission power of the wireless nodes is varied. Based on these results, we argue that phase transition analysis provides a mechanism for quantifying the critical range of network resources needed for scalable, self-configuring multi-hop wireless networks.
3D models of urban sites with good geometry and facade textures are needed for many planning and visualization applications. Approximate wireframe can be derived from aerial images but detailed textures must be obtain...
详细信息
ISBN:
(纸本)076951695X
3D models of urban sites with good geometry and facade textures are needed for many planning and visualization applications. Approximate wireframe can be derived from aerial images but detailed textures must be obtained from ground level images. Integrating such views with the 3D models is difficult as only small parts of buildings may be visible in a single view. We describe a method that uses two or three vanishing points, not necessarily from orthogonal sets of parallel lines, and a small number of point correspondences to estimate the intrinsic and extrinsic parameters of the ground level cameras. Experimental results from some buildings are presented.
Using other agents' experiences and knowledge, a learning agent may learn faster, make fewer mistakes, and create some rides for unseen situations. These benefits will be gained if the learning agents know the are...
详细信息
Using other agents' experiences and knowledge, a learning agent may learn faster, make fewer mistakes, and create some rides for unseen situations. These benefits will be gained if the learning agents know the area of expertise and the expertness values of each other. In this paper, some Q-learning agents with different skills and expertness levels cooperate in learning. The agents use some criteria to judge others information and knowledge. Four expertness criterion, certainty and entropy measures are used to assign degrees of importance to others' Q-Tables. Effects of measuring these values based on their whole Q-Table, a portion of Q-Tables that reflects their proficiencies, and the states in Q-Tables on the learning quality are studied. Simple strategy sharing and two different weighted strategy-sharing methods are used to combine the acquired knowledge from different agents.
This paper presents an attempt to integrate attention and navigation skills in 3D embodied agents (virtual humanoids). The neural model presented has been divided in two main phases. Firstly the environment-categoriza...
详细信息
This paper presents an attempt to integrate attention and navigation skills in 3D embodied agents (virtual humanoids). The neural model presented has been divided in two main phases. Firstly the environment-categorization phase, where an online pattern recognition and categorization of the agent current sensory input data is carried out by an adaptive resonance driven self organizing neural network, which will finally simulate the agent's short term memory (STM). Secondly, the model must also learn how and when to map its current STM state into the navigation and attention motor layers of the 3D agent. We also review the world modelling and the agent vision system, and finally we present the first results extracted from two of the subsystems which conforms the complete neural model, such as, the environment categorization subsystem and the base navigation neural model.
暂无评论