The Experts below are selected from a list of 3033 Experts worldwide ranked by ideXlab platform

Philippe Salembier - One of the best experts on this subject based on the ideXlab platform.

  • ICASSP (2) - A Common Formulation for Interpolation, Prediction, and Update Lifting Design
    2006 IEEE International Conference on Acoustics Speed and Signal Processing Proceedings, 2006
    Co-Authors: Jesús Solé, Philippe Salembier
    Abstract:

    The optimization of a quadratic objective function with linear constraints is useful for interpolation purposes. This formulation may be employed to derive an Initial Prediction in the lifting scheme domain in order to construct wavelet transforms. We modify the formulation to design final Prediction and update lifting steps. The linear constraints relate wavelet bases and coefficients with the underlying signal. The objective function is the detail signal energy for the Prediction lifting design and the gradient of the approximation signal for the update. To report concrete results and the power of the approach, we derive update steps using an auto-regressive image model that show better performance than the 5/3 wavelet for the compression of several image classes.

  • A Common Formulation for Interpolation, Prediction, and Update Lifting Design
    2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings, 2006
    Co-Authors: Jesús Solé, Philippe Salembier
    Abstract:

    The optimization of a quadratic objective function with linear constraints is useful for interpolation purposes. This formulation may be employed to derive an Initial Prediction in the lifting scheme domain in order to construct wavelet transforms. We modify the formulation to design final Prediction and update lifting steps. The linear constraints relate wavelet bases and coefficients with the underlying signal. The objective function is the detail signal energy for the Prediction lifting design and the gradient of the approximation signal for the update. To report concrete results and the power of the approach, we derive update steps using an auto-regressive image model that show better performance than the 5/3 wavelet for the compression of several image classes

J. Fang - One of the best experts on this subject based on the ideXlab platform.

  • The accuracy of Initial Prediction in two-phase dynamic binary translators
    International Symposium on Code Generation and Optimization 2004. CGO 2004., 2004
    Co-Authors: Y. Wu, M. Breternitz, J. Quek, O. Etzion, J. Fang
    Abstract:

    Dynamic binary translators use a two-phase approach to identify and optimize frequently executed code dynamically. In the first step (profiling phase), blocks of code are interpreted or quickly translated to collect execution frequency information for the blocks. In the second phase (optimization phase), frequently executed blocks are grouped into regions and advanced optimizations are applied on them. This approach implicitly assumes that the Initial profile of each block is representative of the block throughout its lifetime. We investigate the ability of the Initial profile to predict the average program behavior. We compare the predicted behavior of varying lengths of the Initial execution with the average program behavior for the whole program execution, and use the Prediction from the training input as the reference. Our result indicates that, for the SPEC2000 benchmarks, even very short Initial profiles have comparable Prediction accuracy to the traditional profile-guided optimizations using the training input, although the Initial profile is inadequate for predicting loop trip count information for some integer programs and several benchmarks can benefit from phase-awareness during dynamic binary translation.

  • CGO - The accuracy of Initial Prediction in two-phase dynamic binary translators
    International Symposium on Code Generation and Optimization 2004. CGO 2004., 2004
    Co-Authors: Y. Wu, M. Breternitz, J. Quek, O. Etzion, J. Fang
    Abstract:

    Dynamic binary translators use a two-phase approach to identify and optimize frequently executed code dynamically. In the first step (profiling phase), blocks of code are interpreted or quickly translated to collect execution frequency information for the blocks. In the second phase (optimization phase), frequently executed blocks are grouped into regions and advanced optimizations are applied on them. This approach implicitly assumes that the Initial profile of each block is representative of the block throughout its lifetime. We investigate the ability of the Initial profile to predict the average program behavior. We compare the predicted behavior of varying lengths of the Initial execution with the average program behavior for the whole program execution, and use the Prediction from the training input as the reference. Our result indicates that, for the SPEC2000 benchmarks, even very short Initial profiles have comparable Prediction accuracy to the traditional profile-guided optimizations using the training input, although the Initial profile is inadequate for predicting loop trip count information for some integer programs and several benchmarks can benefit from phase-awareness during dynamic binary translation.

Jesús Solé - One of the best experts on this subject based on the ideXlab platform.

