Find research datasets worth reusing

opencc-by-4.0Feb 2020View details →

An Open-set Recognition and Few-Shot Learning Dataset for Audio Event Classification in Domestic Environments

The problem of training a deep neural network with a small set of positive samples is known as few-shot learning (FSL). It is widely known that traditional deep learning (DL) algorithms usually show very good performance when trained with large datasets. However, in many applications, it is not possible to obtain such a high number of samples. In the image domain, typical FSL applications are those related to face recognition. In the audio domain, music fraud or speaker recognition can be clearly benefited from FSL methods. This paper deals with the application of FSL to the detection of specific and intentional acoustic events given by different types of sound alarms, such as door bells or fire alarms, using a limited number of samples. These sounds typically occur in domestic environments where many events corresponding to a wide variety of sound classes take place. Therefore, the detection of such alarms in a practical scenario can be considered an open-set recognition (OSR) problem. To address the lack of a dedicated public dataset for audio FSL, researchers usually make modifications on other available datasets. This paper is aimed at providing the audio recognition community with a carefully annotated dataset for FSL and OSR comprised of 1360 clips from 34 classes divided into pattern sounds and unwanted sounds. To facilitate and promote research in this area, results with two baseline systems (one trained from scratch and another based on transfer learning), are presented.

opencc-by-4.0Jul 2022View details →

Improving students' privacy awareness – Analysis of a pilot survey to design a VR environment for self-paced learning

In this research, we measured the knowledge of students at the University of Debrecen in the field of data privacy awareness, online and password security. Description <ul> <li>In the questionnaire, green-highlighted answer signs the correct answer to each question.</li> <li>Total data set contains the answers to each question and the respondent's age.</li> <li>Correct/incorrect data set contains information if the answer is correct to each question, and it also contains the respondent's age. One means the answer was correct, and zero means the answer was incorrect.</li> </ul>

opencc-by-4.0Jun 2019View details →

User stories and xAPI statements for "A mobile campus application as a sensor node for Personal Learning Environments"

This dataset provides the full user stories and xAPI statements as used in the prototype described in the article "A mobile campus application as a sensor node for Personal Learning Environments". It consists of two PDF documents described below. The files were created as part of the master thesis of Hendrik Geßner. "User Stories.pdf" contains a complete list of user stories with required context information, existing portlets and a category. The process that led to this collection is described very briefly in the article mentioned above, a graphical explanation is available in the attached image "Use case process complete.jpg" "xAPI Statements.pdf" contains all xAPI statements used in the prototype described in the article mentioned above. Dynamic elements such as names or IDs are highlighted on color. The statements appear in the following order: attended, used, loggedin, wasat, opened, closed, joined, left.

opencc-by-4.0Aug 2024View details →

User survey data of learning environment eTKMY3, Turku School of Economics, Finland.

User survey data of learning environment (eTKMY3) of an introduction to statistics course, Turku School of Economics.  Variables question_11_row_1                  Starting eTKMY3 was difficult question_11_row_3                  eTKMY3 is a successful system question_11_row_4                  I can manage my studying and exercises easily question_11_row_5                  Navigation was easy question_11_row_7                  eTKMY3 was complex question_12_row_8                  individual starting values of most exercises as a good way to promote independent working Observations: Students of Turku School of Economics taking the course "TKMY3 Introduction of Statistics", Spring 2024.

opencc-by-4.0Oct 2024View details →

Structured Power Grid Simulation Dataset for Machine Learning: Failure and Survival Events in Grid2Op's L2RPN WCCI 2022 Environment

This dataset was developed for and used in the paper titled "Fault Detection for Agents in Power Grid Topology Optimization: A Comprehensive Analysis" by Malte Lehna, Mohamed Hassouna, Dmitry Degtyar, Sven Tomforde, and Christoph Scholz, presented at the Workshop on Machine Learning for Sustainable Power Systems (ML4SPS), part of ECML PKDD 2024. While the paper is pending formal publication, a preprint version is available on arXiv. The dataset contains structured training, validation, and test data comprising failure and survival events observed in transmission power grid simulations. These were generated using Grid2Op with the WCCI 2022 L2RPN environment. Each data instance is labeled with one of four classes, representing survival or impending failure in 1, 3, and 5 timesteps. This dataset was used to train, validate and test machine learning models that predict grid agent failures in topology optimization tasks.

opencc-by-4.0Feb 2022View details →

What can we learn about transients from the 21 cm line emission in their environments?

