In this paper, we present an approach in the Multimodal Learning Analytics field. Within this approach, we have developed a tool to visualize and analyze eye movement data collected during learning sessions in online ...
详细信息
ISBN:
(纸本)9798350376623;9798350376616
In this paper, we present an approach in the Multimodal Learning Analytics field. Within this approach, we have developed a tool to visualize and analyze eye movement data collected during learning sessions in online courses. The tool is named VAAD -an acronym for Visual Attention analysis Dashboard-. These eye movement data have been gathered using an eye-tracker and subsequently processed and visualized for interpretation. The purpose of the tool is to conduct a descriptive analysis of the data by facilitating its visualization, enabling the identification of differences and learning patterns among various learner populations. Additionally, it integrates a predictive module capable of anticipating learner activities during a learning session. Consequently, VAAD holds the potential to offer valuable insights into online learning behaviors from both descriptive and predictive perspectives.
The integrity and security of integrated circuits (ICs) are crucial in the digital world, as hardware Trojans (HTs) can allow unauthorized access and cause data breaches or malfunctions. Traditional HT detection appro...
详细信息
ISBN:
(纸本)9798350387186;9798350387179
The integrity and security of integrated circuits (ICs) are crucial in the digital world, as hardware Trojans (HTs) can allow unauthorized access and cause data breaches or malfunctions. Traditional HT detection approaches, sometimes considered similar to a "golden chip", have difficulties because of their covert nature and complicated designs. This study presents a machine learning-assisted approach for analyzing power side-channel data that overcomes these limitations. Our analysis shows the detection of HT accurately without the requirement for a golden reference by evaluating a largedataset containing different Trojan states. ML-assisted deep learning approach has significantly improved detection accuracy, providing a new direction for real-time HT monitoring and improving IC security throughout their lifecycle.
The proceedings contain 26 papers. The topics discussed include: understanding people’s needs in viewing diverse social opinions about controversial topics;a study of zooming, interactive lenses and overview+detail t...
ISBN:
(纸本)9798350321241
The proceedings contain 26 papers. The topics discussed include: understanding people’s needs in viewing diverse social opinions about controversial topics;a study of zooming, interactive lenses and overview+detail techniques in collaborative map-based tasks;parallel assemblies plot, a visualization tool to explore categorical and quantitative data: application to digital mobility outcomes;an empirical guide for visualization consistency in multiple coordinated views;visualizing interaction networks and evidence in biomedical corpora;efficient raycasting of volumetric depth images for remote visualization of large volumes at high frame rates;investigating animal infectious diseases with visual analytics;interactive transformations and visual assessment of noisy event sequences: an application in en-route air traffic control;and a visual analytics inspired approach to correlate and understand multiple mechanical tensor fields.
The automatic prediction of semantic maps, e.g. for land use/cover, is often addressed via supervised learning approaches that rely on large training datasets. Obtaining the required reference annotations for a large ...
详细信息
ISBN:
(纸本)9798350320107
The automatic prediction of semantic maps, e.g. for land use/cover, is often addressed via supervised learning approaches that rely on large training datasets. Obtaining the required reference annotations for a large amount of data is tedious and time consuming work, in particular for sensors such as synthetic aperture radar. Self-supervised learning aims to mitigate this issue by leveraging the large amount of usually available unlabelled data by constructing a pretext task that does not require annotations. We evaluate DINO, a recent self-supervised learning approach, and show that it outperforms segmentation models that have been initialized randomly or by weights obtained from pretraining on ImageNet. Furthermore, we show that the inclusion of multimodal data during the pretraining phase improves results significantly.
With the growing penetration of inverter-based distributed energy resources and increased loads through electrification, power systems analyses are becoming more important and more complex. Moreover, these analyses in...
详细信息
The traditional vicarious calibration methods for long-term radiometric trending of sensors are primarily based on invariant or pseudo-invariant sites on Earth, which typically rely on large spatially homogeneous area...
