We extend the result model for precedent-based reasoning with incomplete case bases. In contrast to regular case bases, these consist of incomplete cases for which not all dimension values need to be specified, but ra...
详细信息
ISBN:
(纸本)9781643684727;9781643684734
We extend the result model for precedent-based reasoning with incomplete case bases. In contrast to regular case bases, these consist of incomplete cases for which not all dimension values need to be specified, but rather each dimension is assigned a set of possible values. The outcome of cases then applies for each (combination of) the possible dimension values. Building on earlier proposed notions of justification and stability for incomplete focus cases, we introduce the notion of possible justification statuses, which are required to maintain consistency of the incomplete case base. We demonstrate how these theoretic notions can be applied in practice for human-in-the-loop decision support, discuss their computational complexity and provide efficient algorithms.
An algorithm of fast texture synthesis by block tiling is proposed after summarizing several related *** approach uses displacement technology based on correlation to get the mapping addresses of synthesized blocks in...
详细信息
An algorithm of fast texture synthesis by block tiling is proposed after summarizing several related *** approach uses displacement technology based on correlation to get the mapping addresses of synthesized blocks in the sample image and integrates the method of searching feasible blocks along a spiral *** the process,boundary matching algorithm is adopted to find the optimal matching block which meets the threshold,then continues this process in a recursive manner until the synthesizing image is fully filled. Experimental results show that this algorithm has achieved better desired effect and used much less time when dealing with the random texture and the structural texture compared with the previous algorithms.
This paper describes an end-of-semester day-long required project used as a capstone to a junior electrical engineering technology course in Analog Integrated Circuits Applications. The motivation for the project is p...
详细信息
This paper describes an end-of-semester day-long required project used as a capstone to a junior electrical engineering technology course in Analog Integrated Circuits Applications. The motivation for the project is presented in the Introduction. The Project Description explains both the problem presented to the students and the implementation constraints. The Evaluation section has three parts;the subjective evaluation of the project by the students, the subjective evaluation of the project by the course instructors, and the objective results, including the students' performance evaluation, and a comparison of the students' overall course performance as compared with previous semesters which did not include a project. Finally, continuing trends in the application of teaming are presented.
We developed a label-designing and restoration method for end-to-end automatic speech recognition based on connectionist temporal classification (CTC). With an end-to-end speech-recognition system including thousands ...
详细信息
ISBN:
(纸本)9789881476852
We developed a label-designing and restoration method for end-to-end automatic speech recognition based on connectionist temporal classification (CTC). With an end-to-end speech-recognition system including thousands of output labels such as words or characters, it is difficult to train a robust model because of data sparsity. With our proposed method, characters with less training data are estimated using the context of a language model rather than the acoustic features. Our method involves two steps. First, we train acoustic models using 70 class labels instead of thousands of low-frequency labels. Second, the class labels are restored to the original labels by using a weighted finite state transducer and n-gram language model. We applied the proposed method to a Japanese end-to-end automatic speechr-ecognition system including labels of over 3,000 characters. Experimental results indicate that the word error rate relatively improved with our method by a maximum of 15.5% compared with a conventional CTC-based method and is comparable to state-of-the-art hybrid DNN methods.
Deep learning neural network is very powerful to deal with signal processing, computer vision and many other recognition problems since its high accuracy. The CUDA based framework is a mainstream in deep learning mode...
详细信息
ISBN:
(纸本)9781538622858
Deep learning neural network is very powerful to deal with signal processing, computer vision and many other recognition problems since its high accuracy. The CUDA based framework is a mainstream in deep learning models. In this paper, we present OpenCL based implementation of the most advanced deep learning framework. Our work shows following contributions: (1) It is a real-time object recognition framework. (2) Our framework can work across variety of compute devices, e.g., Mali and AMD, etc. (3) The framework can be applied in various constraint situations, such as portable devices (embedded system) or low power applications (embedded system and FPGA).
暂无评论