Combining Implicit and Explicit Feature Extraction for Eye Tracking: Attention Classification Using a Heterogeneous Input

Statistical measurements of eye movement-specific properties, such as fixations, saccades, blinks, or pupil dilation, are frequently utilized as input features for machine learning algorithms applied to eye tracking recordings. These characteristics are intended to be interpretable aspects of eye gazing behavior. However, prior research has demonstrated that when trained on implicit representations of raw eye tracking data, neural networks outperform these traditional techniques. To leverage the strengths and information of both feature sets, we integrated implicit and explicit eye tracking features in one classification approach in this work. A neural network was adapted to process the heterogeneous input and predict the internally and externally directed attention of 154 participants. We compared the accuracies reached by the implicit and combined features for different window lengths and evaluated the approaches in terms of person- and task-independence. The results indicate that combining implicit and explicit feature extraction techniques for eye tracking data improves classification results for attentional state detection significantly. The attentional state was correctly classified during new tasks with an accuracy better than chance, and person-independent classification even outperformed person-dependently trained classifiers for some settings. For future experiments and applications that require eye tracking data classification, we suggest to consider implicit data representation in addition to interpretable explicit features.

A Measurement Framework for Explicit and Implicit Urban Traffic Sensing

Urban traffic sensing has been investigated extensively by different real-time sensing approaches due to important applications such as navigation and emergency services. Basically, the existing traffic sensing approaches can be classified into two categories by sensing natures, i.e., explicit and implicit sensing. In this article, we design a measurement framework called EXIMIUS for a large-scale data-driven study to investigate the strengths and weaknesses of two sensing approaches by using two particular systems for traffic sensing as concrete examples. In our investigation, we utilize TB-level data from two systems: (i) GPS data from five thousand vehicles, (ii) signaling data from three million cellphone users, from the Chinese city Hefei. Our study adopts a widely used concept called crowdedness level to rigorously explore the impacts of contexts on traffic conditions including population density, region functions, road categories, rush hours, holidays, weather, and so on, based on various context data. We quantify the strengths and weaknesses of these two sensing approaches in different scenarios and then we explore the possibility of unifying two sensing approaches for better performance by using a truth discovery-based data fusion scheme. Our results provide a few valuable insights for urban sensing based on explicit and implicit data from transportation and telecommunication domains.

Collaborative Filtering Technical Comparison in Implicit Data

Recommendation systems have become a necessity due to the mass of information accumulated for each site. For this purpose, there are several methods including collaborative filtering and content-based filtering. For each approach there is a vast list of procedural choices. The work studies the different methods and algorithms in the field of collaborative filtering recommendation. The objective of the work is to implement these algorithms in order to compare the different performances of each one; the tests were carried out in two datasets, book crossing and Movieslens. The use of a data set benchmark is crucial for the proper evaluation of collaborative filtering algorithms in order to draw a conclusion on the performance of the algorithms.

A Hybrid Recommendation System for Video Games: Combining Content-based & Collaborative Filtering

Recommending video games can be trickier than movies. When it comes to selecting a video game, many factors are involved such as its genre, platform on which it’s played, duration of main and side quests, and more. However, recommending games based on just these features won’t suffice as a person who, for example, enjoys a certain genre of game can equally enjoy a vastly different genre. Therefore, a scoring mechanism is required which takes into account both, features of a game (contentbased filtering) and also studies the buying patterns of people playing a particular game (collaborative filtering). In this paper I have proposed a way to take into account both content-based and collaborative filtering into the final recommendation. I have used cosine similarity to quantify the similarity between the features of games. Along with this, I have employed a Deep fullyconnected AutoEncoder (DAE) to generalize the implicit data representation of an user’s buying patterns. Finally, I present a novel approach to combine the scores of these filtering techniques in such a way that it gives equal weightage to both. In other words, they both have equal influence over the final list of the top 10 games recommended to the user. Keywords: Hybrid Recommender, Collaborative filtering, Content-based filtering, Cosine similarity, AutoEncoder.

Force Prediction in Robotized Ultrasonic Sheet Metal Forming Based on Finite Element Analysis and Artificial Neural Networks

The forces acting in the process of single point incremental forming (SPIF) change the geometry of the sheet metal. The tool-workpiece interaction process is non-linear due to the large deformations of the sheet metal, which determine the plastic behaviour, as well as the evolutionary boundary conditions resulting from the contact between the tool and the sheet. Instead of lubricating the contact surface of the forming tool and the sheet metal, an innovative environmentally friendly method to reduce the coefficient of friction by vibrating the sheet has been proposed. The finite element method (FEM) allowed a virtual evaluation of the deformation parameters of the SPIF process in order to determine the destructive loads. The FEM was chosen as a deterministic numerical tool to evaluate the set of defect parameters induced by forming forces. The paper also proposes a method for predicting the formation force using an artificial neural network (ANN), assuming that such a model is generalized to implicit data. In this context, an empirical analysis of the implementation of the ANN technique is performed.

Temporal Fluctuations in Context Ethnic Diversity Over Three Decades Predict Implicit National Inclusion of Asian Americans

The present research examined whether temporal fluctuations in context ethnic diversity account for current levels of implicit ethnic-American associations. Temporal fluctuations in ethnic diversity at the metropolitan level were assessed using data from four decennial U.S. censuses (1980-2010) and distinguishing three dimensions of context ethnic diversity (minority representation, variety, and integration). Project Implicit data (2011-2017) indexed the extent to which American identity was implicitly associated with European Americans over Asian Americans (i.e., American = White associations). Data were analyzed using multilevel modeling (N = 152,011, nested within 226 metropolitan areas). Steeper increases in the proportion of Asian Americans were related to weaker implicit (but stronger explicit) American = White associations. Increases in ethnic integration accounted for stronger implicit American = White associations when integration fluctuations reflected accelerating rather than decelerating trends. These results suggest that current levels of implicit ethnic-national associations are linked to complex patterns of ethnic diversity fluctuations.

Detecting and Classifying Typhoon Information from Chinese News Based on a Neural Network Model

Typhoons are major natural disasters in China. Much typhoon information is contained in a large number of network media resources, such as news reports and volunteered geographic information (VGI) data, and these are the implicit data sources for typhoon research. However, two problems arise when using typhoon information from Chinese news reports. Since the Chinese language lacks natural delimiters, word segmentation error results in trigger mismatches. Additionally, the polysemy of Chinese affects the classification of triggers. Second, there is no authoritative classification system for typhoon events. This paper defines a classification system for typhoon events, and then uses the system in a neural network model, lattice-structured bidirectional long–short-term memory with a conditional random field (BiLSTM-CRF), to detect these events in Chinese online news. A typhoon dataset is created using texts from the China Weather Typhoon Network. Three other datasets are generated from general Chinese web pages. Experiments on these four datasets show that the model can tackle the problems mentioned above and accurately detect typhoon events in Chinese news reports.

PandAna: A Python Analysis Framework for Scalable High Performance Computing in High Energy Physics

Modern experiments in high energy physics analyze millions of events recorded in particle detectors to select the events of interest and make measurements of physics parameters. These data can often be stored as tabular data in files with detector information and reconstructed quantities. Most current techniques for event selection in these files lack the scalability needed for high performance computing environments. We describe our work to develop a high energy physics analysis framework suitable for high performance computing. This new framework utilizes modern tools for reading files and implicit data parallelism. Framework users analyze tabular data using standard, easy-to-use data analysis techniques in Python while the framework handles the file manipulations and parallelism without the user needing advanced experience in parallel programming. In future versions, we hope to provide a framework that can be utilized on a personal computer or a high performance computing cluster with little change to the user code.