Search

While several methods for predicting uncertainty on deep networks have been recently proposed, they do not readily translate to large and complex datasets. In this paper we utilize a simplified form of the Mixture Density Networks (MDNs) to produce a one-shot approach to quantify uncertainty in regression problems. We show that our uncertainty bounds are on-par or better than other reported existing methods. When applied to standard regression benchmark datasets, we show an improvement in predictive log-likelihood and root-mean-square-error when compared to existing state-of-the-art methods. We also demonstrate this method's efficacy on stochastic, highly volatile time-series data where stock prices are predicted for the next time interval. The resulting uncertainty graph summarizes significant anomalies in the stock price chart. Furthermore, we apply this method to the task of age estimation from the challenging IMDb-Wiki dataset of half a million face images. We successfully predict the uncertainties associated with the prediction and empirically analyze the underlying causes of the uncertainties. This uncertainty quantification can be used to pre-process low quality datasets and further enable learning.

As with many problems solved by deep neural networks, existing solutions rarely explain, precisely, the important factors responsible for the predictions made by the model. This work looks to investigate how different spatial regions and landmark points change in position over time, to better explain the underlying factors responsible for various facial emotion expressions. By pinpointing the specific regions or points responsible for the classification of a particular facial expression, we gain better insight into the dynamics of the face when displaying that emotion. To accomplish this, we examine two spatiotemporal representations of moving faces, while expressing different emotions. The representations are then presented to a convolutional neural network for emotion classification. Class activation maps are used in highlighting the regions of interest and the results are qualitatively compared with the well known facial action units, using the facial action coding system. The model was originally trained and tested on the CK+ dataset for emotion classification, and then generalized to the SAMM dataset. In so doing, we successfully present an interpretable technique for understanding the dynamics that occur during convolutional-based prediction tasks on sequences of face data.

Expression neutralization is the process of synthetically altering an image of a face so as to remove any facial expression from it without changing the face’s identity. Facial expression neutralization could have a variety of applications, particularly in the realms of facial recognition, in action unit analysis, or even improving the quality of identification pictures for various types of documents. Our proposed model, StoicNet, combines the robust encoding capacity of variational autoencoders, the generative power of generative adversarial networks, and the enhancing capabilities of super resolution networks with a learned encoding transformation to achieve compelling expression neutralization, while preserving the identity of the input face. Objective experiments demonstrate that StoicNet successfully generates realistic, identity-preserved faces with neutral expressions, regardless of the emotion or expression intensity of the input face.

Emotion regulation can be characterized by different activities that attempt to alter an emotional response, whether behavioral, physiological or neurological. The two most widely adopted strategies, cognitive reappraisal and expressive suppression are explored in this study, specifically in the context of disgust. Study participants (N = 21) experienced disgust via video exposure, and were instructed to either regulate their emotions or express them freely. If regulating, they were required to either cognitively reappraise or suppress their emotional experiences while viewing the videos. Video recordings of the participants’ faces were taken during the experiment and electrocardiogram (ECG), electromyography (EMG), and galvanic skin response (GSR) readings were also collected for further analysis. We compared the participants behavioral (facial musculature movements) and physiological (GSR and heart rate) responses as they aimed to alter their emotional responses and computationally determined that when responding to disgust stimuli, the signals recorded during suppression and free expression were very similar, whereas those recorded during cognitive reappraisal were significantly different,. Thus, in the context of this study, from a signal analysis perspective, we conclude that emotion regulation via cognitive reappraisal significantly alters participants’ physiological responses to disgust, unlike regulation via suppression.

To cope with the data requirements associated with networks of the future, heterogeneous network architectures comprising of a coalition of low powered access nodes with macro cells have been proffered as a solution. Most Mobile Network Operators (MNO) are starting to upgrade their infrastructure to include cloud-Radio Access (C-RAN) services to support higher processing capacities, ease in network scalability and efficient resource management. These upgrades to the network architecture come with their unique challenges in the form of user associations to cells, efficient load balancing among multi-tiered cells, as well as Remote Radio Head (RRH) to Base Band Unit (BBU) association and re-association. In this paper we attempt to address and proffer a solution by proposing a joint optimisation algorithm consisting of two stages: the first being an upgrade to user association by way of our novel load balancing/user association scheme, and the second being in the form of an exact potential game among RRHs to associate to available BBU pool resources optimally, with the help of a centralized controller.

