Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 20 of 49 for “"feature representations"”.

  1. Towards robust and domain invariant feature representations in Deep Learning

    … perception-based systems is to define and learn representations of the scene that are more robust and adaptive to several nuisance factors. Over the recent past, for a variety of tasks involving images, learned representations have been empirically shown to outperform handcrafted ones. However, …

    maryland Repository record for Towards robust and domain invariant feature representations in Deep Learning (opens in a new tab)

  2. Understanding the effects of higher order sequence features on peptide MHC binding

    … In this thesis we propose the use of sequence feature representations as a means of capturing and categorizing these factors, and we develop the theoretical framework and justification for their use. We then apply sequence feature representations to analyze data derived from yeast display …

    mit Repository record for Understanding the effects of higher order sequence features on peptide MHC binding (opens in a new tab)

  3. Representation and transfer learning using information-theoretic approximations

    Learning informative and transferable feature representations is a key aspect of machine learning systems. Mutual information and Kullback-Leibler divergence are principled and very popular metrics to measure feature relevance and perform distribution matching, respectively. However, clean …

    mit Repository record for Representation and transfer learning using information-theoretic approximations (opens in a new tab)

  4. Bridging Deep Learning and Probabilistic Inference: Towards Data Efficiency, Identifiability, and Sampling Scalability

    … the theoretical properties of neural network representations learned across multiple tasks within a probabilistic framework, establishing conditions under which neural networks can recover canonical feature representations that reflect the underlying ground-truth data generating process. Our …

    cambridge Repository record for Bridging Deep Learning and Probabilistic Inference: Towards Data Efficiency, Identifiability, and Sampling Scalability (opens in a new tab)

  5. Learning Gaussisan noise models from high-dimensional sensor data with deep neural networks

    … thesis describes a method of learning compact feature representations for real-time covariance estimation. A direct log-likelihood optimization technique is used to train a deep convolutional neural network to predict the covariance matrix of a Gaussian measurement model, given representative …

    mit Repository record for Learning Gaussisan noise models from high-dimensional sensor data with deep neural networks (opens in a new tab)

  6. Using Machine Learning for Description and Inference of Cyber Threats, Vulnerabilities, and Mitigations

    … cyber threats. We experiment with different feature representations and subsets of the data, and show that machine learning and NLP can effectively classify edges between entries from different data sources as well as predict possible edge candidates. Experts agree that several of our …

    mit Repository record for Using Machine Learning for Description and Inference of Cyber Threats, Vulnerabilities, and Mitigations (opens in a new tab)

  7. DEEP LEARNING FOR FOR LARGE-SCALE FACIAL EXPRESSION RECOGNITION IN THE WILD

    … via label fusion in order to obtain better feature representations. Experiments on in-the-wild AffectNet as well as a collection of laboratory based FER datasets suggests that jointly learning both dimensional and categorical models of affect significantly reduce redundancy and computational …

    nus Repository record for DEEP LEARNING FOR FOR LARGE-SCALE FACIAL EXPRESSION RECOGNITION IN THE WILD (opens in a new tab)

  8. Benchmarking Methods For Predicting Phenotype Gene Associations

    … on multiple protein interaction networks and feature representations. We empirically evaluate the performance of multiple prediction tasks using two evaluation experiments: cross-fold validation and the more stringent temporal holdout. We demonstrate that all of the prediction methods …

    vt Repository record for Benchmarking Methods For Predicting Phenotype Gene Associations (opens in a new tab)

  9. Scalability and interpretability of graph neural networks for small molecules

    … network (SMNN), which is designed to have all feature representations and weights be human interpretable. I show that this network can achieve competitive performance with common graph neural network baselines. I also show that the network is capable of learning features that allow for transfer …

    mit Repository record for Scalability and interpretability of graph neural networks for small molecules (opens in a new tab)

  10. Detecting cells and analyzing their behaviors in microscopy images using deep neural networks

    … of these advances is exploiting hierarchical feature representations by various deep learning models, instead of handcrafted features based on domain-specific knowledge.</p> <p>In the work presented in this dissertation, we are particularly interested in exploring the power of deep neural …

    must-thes Repository record for Detecting cells and analyzing their behaviors in microscopy images using deep neural networks (opens in a new tab)