  • ICASSP (2) - A Common Formulation for Interpolation, Prediction, and Update Lifting Design
    2006 IEEE International Conference on Acoustics Speed and Signal Processing Proceedings, 2006
    Co-Authors: Jesús Solé, Philippe Salembier
    Abstract:

    The optimization of a quadratic objective function with linear constraints is useful for interpolation purposes. This formulation may be employed to derive an Initial Prediction in the lifting scheme domain in order to construct wavelet transforms. We modify the formulation to design final Prediction and update lifting steps. The linear constraints relate wavelet bases and coefficients with the underlying signal. The objective function is the detail signal energy for the Prediction lifting design and the gradient of the approximation signal for the update. To report concrete results and the power of the approach, we derive update steps using an auto-regressive image model that show better performance than the 5/3 wavelet for the compression of several image classes.

  • A Common Formulation for Interpolation, Prediction, and Update Lifting Design
    2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings, 2006
    Co-Authors: Jesús Solé, Philippe Salembier
    Abstract:

    The optimization of a quadratic objective function with linear constraints is useful for interpolation purposes. This formulation may be employed to derive an Initial Prediction in the lifting scheme domain in order to construct wavelet transforms. We modify the formulation to design final Prediction and update lifting steps. The linear constraints relate wavelet bases and coefficients with the underlying signal. The objective function is the detail signal energy for the Prediction lifting design and the gradient of the approximation signal for the update. To report concrete results and the power of the approach, we derive update steps using an auto-regressive image model that show better performance than the 5/3 wavelet for the compression of several image classes

Y. Wu - One of the best experts on this subject based on the ideXlab platform.

  • The accuracy of Initial Prediction in two-phase dynamic binary translators
    International Symposium on Code Generation and Optimization 2004. CGO 2004., 2004
    Co-Authors: Y. Wu, M. Breternitz, J. Quek, O. Etzion, J. Fang
    Abstract:

    Dynamic binary translators use a two-phase approach to identify and optimize frequently executed code dynamically. In the first step (profiling phase), blocks of code are interpreted or quickly translated to collect execution frequency information for the blocks. In the second phase (optimization phase), frequently executed blocks are grouped into regions and advanced optimizations are applied on them. This approach implicitly assumes that the Initial profile of each block is representative of the block throughout its lifetime. We investigate the ability of the Initial profile to predict the average program behavior. We compare the predicted behavior of varying lengths of the Initial execution with the average program behavior for the whole program execution, and use the Prediction from the training input as the reference. Our result indicates that, for the SPEC2000 benchmarks, even very short Initial profiles have comparable Prediction accuracy to the traditional profile-guided optimizations using the training input, although the Initial profile is inadequate for predicting loop trip count information for some integer programs and several benchmarks can benefit from phase-awareness during dynamic binary translation.

  • CGO - The accuracy of Initial Prediction in two-phase dynamic binary translators
    International Symposium on Code Generation and Optimization 2004. CGO 2004., 2004
    Co-Authors: Y. Wu, M. Breternitz, J. Quek, O. Etzion, J. Fang
    Abstract:

    Dynamic binary translators use a two-phase approach to identify and optimize frequently executed code dynamically. In the first step (profiling phase), blocks of code are interpreted or quickly translated to collect execution frequency information for the blocks. In the second phase (optimization phase), frequently executed blocks are grouped into regions and advanced optimizations are applied on them. This approach implicitly assumes that the Initial profile of each block is representative of the block throughout its lifetime. We investigate the ability of the Initial profile to predict the average program behavior. We compare the predicted behavior of varying lengths of the Initial execution with the average program behavior for the whole program execution, and use the Prediction from the training input as the reference. Our result indicates that, for the SPEC2000 benchmarks, even very short Initial profiles have comparable Prediction accuracy to the traditional profile-guided optimizations using the training input, although the Initial profile is inadequate for predicting loop trip count information for some integer programs and several benchmarks can benefit from phase-awareness during dynamic binary translation.

Hideyoshi Yanagisawa - One of the best experts on this subject based on the ideXlab platform.

