Computer Sciences Commons^™

Generating Diverse And Meaningful Captions: Unsupervised Specificity Optimization For Image Captioning, Annika Lindh, Robert J. Ross, Abhijit Mahalunkar, Giancarlo Salton, John D. Kelleher

Conference papers

Image Captioning is a task that requires models to acquire a multi-modal understanding of the world and to express this understanding in natural language text. While the state-of-the-art for this task has rapidly improved in terms of n-gram metrics, these models tend to output the same generic captions for similar images. In this work, we address this limitation and train a model that generates more diverse and specific captions through an unsupervised training approach that incorporates a learning signal from an Image Retrieval model. We summarize previous results and improve the state-of-the-art on caption diversity and novelty.

We make our …

Graph-Based Latent Embedding, Annotation And Representation Learning In Neural Networks For Semi-Supervised And Unsupervised Settings, Ismail Ozsel Kilinc

USF Tampa Graduate Theses and Dissertations

Machine learning has been immensely successful in supervised learning with outstanding examples in major industrial applications such as voice and image recognition. Following these developments, the most recent research has now begun to focus primarily on algorithms which can exploit very large sets of unlabeled examples to reduce the amount of manually labeled data required for existing models to perform well. In this dissertation, we propose graph-based latent embedding/annotation/representation learning techniques in neural networks tailored for semi-supervised and unsupervised learning problems. Specifically, we propose a novel regularization technique called Graph-based Activity Regularization (GAR) and a novel output layer modification called …

Modular Mechanistic Networks: On Bridging Mechanistic And Phenomenological Models With Deep Neural Networks In Natural Language Processing, Simon Dobnik, John D. Kelleher

Books/Book chapters

Natural language processing (NLP) can be done using either top-down (theory driven) and bottom-up (data driven) approaches, which we call mechanistic and phenomenological respectively. The approaches are frequently considered to stand in opposition to each other. Examining some recent approaches in deep learning we argue that deep neural networks incorporate both perspectives and, furthermore, that leveraging this aspect of deep learning may help in solving complex problems within language technology, such as modelling language and perception in the domain of spatial cognition.

What Is Not Where: The Challenge Of Integrating Spatial Representations Into Deep Learning Architectures, John D. Kelleher, Simon Dobnik

Books/Book chapters

This paper examines to what degree current deep learning architectures for image caption generation capture spatial lan- guage. On the basis of the evaluation of examples of generated captions from the literature we argue that systems capture what objects are in the image data but not where these objects are located: the cap- tions generated by these systems are the output of a language model conditioned on the output of an object detector that cannot capture fine-grained location information. Although language models provide useful knowledge for image captions, we argue that deep learning image captioning architectures should also model geometric …

Comparing Tensorflow Deep Learning Performance Using Cpus, Gpus, Local Pcs And Cloud, John Lawrence, Jonas Malmsten, Andrey Rybka, Daniel A. Sabol, Ken Triplin

Publications and Research

Deep learning is a very computational intensive task. Traditionally GPUs have been used to speed-up computations by several orders of magnitude. TensorFlow is a deep learning framework designed to improve performance further by running on multiple nodes in a distributed system. While TensorFlow has only been available for a little over a year, it has quickly become the most popular open source machine learning project on GitHub. The open source version of TensorFlow was originally only capable of running on a single node while Google’s proprietary version only was capable of leveraging distributed systems. This has now changed. In this …

A Deep Learning Based Pipeline For Metastatic Breast Cancer Classification From Whole Slide Images (Wsi), Arjun Punabhai Vekariya

Computer Science and Engineering Theses - Archive

Pathology is a 150-year-old medical specialty that has seen a paradigm shift over the past few years with the advent of Digital Pathology. Digital Pathology is a very promising approach to diagnostic medicine to accomplish better, faster and cheaper diagnosis, prognosis and prediction of cancer and other important diseases. Historical approaches in Digital Pathology have focused primarily on low-level image analysis tasks (e.g., color normalization, nuclear segmentation, and feature extraction) hence they are not generalized, thus not useful for practical use in clinical practices. In this thesis, a general Deep Learning based classification pipeline for identifying cancer metastases from histological …

Deep Learning Implementation For Comparison Of User Reviews And Ratings, Nishit Shrestha