Research in social psychology has extensively shown that in cohesive groups, individuals often mirror each other’s prosody, facial expressions, and body movements. This mirroring effect can help determine the level of comfort or the extent of engagement and genuine interest between two or more interlocutors. In this work, using an annotated dataset consisting of videos of three-person conversations, we aim to analyze the extent of rapport in each of the triadic groups. We generate behavioral curves from features extracted from the participants’ face and body movements. These are the sampled time series signals resulting from their multimodal features. Next, the extents of synchrony are analyzed by aligning the behavioral curves of pairs of participants. The alignment tests show that basic correlation coefficient measures outperform more advanced curve matching techniques when used to estimate the similarities between multidimensional behavior curves. They also show that in this dataset, synchrony is better observed from facial expressions than body movements. For this reason, using facial action units, we show that an end-to-end recursive neural network (RNN) trained using a regression loss yields good results in predicting the extent of synchrony in small groups.

In this paper we propose a three-way partitioned resource reservation approach to deal with congestion in heterogeneous cellular networks as a result of densely populated mixed traffic scenarios analogous with the Fifth Generation (5G) networks which will host Factories-of-the-Future (FoF) environments. Our approach lends to the reduction in random access delay budget through a combined static/dynamic resource usage approach with respect to traffic types with different latency requirements. In addition, we obtain figures for optimum performance random access (RA) resource partitioning for ultra Reliable Low Latency (uRLLC) and non-uRLLC traffic.

In-home monitoring technologies have the potential to transform the healthcare system. With the increasing number and variety of connected devices that can be accessed by multiple users, the need for seamless authentication and identification mechanisms is greater now than ever. Seamless, non-interruptive mechanisms are required to personalize these devices and non-intrusive biometric technologies are gaining popularity as an answer to this automated personalization challenge. In this work, we evaluate the biometric efficacies of a fully integrated toilet (FIT) seat, designed for monitoring a subjects cardiac health parameters such as electrocardiogram (ECG), ballistocardiogram (BCG) and weight measures, all in the home without any change in daily habits. We assess the system on data obtained from 22 healthy subjects with measurements taken over an 8-week period, in order to obtain the optimal combination of features, classifiers and enrollment record size. We found the multi-class SVM classifier along with all features extracted from the raw measurements to be the best performing combination. This combination produced an area under curve (AUC) value of 0.86. When tested for its person identification performance, we obtained an average accuracy of 75%. We have therefore demonstrated that the biometric capacities of the FIT seat shows a strong potential for deployment in real-life, multi-user, in-home settings.

Video streaming services are prominent in people’s lives and there is a need for improved video recommendation systems that adapt to their users in a personalized way. This project uses affective computing and non-invasive sensing to address this issue. Our objective is to develop an approach that uses the viewer’s emotional reactions as the basis for recommending new content. To achieve this goal, we must first understand how viewers react to videos. We conducted a study where subjects’ facial expressions and skin-estimated pulse were monitored while watching videos. Results showed that our approach can estimate dominant emotions 70% of the time. We found no correlation between the number of emotional reactions people have and how they rate the videos they watch. The pulse estimation is reliable to measure important changes in pulse, however it can still be improved.

Medical diagnosis requires extensive knowledge of pathological characteristics and the clinical training necessary to master specific domains. In the domain of dermatology, these properties include the size, shape, distribution, color, and location of symptoms. We have built and trained a tag recommender with expert narratives of dermatological disease, then explored the effectiveness of this system on novice narratives. The system mitigates the prerequisite of domain knowledge, empowering a novice to enhance their medical descriptions or even reach an accurate diagnosis. After collecting novice narratives, we explored word alignment to provide a mapping between expert and novice vocabulary allowing novice input to be augmented with expert terminology. Ultimately, we found that our system is an effective educational tool, which could be improved by word alignment and other techniques.