  • SMC - Mathematical Model of Emotional Habituation to Novelty: Modeling with Bayesian Update and Information Theory *
    2019 IEEE International Conference on Systems Man and Cybernetics (SMC), 2019
    Co-Authors: Takahiro Sekoguchi, Yuki Sakai, Hideyoshi Yanagisawa
    Abstract:

    Novelty is an important factor of creativity in product design. Acceptance of novelty, however, depends on one’s emotions. Yanagisawa, the last author, and his colleagues previously developed a mathematical model of emotional dimensions associated with novelty such as arousal (surprise) and valence (positivity and negativity). The model formalized arousal as Bayesian information gain and valence as a function of arousal based on Berlyne’s arousal potential theory. One becomes accustomed to novelty by repeated exposure. This so-called habituation to novelty is important in the design of long-term product experience. We herein propose a mathematical model of habituation to novelty based on the emotional dimension model. We formalized the habituation as a decrement in information gain from a novel event through Bayesian update. We derived the information gained from the repeated exposure of a novel stimulus as a function of three parameters: Initial Prediction error, Initial uncertainty, and noise of sensory stimulus. With the proposed model, we discovered an interaction effect of the Initial Prediction error and Initial uncertainty on habituation. Furthermore, we demonstrate that a range of positive emotions on Prediction errors shift toward becoming more novel by repeated exposure. We hypothesize that the ease to become accustomed to novelty depends on the Initial uncertainty. To verify this hypothesis, we conducted an experiment with several short videos in which different percussion instruments were played. We manipulated the uncertainty level by the popularity of instruments and Prediction error by the congruity between sounds and videos. We used event-related potential P300 amplitudes and the subjective reports of surprise in response to the sounds as measures of arousal levels. The experimental results supported our hypothesis; therefore, the decrement in information gain can be decomposed into Initial Prediction error and Initial uncertainty, and is considered as a valid measure of emotional habituation.

  • Mathematical Model of Emotional Habituation to Novelty: Modeling with Bayesian Update and Information Theory
    arXiv: Information Theory, 2019
    Co-Authors: Takahiro Sekoguchi, Yuki Sakai, Hideyoshi Yanagisawa
    Abstract:

    Novelty is an important factor of creativity in product design. Acceptance of novelty, however, depends on one's emotions. Yanagisawa, the last author, and his colleagues previously developed a mathematical model of emotional dimensions associated with novelty such as arousal (surprise) and valence. The model formalized arousal as Bayesian information gain and valence as a function of arousal based on Berlyne's arousal potential theory. One becomes accustomed to novelty by repeated exposure. This so-called habituation to novelty is important in the design of long-term product experience. We herein propose a mathematical model of habituation to novelty based on the emotional dimension model. We formalized the habituation as a decrement in information gain from a novel event through Bayesian update. We derived the information gained from the repeated exposure of a novel stimulus as a function of three parameters: Initial Prediction error, Initial uncertainty, and noise of sensory stimulus. With the proposed model, we discovered an interaction effect of the Initial Prediction error and Initial uncertainty on habituation. Furthermore, we demonstrate that a range of positive emotions on Prediction errors shift toward becoming more novel by repeated exposure.

  • Mathematical Model of Emotional Habituation to Novelty: Modeling with Bayesian Update and Information Theory*
    2019 IEEE International Conference on Systems Man and Cybernetics (SMC), 2019
    Co-Authors: Takahiro Sekoguchi, Yuki Sakai, Hideyoshi Yanagisawa
    Abstract:

    Novelty is an important factor of creativity in product design. Acceptance of novelty, however, depends on one's emotions. Yanagisawa, the last author, and his colleagues previously developed a mathematical model of emotional dimensions associated with novelty such as arousal (surprise) and valence (positivity and negativity). The model formalized arousal as Bayesian information gain and valence as a function of arousal based on Berlyne's arousal potential theory. One becomes accustomed to novelty by repeated exposure. This so-called habituation to novelty is important in the design of long-term product experience. We herein propose a mathematical model of habituation to novelty based on the emotional dimension model. We formalized the habituation as a decrement in information gain from a novel event through Bayesian update. We derived the information gained from the repeated exposure of a novel stimulus as a function of three parameters: Initial Prediction error, Initial uncertainty, and noise of sensory stimulus. With the proposed model, we discovered an interaction effect of the Initial Prediction error and Initial uncertainty on habituation. Furthermore, we demonstrate that a range of positive emotions on Prediction errors shift toward becoming more novel by repeated exposure. We hypothesize that the ease to become accustomed to novelty depends on the Initial uncertainty. To verify this hypothesis, we conducted an experiment with several short videos in which different percussion instruments were played. We manipulated the uncertainty level by the popularity of instruments and Prediction error by the congruity between sounds and videos. We used event-related potential P300 amplitudes and the subjective reports of surprise in response to the sounds as measures of arousal levels. The experimental results supported our hypothesis; therefore, the decrement in information gain can be decomposed into Initial Prediction error and Initial uncertainty, and is considered as a valid measure of emotional habituation.