UNLV Theses, Dissertations, Professional Papers, and Capstones

Sentiment Analysis is the task of identifying and classifying the sentiment expressed in a piece of text as of positive or negative sentiment and has wide application in E-Commerce. In present time, most e-commerce websites have product review sections, which can be used to identify customer satisfaction/dissatisfaction for their product. In E-COMMERCE websites such as Amazon.com, E-bay.com etc, consumers can submit their reviews along with a specific polarity rating (e.g. 1 to 5 stars at Amazon.com). There is a possibility of mismatch between review submitted and polarity of rating. For Amazon.com, a customer can submit a strongly positive review but …

Impact Of Artificial Intelligence, Robotics, And Machine Learning On Sales And Marketing, Keng Siau, Y. Yang

Research Collection School Of Computing and Information Systems

AI, robotics, and machine learning are impacting the field of sales and marketing in an unprecedented way. A perfect storm is brewing! On one hand, online retail stores like Amazon are crushing the bricks and mortar stores. Sales and marketing professionals in bricks and mortar stores are facing a grim future. On the other hand, AI, robotics, and machine learning are replacing sales and marketing professionals in online stores. In fact, salespersons and marketers are predicted to be among the first to be replaced by robots. In a face-to-face environment, human may still prefer to interact with another human. In …

Deep Neural Networks With Confidence Sampling For Electrical Anomaly Detection, Norman L. Tasfi, Wilson A. Higashino, Katarina Grolinger, Miriam A. M. Capretz

Electrical and Computer Engineering Publications

The increase in electrical metering has created tremendous quantities of data and, as a result, possibilities for deep insights into energy usage, better energy management, and new ways of energy conservation. As buildings are responsible for a significant portion of overall energy consumption, conservation efforts targeting buildings can provide tremendous effect on energy savings. Building energy monitoring enables identification of anomalous or unexpected behaviors which, when corrected, can lead to energy savings. Although the available data is large, the limited availability of labels makes anomaly detection difficult. This research proposes a deep semi-supervised convolutional neural network with confidence sampling for …

Deep Learning Approach For Intrusion Detection System (Ids) In The Internet Of Things (Iot) Network Using Gated Recurrent Neural Networks (Gru), Manoj Kumar Putchala

Browse all Theses and Dissertations

The Internet of Things (IoT) is a complex paradigm where billions of devices are connected to a network. These connected devices form an intelligent system of systems that share the data without human-to-computer or human-to-human interaction. These systems extract meaningful data that can transform human lives, businesses, and the world in significant ways. However, the reality of IoT is prone to countless cyber-attacks in the extremely hostile environment like the internet. The recent hack of 2014 Jeep Cherokee, iStan pacemaker, and a German steel plant are a few notable security breaches. To secure an IoT system, the traditional high-end security …

Analysing The Effects Of Data Augmentation And Free Parameters For Text Classification With Recurrent Convolutional Neural Networks, Jonathan Quijas

Open Access Theses & Dissertations

Convolutional neural networks have seen much success in computer vision and natural language processing tasks. When training convolutional neural networks for text classification tasks, a common technique is to transform an input sequence of words into a dense matrix of word embeddings, or words represented as dense vectors, using table lookup operations. This enables the inputs to be represented in a way that the well-known convolution/pooling operations can be applied to them in a manner similar to images. These word embeddings may be further incorporated into the neural network itself as a trainable layer to allow fine-tuning, usually leading to …

Investigating The Impact Of Unsupervised Feature-Extraction From Multi-Wavelength Image Data For Photometric Classification Of Stars, Galaxies And Qsos, Annika Lindh

Conference papers

Accurate classification of astronomical objects currently relies on spectroscopic data. Acquiring this data is time-consuming and expensive compared to photometric data. Hence, improving the accuracy of photometric classification could lead to far better coverage and faster classification pipelines. This paper investigates the benefit of using unsupervised feature-extraction from multi-wavelength image data for photometric classification of stars, galaxies and QSOs. An unsupervised Deep Belief Network is used, giving the model a higher level of interpretability thanks to its generative nature and layer-wise training. A Random Forest classifier is used to measure the contribution of the novel features compared to a set …

Investigating The Impact Of Unsupervised Feature-Extraction From Multi-Wavelength Image Data For Photometric Classification Of Stars, Galaxies And Qsos, Annika Lindh

Dissertations

This thesis reviews the current state of photometric classification in Astronomy and identifies two main gaps: a dependence on handcrafted rules, and a lack of interpretability in the more successful classifiers. To address this, Deep Learning and Computer Vision were used to create a more interpretable model, using unsupervised training to reduce human bias.