详细信息
ISBN:
(纸本)9798350360332;9798350360325
The traditional vicarious calibration methods for long-term radiometric trending of sensors are primarily based on invariant or pseudo-invariant sites on Earth, which typically rely on large spatially homogeneous areas and require delicate selection of samples. This study employs big dataanalysis of global pseudo-invariant pixels (PIPs) to monitor the long-term degradation of the Medium Resolution Spectral Imager (MERSI) onboard the FengYun(FY)-3 series satellites. Specifically, the PIPs are determined from image pairs acquired at different times by the iteratively re-weighted multivariate alteration detection (IR-MAD) technique. The global PIPs from several regions located in the middle to low latitude lands are obtained through the big dataanalysis of long-term time series data. This method requires no prior knowledge of the targets and extend the calibration sites to the pixel-level targets. Results of instrument degradation trending derived from the big dataanalysis of PIPs from different regions are consistent, which also align well with those obtained through traditional vicarious calibration methods. This method is not sensor-specific and can significantly enhance the frequency of vicarious calibration.
Investing capital requires cautious thought when navigating the stock market. analysis carried out throughout the project's design phase concentrated on handling large amounts of data and the effects of pattern re...
详细信息
General-purpose large language models (LLMs) have become a versatile tool in software development and maintenance. These models offer support in tasks such as understanding, writing, and summarizing code. While these ...
详细信息
ISBN:
(纸本)9798331528492;9798331528485
General-purpose large language models (LLMs) have become a versatile tool in software development and maintenance. These models offer support in tasks such as understanding, writing, and summarizing code. While these LLMs can generate code quickly their use in software modeling is relatively under-explored. This research aims to study ChatGPT's ability to generate class diagrams from source code. To do this, we engineered a prompt that takes in source code to create a UML class diagram for that system. We used this prompt to create class diagrams for 40 systems and assessed the diagrams on their correctness and structure. Our results show that ChatGPT creates class diagrams that accurately capture 90% of the classes and their attributes and 66% of the associations. While ChatGPT performed nearly flawlessly for smaller projects, the diagrams for larger projects had more issues. We conclude that ChatGPT is best utilized as a complementary tool rather than the sole resource for software modeling and maintenance.
A critical yet unresolved challenge in designing space-adaptive narratives for Augmented Reality (AR) is to provide consistently immersive user experiences anywhere, regardless of physical features specific to a space...
详细信息
A critical yet unresolved challenge in designing space-adaptive narratives for Augmented Reality (AR) is to provide consistently immersive user experiences anywhere, regardless of physical features specific to a space. For this, we present a comprehensive analysis on a series of user studies investigating how the size, density, and layout of real indoor spaces affect users playing Fragments, a space-adaptive AR detective game. Based on the studies, we assert that moderate levels of traversability and visual complexity afforded in counteracting combinations of size and complexity are beneficial for narrative experience. To confirm our argument, we combined the experimental data of the studies (n=112) to compare how five different spatial complexity conditions impact narrative experience when applied to contrasting room sizes. Results show that whereas factors of narrative experience are rated significantly higher in relatively simple settings for a small space, they are less affected by complexity in a large space. Ultimately, we establish guidelines on the design and placement of space-adaptive augmentations in location-independent AR narratives to compensate for the lack or excess of affordances in various real spaces and enhance user experiences therein.
Today, deep learning detectors for autonomous driving are delivering impressive results on public datasets and in real-world applications. However, these detectors require large amounts of data, especially labeled dat...
详细信息
ISBN:
(纸本)9798350348811;9798350348828
Today, deep learning detectors for autonomous driving are delivering impressive results on public datasets and in real-world applications. However, these detectors require large amounts of data, especially labeled data, to achieve the performance needed to ensure safe driving. The process of collecting and tagging data is expensive and cumbersome. Therefore, the recent focus of the industry has been on how to achieve similar performance while limiting the amount of labeled data required to train such models. Within the cross-modal active learning paradigm, we propose and analyze new strategies to exploit the inconsistencies between camera and LiDAR detectors to improve sampling efficiency and label only the samples that promise improvements for model training. For this, we leverage the 2D projection of the bounding boxes to equalize the output quality of camera and LiDAR detections. Finally, we achieve up to 0.6% AP improvement for camera and 2% improvement for LiDAR over random sampling on the KITTI dataset using a sampling strategy based on the number of detected objects.
暂无评论