I present the result of an observational campaign to map the 21 cm atomic hydrogen line in host galaxies of supernovae (SNe), long gamma-ray bursts (GRBs), and fast radio bursts (FRBs). For all analysed hosts of type Ic-BL SNe and GRBs we found off-centre gas concentrations close to the GRB/SN positions and irregular velocity fields. This suggests a recent gas inflow. This is consistent with GRB/SN progenitors being born when a galaxy accretes metal-poor gas from the intergalactic medium, and opens a possibility to use GRB/SN hosts to study gas accretion. This supports a very massive model for their progenitors. Similarly, the host galaxies of FRBs exhibit very asymmetric 21 cm lines, suggesting a connection between the birth of FRB progenitors and galaxy interactions. On the other hand, the host galaxy of the unusual transient AT 2018cow (the first fast blue optical transient, FBOT) does not exhibit such features, which suggests that its progenitor may not have been a massive star.

BRAIN Journal-Developing Distance Learning Environments in the Context of Cross-Border Cooperation-Figure 5. Weekly session stats for 2016

On following figures, statistic usage is given as Monthly and Weekly statistic for individual users and sessions. The number of the individual users and of the sessions is far better for the 2015 period, especially for the extent of time until the end of June. This is reasonable because it was during the project and the site is frequently put forward in promotional conferences, in press, in direct contacts with schools, and companies. After that period the site was not promoted additionally, and the only pointer to the site is a number of links found on our institutional websites as one of the services we are offering to the students. Considering all that, the figures for 2015 and even 2016, seems to be quite satisfactory. (Figure 1, Figure 4, and Figure 5).

BRAIN Journal-Developing Distance Learning Environments in the Context of Cross-Border Cooperation-Figure 3. Monthly user/session stats for 2016

It is also interesting to observe that in Figure 3, where the usage for 2016 is presented, that in the non-promotional period, the site is mostly used in January (before January exam term), in April, May, and June (before the June exam session and during colloquial exams) or in July and August (before the September exam session).

BRAIN Journal-Developing Distance Learning Environments in the Context of Cross-Border Cooperation-Figure 2. Monthly user/session stats for 2015

BRAIN Journal-Developing Distance Learning Environments in the Context of Cross-Border Cooperation-Figure 1. Components of EduWebCast System

The aim of this partnership would be to implement an infrastructure for live and on-demand video streaming of learning material for the targeted groups and, to this purpose, to establish a long and fruitful cooperation between teachers, pupils, and students on both sides of the border. The joint creation and administration of the webcast project is the ground stone of the partnership between the two universities and will result in more common projects based on the materials obtained through the project, contests between pupils and students, possible periodic educational exchanges.

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 10. Room model generated with Autodesk 123D Catch - the 3D model (screen capture from GLC Player)

Structure from motion was used for rapid modeling of a small room with all its objects. Two files were generated, a Wavefront obj and mtl (corresponding to the texture). The 3D model was post-processed with MeshLab, during which several filters were applied to clean up the model. The mesh model was also connected with the scanned model, by choosing at least 4 connection points. The 2D and 3D results are shown in Figures 9, 10. A post-processing could also be performed using the Autodesk 123D Catch web application.

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 2. The model of the 3D virtual campus - details from the building interior (3D modeling by Marius Hodea)

The processing workflow for 3D modeling and design for a 3DVLE represents a time- consuming stage in the overall pipeline production. One reason is that a range of technologies and tools are typically used. In (Cudworth 2014) a 3-week period is indicated for experienced users to perform the 3D modeling of a virtual space. In our case, a 3-month work was needed for designing a working model of a 3D virtual campus (see Figure 1 and Figure 2 for final results).

opencc-by-4.0Dec 2017View details →

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 1. The model of the 3D virtual campus - an outdoor view (3D modeling by Marius Hodea)

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 9. Room model generated with Autodesk 123D Catch - the 2D model

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 8. The 3D model of the faculty building in a HTML page

The pipeline processing was the following: a) the 3D model from Sketchup was saved as a Collada file; b) this file has been imported in MeshLab (MESHLAB 2017) and converted to VRML97 format (wrl); c) aopt utility was used to convert wrl files to X3D and HTML5 files. The model was visualized in the OpenSim virtual world setting using an external browser (Figure 8).

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 5. Iteration of the 3D modeling of the virtual faculty building (3D modeling by Marius Hodea)

For our research, several iterations and methods were employed for the 3D design of an online campus.  Different virtual models of a faculty building (see Figures 4,5) were designed and finally a virtual model of a 3D campus comprising a simplified 3-story faculty building (Figure 6) was created. The objective was the optimization of the 3D model and the demonstration of the desired functionalities. For these purposes two 3D modeling and post-processing software were used, i.e. 3DSMax and Trimble Sketchup. The model of the building resulted in 5962 vertices and 4528 faces. The textures and illumination were applied using OpenSim’s in-world tools. Furniture objects (tables, chair, computer monitors) were taken from the Google 3D Warehouse, distributed and shared under Trimble General Model License.

Prototyping 3D Virtual Learning Environments with X3D-based Content and Visualization Tools-Figure 12. Diagram of the proposed working methodology

The chart below (Figure 12) summarizes the workflow recommended for the implementation of a prototype of a 3D online campus.