The main contribution is the investigation into the impact of using unsupervised feature-extraction from multi-wavelength image data for the classification task. The feature-extraction is achieved by implementing an unsupervised Deep Belief Network to extract lower-dimensionality features from the multi-wavelength image data captured by the Sloan Digital …

Learning With Scalability And Compactness, Wenlin Chen

McKelvey School of Engineering Graduate Student Theses & Dissertations

Artificial Intelligence has been thriving for decades since its birth. Traditional AI features heuristic search and planning, providing good strategy for tasks that are inherently search-based problems, such as games and GPS searching. In the meantime, machine learning, arguably the hottest subfield of AI, embraces data-driven methodology with great success in a wide range of applications such as computer vision and speech recognition. As a new trend, the applications of both learning and search have shifted toward mobile and embedded devices which entails not only scalability but also compactness of the models. Under this general paradigm, we propose a series …

Visualization Of Deep Convolutional Neural Networks, Dingwen Li

McKelvey School of Engineering Graduate Student Theses & Dissertations

Deep learning has achieved great accuracy in large scale image classification and scene recognition tasks, especially after the Convolutional Neural Network (CNN) model was introduced. Although a CNN often demonstrates very good classification results, it is usually unclear how or why a classification result is achieved. The objective of this thesis is to explore several existing visualization approaches which offer intuitive visual results. The thesis focuses on three visualization approaches: (1) image masking which highlights the region of image with high influence on the classification, (2) Taylor decomposition back-propagation which generates a per pixel heat map that describes each pixel's …

The Effect Of Hyperparameters In The Activation Layers Of Deep Neural Networks, Clay Lafayette Mcleod

Electronic Theses and Dissertations

Deep neural networks (DNNs), and artificial neural networks (ANNs) in general, have recently received a great amount of attention from both the media and the machine learning community at large. DNNs have been used to produce world-class results in a variety of domains, including image recognition, speech recognition, sequence modeling, and natural language processing. Many of most exciting recent deep neural network studies have made improvements by hardcoding less about the network and giving the neural network more control over its own parameters, allowing flexibility and control within the network. Although much research has been done to introduce trainable hyperparameters …

Applying Deep Learning Techniques To The Analysis Of Android Apks, Robin Andrew Nix

LSU Master's Theses

Malware targeting mobile devices is a pervasive problem in modern life and as such tools to detect and classify malware are of great value. This paper seeks to demonstrate the effectiveness of Deep Learning Techniques, specifically Convolutional Neural Networks, in detecting and classifying malware targeting the Android operating system. Unlike many current detection techniques, which require the use of relatively rigid features to aid in detection, deep neural networks are capable of automatically learning flexible features which may be more resilient to obfuscation. We present a parsing for extracting sequences of API calls which can be used to describe a …

Deep Learning For Just-In-Time Defect Prediction, Xinli Yang, David Lo, Xin Xia, Yun Zhang, Jianling Sun

Research Collection School Of Computing and Information Systems

Defect prediction is a very meaningful topic, particularly at change-level. Change-level defect prediction, which is also referred as just-in-time defect prediction, could not only ensure software quality in the development process, but also make the developers check and fix the defects in time. Nowadays, deep learning is a hot topic in the machine learning literature. Whether deep learning can be used to improve the performance of just-in-time defect prediction is still uninvestigated. In this paper, to bridge this research gap, we propose an approach Deeper which leverages deep learning techniques to predict defect-prone changes. We first build a set of …

An Evolutionary Approximation To Contrastive Divergence In Convolutional Restricted Boltzmann Machines, Ryan R. Mccoppin

Browse all Theses and Dissertations

Deep learning is an emerging area in machine learning that exploits multi-layered neural networks to extract invariant relationships from large data sets. Deep learning uses layers of non-linear transformations to represent data in abstract and discrete forms. Several different architectures have been developed over the past few years specifically to process images including the Convolutional Restricted Boltzmann Machine. The Boltzmann Machine is trained using contrastive divergence, a depth-first gradient based training algorithm. Gradient based training methods have no guarantee of reaching an optimal solution and tend to search a limited region of the solution space. In this thesis, we present …