Abstract

Keeping time is fundamental for our everyday existence. Various isochronous activities, such as locomotion, require us to use internal timekeeping. This phenomenon comes into play also in other human pursuits such as dance and music. When listening to music, we spontaneously perceive and predict its beat. The process of beat perception comprises both beat inference and beat maintenance, their relative importance depending on the salience of beat in the music. To study functional connectivity associated with these processes in a naturalistic situation, we used functional magnetic resonance imaging to measure brain responses of participants while they were listening to a piece of music containing strong contrasts in beat salience. Subsequently, we utilized dynamic graph analysis and psychophysiological interactions (PPI) analysis in connection with computational modelling of beat salience to investigate how functional connectivity manifests these processes. As the main effect, correlation analyses between the obtained dynamic graph measures and the beat salience measure revealed increased centrality in auditory-motor cortices, cerebellum, and extrastriate visual areas during low beat salience, whereas regions of the default mode- and central executive networks displayed high centrality during high beat salience. PPI analyses revealed partial dissociation of functional networks belonging to this pathway indicating complementary neural mechanisms crucial in beat inference and maintenance, processes pivotal for extracting and predicting temporal regularities in our environment.

Abstract

Variational autoencoders were proven successful in domains such as computer vision and speech processing. Their adoption for modeling user preferences is still unexplored, although recently it is starting to gain attention in the current literature. In this work, we propose a model which extends variational autoencoders by exploiting the rich information present in the past preference history. We introduce a recurrent version of the VAE, where instead of passing a subset of the whole history regardless of temporal dependencies, we rather pass the consumption sequence subset through a recurrent neural network. At each time-step of the RNN, the sequence is fed through a series of fully-connected layers, the output of which models the probability distribution of the most likely future preferences. We show that handling temporal information is crucial for improving the accuracy of the VAE: In fact, our model beats the current state-of-the-art by valuable margins because of its ability to capture temporal dependencies among the user-consumption sequence using the recurrent encoder still keeping the fundamentals of variational autoencoders intact.

Abstract

Clustering has become a very popular machine learning technique for identifying groups of data points with common features in a set of data points. In several applications, there is a need to explain the clusters so that the user can understand the underlying commonalities. One such application is in the area of building energy simulation. There is a need to cluster solutions obtained by parametric energy simulation runs and explain the characteristics of each cluster for human consumption. This paper demonstrates how the axis-aligned hyper-rectangles based clustering, on building energy simulation data, can help identify clusters and describe the governing rules for each cluster. We are calling these rules design strategies. Instead of the distance-based clustering methods that are unable to extract simple rules from the underlying commonalities in each cluster, this method is able to overcome this limitation. This method is applied to identify design strategies from a parametric run of a simple five-zone rectangular building model. Based on a user-given threshold, low energy solutions are selected for clustering. Each axis-aligned hyper-rectangle cluster is a unique design strategy that can be easily communicated to the user.

Abstract

This paper describes our system (Fermi) for Task 6: OffensEval: Identifying and Categorizing Offensive Language in Social Media of SemEval-2019. We participated in all the three sub-tasks within Task 6. We evaluate multiple sentence embeddings in conjunction with various supervised machine learning algorithms and evaluate the performance of simple yet effective embedding-ML combination algorithms. Our team Fermi’s model achieved an F1-score of 64.40%, 62.00% and 62.60% for sub-task A, B and C respectively on the official leaderboard. Our model for sub-task C which uses pre-trained ELMo embeddings for transforming the input and uses SVM (RBF kernel) for training, scored third position on the official leaderboard. Through the paper we provide a detailed description of the approach, as well as the results obtained for the task.

Abstract

This paper describes our system (Fermi) for Task 5 of SemEval-2019: HatEval: Multilingual Detection of Hate Speech Against Immigrants and Women on Twitter. We participated in the subtask A for English and ranked first in the evaluation on the test set. We evaluate the quality of multiple sentence embeddings and explore multiple training models to evaluate the performance of simple yet effective embedding-ML combination algorithms. Our team-Fermi’s model achieved an accuracy of 65.00% for English language in task A. Our models, which use pretrained Universal Encoder sentence embeddings for transforming the input and SVM (with RBF kernel) for classification, scored first position (among 68) in the leaderboard on the test set for Subtask A in English language. In this paper we provide a detailed description of the approach, as well as the results obtained in the task.