In recent years there has been a growing interest in digitizing the extensive amounts of books and documents that existed preceding the widespread adoption of digital technologies. Many of these digitizing initiatives deal with huge collections of handwritten documents, for which document image analysis techniques (page segmentation, keyword-spotting, optical character recognition (OCR), etc) are not yet as mature as for printed text. Thus, there is an imminent need to develop techniques to understand, archive, index and search the manuscripts. The antiquated approach of manually transcribing handwritten collections and then using standard text retrieval techniques can be very expensive for large collections. But many of the manuscripts in these collections, unlike machine-printed texts, contain unstructured information, cluttered group of texts and graphics that do not necessarily follow a pre-specified format, thus making it quite challenging to automatically process. Thus, in this paper we present a convolutional neural network (CNN) based implementation that is used to segment pages of handwritten documents into their constituent sections. We showcase a multiscale sliding window based network that is trained to predict the sections of the pages in handwritten manuscripts. The results of the network are post-processed with a novel region growing technique to further improve the segmentation results. The implementation is applied on the Marianne Moore archival collection, a body of handwritten notes and memos by the renowned author Marianne Moore (1887–1972), one of the foremost modernist poets of the early twentieth-century. We present our segmentation results both quantitatively and qualitatively.

Information graphics, such as bar charts, graphs, plots etc. in scientific documents primarily facilitate better understanding of information. Graphics are a key component in technical documents as they are simplified representations of complex ideas. When the traditional optical character recognition (OCR) systems is used on digitized documents, we lose the ideas conveyed in these information graphics since OCRs typically work only on text. And although in more recent times, tools have been developed to extract information graphics from pdf files, they still do not intelligently interpret the contents of the extracted graphics. We therefore propose a method for identifying the intended messages of line plots using a Bayesian network. We accomplish this by first extracting a dense set of points in from a line plot and then represent the entire line plot as a sequence of trends. We then implement a Bayesian network for reasoning about the messages conveyed by the line plots and their trends. We validate our approach by performing experiments on a dataset obtained from computer science conference publications and evaluate the performance of the network against the messages generated by human end users. The resulting intended message gives holistic information about the line plot(s) as well as lower level information about the trends that make up the plot.

Information graphics, such as graphs and plots, are used in technical documents to convey information to humans and to facilitate greater understanding. Usually, graphics are a key component in a technical document, as they enable the author to convey complex ideas in a simplified visual format. However, in an automatic text recognition system, which are typically used to digitize documents, the ideas conveyed in a graphical format are lost. We contend that the message or extracted information can be used to help better understand the ideas conveyed in the document. In scientific papers, line plots are the most commonly used graphic to represent experimental results in the form of correlation present between values represented on the axes. The contribution of our work is in the series of image processing algorithms that are used to automatically extract relevant information, including text and plot from graphics found in technical documents. We validate the approach by performing the experiments on a dataset of line plots obtained from scientific documents from computer science conference papers and evaluate the variation of a reconstructed curve from the original curve. Our algorithm achieves a classification accuracy of 91% across the dataset and successfully extracts the axes from 92% of line plots. Axes label extraction and line curve tracing are performed successfully in about half the line plots as well.

Dissemination of malicious code, also known as malware, poses severe challenges to cyber security. Malware authors embed software in seemingly innocuous executables, unknown to a user. The malware subsequently interacts with security-critical OS resources on the host system or network, in order to destroy their information or to gather sensitive information such as passwords and credit card numbers. Malware authors typically use Application Programming Interface (API) calls to perpetrate these crimes. We present a model that uses text mining and topic modeling to detect malware, based on the types of API call sequences. We evaluated our technique on two publicly available datasets. We observed that Decision Tree and Support Vector Machine yielded significant results. We performed t-test with respect to sensitivity for the two models and found that statistically there is no significant difference between these models. We recommend Decision Tree as it yields ‘if-then’ rules, which could be used as an early warning expert system.

This paper discusses methods for using measurements of the backscattered power and the Doppler shift of radar signals scattered from the ocean surface to compute maps of phase resolved ocean wave fields. Results are compared with buoy and lidar measurements of ocean surface waves off the coast of southern California.

In this paper we describe the development of a low cost, high power coherent-on-receiver radar. The unit has a 25kW peak transmit power and is capable of accurately measuring velocity as well as range. The design of the radar is optimized for marine surveillance, although the techniques developed have general application. The radar system is validated through a series of tests that culminate with sea surface measurements.