Reinforcement learning involves learning to adapt to environments through the presentation of rewards - special input - serving as clues. To obtain quick rational policies, profit sharing, rational policy making algor...
详细信息
ISBN:
(纸本)9783642153808
Reinforcement learning involves learning to adapt to environments through the presentation of rewards - special input - serving as clues. To obtain quick rational policies, profit sharing, rational policy making algorithm, penalty avoiding rational policy making algorithm (PARP), PS-r* and PS-r# are used. they are called Exploitation-oriented learning (XoL). When applying reinforcement learning to actual problems, treatment of continuous-valued input and output are sometimes required. A method based on PARP is proposed as a XoL method corresponding to the continuous-valued input, but continuous-valued output cannot be treated. We study the treatment of continuous-valued output suitable for a XoL method in which the environment includes both a reward and a penalty. We extend PARP in the continuous-valued input to continuous-valued output. We apply our proposal to the pole-cart balancing problem and confirm its validity.
Improving energy efficiency by monitoring household electrical consumption is of significant importance withthe present-day climate change concerns. A solution for the electrical consumption management problem is the...
详细信息
ISBN:
(纸本)9783642153808
Improving energy efficiency by monitoring household electrical consumption is of significant importance withthe present-day climate change concerns. A solution for the electrical consumption management problem is the use of a non-intrusive load monitoring system (NILM). this system captures the signals from the aggregate consumption, extracts the features from these signals and classifies the extracted features in order to identify the switched on appliances. An effective device identification (ID) requires a signature to be assigned for each appliance. Moreover, to specify an ID for each device, signal processing techniques are needed for extracting the relevant features. this paper describes a technique for the steady-states recognition in an electrical digital signal as the first stage for the implementation of an innovative NILM. Furthermore, the final goal is to develop an intelligent system for the identification of the appliances by automatedlearning. the proposed approach is based on the ratio value between rectangular areas defined by the signal samples. the computational experiments show the method effectiveness for the accurate steady-states identification in the electrical input signals.
OLAP systems depend heavily on the materialization of multidimensional structures to speed-up queries, whose appropriate selection constitutes the cube selection problem. However, the recently proposed distribution of...
详细信息
ISBN:
(纸本)9783642153808
OLAP systems depend heavily on the materialization of multidimensional structures to speed-up queries, whose appropriate selection constitutes the cube selection problem. However, the recently proposed distribution of OLAP structures emerges to answer new globalization's requirements, capturing the known advantages of distributed databases. But this hardens the search for solutions, especially due to the inherent heterogeneity, imposing an extra characteristic of the algorithm that must be used: adaptability. Here the emerging concept known as hyper-heuristic can be a solution. In fact, having an algorithm where several (meta-)heuristics may be selected under the control of a heuristic has an intrinsic adaptive behavior. this paper presents a hyper-heuristic polymorphic algorithm used to solve the extended cube selection and allocation problem generated in M-OLAP architectures.
In this paper, several different multi-agent based simulation systems are constructed using some types of learning mechanisms for the market entry games which are played by n players. this paper extracts some learning...
详细信息
In this paper, several different multi-agent based simulation systems are constructed using some types of learning mechanisms for the market entry games which are played by n players. this paper extracts some learning and behavioral rules of human in the market entry games by comparing the results of laboratory experiments using human subjects and the simulation experiments.
Head pose estimation is an important area of investigation for understanding human dynamics. Appearance-based methods are one of the popular solutions to this problem. In this paper we present a novel approach using v...
详细信息
ISBN:
(纸本)3540454853
Head pose estimation is an important area of investigation for understanding human dynamics. Appearance-based methods are one of the popular solutions to this problem. In this paper we present a novel approach using vector quantization that adds spatial information to the feature set. We compare this with raw, Gabor filtered and Wavelet features using the Carnegie Mellon PIE database. Our approach shows increased performance over the other methods.
Livestock is the primary source of meat, dairy, eggs, leather, wool, etc. Because of increased demand, increasing herd size maintained by fewer manpower is becoming important. Precision Livestock Farming (PLF) utilizi...
详细信息
Combining audio and image processing for understanding video content has several benefits when compared to using each modality on their own. For the task of context and activity recognition in video sequences, it is i...
详细信息
ISBN:
(纸本)3540454853
Combining audio and image processing for understanding video content has several benefits when compared to using each modality on their own. For the task of context and activity recognition in video sequences, it is important to explore bothdata streams to gather relevant information. In this paper we describe a video context and activity recognition model. Our work extracts a range of audio and visual features, followed by feature reduction and information fusion. We show that combining audio with video based decision making improves the quality of context and activity recognition in videos by 4% over audio data and 18% over image data.
As an Internet application, smart tourism has greatly enriched the tourism information. In this paper, we propose a unified modeling and expression method of attraction texts and images based on the text deep represen...
详细信息
ISBN:
(数字)9783319689357
ISBN:
(纸本)9783319689357;9783319689340
As an Internet application, smart tourism has greatly enriched the tourism information. In this paper, we propose a unified modeling and expression method of attraction texts and images based on the text deep representation model and convolution neural network. According to the cross-media characteristics of tourism big data, we propose a semantic learning and analysis method for cross-media data, and correlate tourism texts with images based on deep features and topic semantics. Experimental results show that the proposed method can achieve better results for semantic analysis and cross-media retrieval of tourism big data.
Context recognition in indoor and outdoor surroundings is an important area of research for the development of autonomous systems. this work describes an approach to the classification of audio signals found in both i...
详细信息
ISBN:
(纸本)3540454853
Context recognition in indoor and outdoor surroundings is an important area of research for the development of autonomous systems. this work describes an approach to the classification of audio signals found in both indoor and outdoor environments. Several audio features are extracted from raw signals. We analyze the relevance and importance of these features and use that information to design a multi-stage classifier architecture. Our results show that the multi-stage classification scheme is superior than a single stage classifier and it generates an 80% success rate on a 7 class problem.
the identification of human activity in video, for example whether a person is walking, clapping, waving, etc. is extremely important for video interpretation. In this paper we present a systematic approach to extract...
详细信息
ISBN:
(纸本)3540454853
the identification of human activity in video, for example whether a person is walking, clapping, waving, etc. is extremely important for video interpretation. In this paper we present a systematic approach to extracting visual features from image sequences that are used for classifying different activities. Furthermore, since different people perform the same action across different number of frames, matching training and test sequences is not a trivial task. We discuss a new technique for video shot matching where the shots matched are of different sizes. the proposed technique is based on frequency domain analysis of feature data and it is shown to achieve very high accuracy of 94.5% on recognizing a number of different human actions.
暂无评论