Abstract

Online Community Question Answering Forums (cQA) have gained massive popularity within recent years. The rise in users for such forums have led to the increase in the need for automated evaluation for question comprehension and fact evaluation of the answers provided by various participants in the forum. Our team, Fermi, participated in sub-task A of Task 8 at SemEval 2019-which tackles the first problem in the pipeline of factual evaluation in cQA forums, ie, deciding whether a posed question asks for a factual information, an opinion/advice or is just socializing. This information is highly useful in segregating factual questions from non-factual ones which highly helps in organizing the questions into useful categories and trims down the problem space for the next task in the pipeline for fact evaluation among the available answers. Our system uses the embeddings obtained from Universal Sentence Encoder combined with XGBoost for the classification sub-task A. We also evaluate other combinations of embeddings and off-the-shelf machine learning algorithms to demonstrate the efficacy of the various representations and their combinations. Our results across the evaluation test set gave an accuracy of 84% and received the first position in the final standings judged by the organizers.

Abstract

Resting-state functional connectivity (FC) analyses have shown atypical connectivity in autism spectrum disorder (ASD) as compared to typically developing (TD). However, this view emerges from investigating static FC overlooking the whole brain transient connectivity patterns. In our study, we investigated how age and disease influence the dynamic changes in functional connectivity of TD and ASD. We used resting-state functional magnetic resonance imaging (rs-fMRI) data stratified into three cohorts: children (7–11 years), adolescents (12–17 years), and adults (18+ years) for the analysis. The dynamic variability in the connection strength and the modular organization in terms of measures such as flexiblity, cohesion strength, and disjointness were explored for each subject to characterize the differences between ASD and TD. In ASD, we observed significantly higher inter-subject dynamic variability in connection strength as compared to TD. This hyper-variability relates to the symptom severity in ASD. We also found that whole-brain flexibility correlates with static modularity only in TD. Further, we observed a core-periphery organization in the resting-state, with Sensorimotor and Visual regions in the rigid core; and DMN and attention areas in the flexible periphery. TD also develops a more cohesive organization of sensorimotor areas. However, in ASD we found a strong positive correlation of symptom severity with flexibility of rigid areas and with disjointness of sensorimotor areas. The regions of the brain showing high predictive power of symptom severity were distributed across the cortex, with stronger bearings in the frontal, motor, and occipital cortices. Our study demonstrates that the dynamic framework best characterizes the variability in ASD.

Abstract

Learning a forward mapping that relates stimuli to the corresponding brain activation measured by functional magnetic resonance imaging (fMRI) is termed as estimating encoding models. Computational tractability usually forces current encoding as well as decoding solutions to typically consider only a small subset of voxels from the actual 3D volume of activation. Further, while reconstructing stimulus information from brain activation (brain decoding) has received wider attention, there have been only a few attempts at constructing encoding solutions in the extant neuro-imaging literature. In this paper, we present StepEncog, a convolutional LSTM autoencoder model trained on fMRI voxels. The model can predict the entire brain volume rather than a small subset of voxels, as presented in earlier research works. We argue that the resulting solution avoids the problem of devising encoding models based on a rulebased selection of informative voxels and the concomitant issue of wide spatial variability of such voxels across participants. The perturbation experiments indicate that the proposed deep encoder indeed learns to predict brain activations with high spatial accuracy. On challenging universal decoder imaging datasets, our model yielded encouraging results.

Abstract

Use of virtual reality (VR) technology is proliferating for designing and upgrading entertainment devices, and creating virtual environments that could be used for research and training. VR is becoming a strong research tool by providing a tighter control on the experimental environment and by allowing almost limitless possibilities of creating ecologically valid stimuli. However, the enhanced fidelity between the real and virtual worlds that VR provides does not always benefit human performance. For a better understanding, and increasing VR’s usability, we need to identify the relevant constituent components of immersive technologies, and differentiate their roles, for example, how visual and interaction fidelity differentially improves human performance. We conducted an experiment to examine how two common VR display modes, head mounted display (HMD) and desktop (DT), would affect spatial learning when we restrict ambulatory locomotion in HMD. This manipulation allowed examining the role of varying visual fidelity with low interaction fidelity. We used a between-group design with 40 naïve participants. They explored a virtual environment and later drew its sketch-map. Our results showed participants spent more time and perceived less motion-sickness and task effort using desktop than HMD VR. With reduced interaction fidelity, the high visual fidelity of HMD as compared to desktop resulted in similar or poorer performance on different spatial learning tasks after accounting for motion-sickness and workload effort. Participants were better in recalling spatial components related to junction and cyclic order of the navigated virtual space in …

Abstract

The aim of the current study is to propose acoustic features for detection of nasals in continuous speech. Acoustic features that represent certain characteristics of speech production are extracted. Features representing excitation source characteristcs are extracted using zero frequency filtering method. Features representing vocal tract system characteristics are extracted using zero time windowing method. Feature sets are formed by combining certain subsets of the features mentioned above. These feature sets are evaluated for their representativeness of nasals in continuous speech in three different languages, namely, English, Hindi and Telugu. Results show that nasal detection is reliable and consistent across all the languages mentioned above.