  11. Multi-theme sentiment analysis with sentiment shifting

    … sentiment by learning embeddings (i.e., vector representations) for both themes and words, and derives the shifter effect learning algorithm by modeling the shifted sentiment in a logistic regression model. Extensive experiments have been conducted on Yelp business reviews and IMDB movie …

    uiuc Repository record for Multi-theme sentiment analysis with sentiment shifting (opens in a new tab)

  12. A machine learning based method for sensitivity estimation for accelerated magnetic resonance spectroscopy imaging using phased array coils

    … prior information in the form of learned image feature representations may be combined with noisy imaging data to produce high-resolution, artifact-free sensitivity profiles. An in-vivo experiment demonstrates the effectiveness of the proposed method. The relative SENSE reconstruction error for …

    uiuc Repository record for A machine learning based method for sensitivity estimation for accelerated magnetic resonance spectroscopy imaging using phased array coils (opens in a new tab)

  13. Learning the Language of Antibody Hypervariability Through Biological Property Prediction

    … and binding specificity. We demonstrate how our feature representations can be applied to the accurate prediction of an antibody’s local and global 3D structures, mutational effects on antigen binding specificity, as well as identification of its paratope. The scalability of AbMAP newly enables …

    mit Repository record for Learning the Language of Antibody Hypervariability Through Biological Property Prediction (opens in a new tab)

  14. Improving and Analyzing Model Merging Methods for Adaptation

    … of combining models by averaging intermediate features, referred to as model merging, and propose a new direction for achieving collective model intelligence through what we call compatible specialization. Current methods for model merging, such as parameter and feature averaging, struggle to …

    mit Repository record for Improving and Analyzing Model Merging Methods for Adaptation (opens in a new tab)

  15. Advances in discriminative dependency parsing

    … second, the ability to use arbitrarily-defined feature representations. This thesis explores three advances in the field of discriminative dependency parsing. First, we show that the classic Matrix-Tree Theorem (Kirchhoff, 1847; Tutte, 1984) can be applied to the problem of non-projective …

    mit Repository record for Advances in discriminative dependency parsing (opens in a new tab)

  16. Effects of automated cartographic generalization on linear map features

    … relates to manual cartographic methods and feature representation is analyzed. It is suggested that the nature of representation of linear features on maps be considered in the analysis of effectiveness of automated generalization. The development of a computer platform for evaluating linear …

    vt Repository record for Effects of automated cartographic generalization on linear map features (opens in a new tab)

  17. Inferring travel activity pattern from smartphone sensing data using deep learning

    … algorithms provide a framework for learning feature representation from raw data. The convolutional neural networks have been particularly effective in learning feature representations on many datasets. These models have achieved significant improvement on many complex problems over other …

    mit Repository record for Inferring travel activity pattern from smartphone sensing data using deep learning (opens in a new tab)

  18. Transfer learning algorithms for image classification

    … be able to exploit complex high dimensional feature representations even when only a few labeled examples are available for training. To achieve this goal we develop transfer learning algorithms that: 1) Leverage unlabeled data annotated with meta-data and 2) Exploit labeled data from related …

    mit Repository record for Transfer learning algorithms for image classification (opens in a new tab)

  19. Romance, revolution and regulation: colonialism and the US-Mexico border in American Cold War film.

    … to focus on films of different genres which feature representations of the US-Mexico border. The thesis’ central contribution therefore lies in its assertion that a study which is attentive to cinematic space and focused on a particular cinematic location can provide new ways of understanding …

    east-anglia Repository record for Romance, revolution and regulation: colonialism and the US-Mexico border in American Cold War film. (opens in a new tab)

  20. How deep learning can help emotion recognition

    … through pre-specified rules or hand-crafted features. However, in the last few years, learned feature representations have experienced a resurgence mainly due to the success of deep neural networks. In this dissertation, we highlight how deep neural networks, when applied to emotion …

    uiuc Repository record for How deep learning can help emotion recognition (opens in a new tab)

Page 1 of 3