Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
13
datasets available to search
ShareScore release 0.9.0
Dataset results
13 results for “generative learning objects”
Transfer learning with generative models for object detection on limited datasets
<p>The provided datasets are used for the analysis in the work "Transfer learning with generative models for object detection on limited datasets" (https://doi.org/10.1088/2632-2153/ad65b5). The availability of data is limited in some fields, especially for object detection tasks, where it is necessary to have correctly labeled bounding boxes around each object. A notable example of such data scarcity is found in the domain of marine biology, where it is useful to develop methods to automatically detect submarine species for environmental monitoring. To address this data limitation, the state-of-the-art machine learning strategies employ two main approaches. The first involves pretraining models on existing datasets before generalizing to the specific domain of interest. The second strategy is to create synthetic datasets specifically tailored to the target domain using methods like copy-paste techniques or ad-hoc simulators. The first strategy often faces a significant domain shift, while the second demands custom solutions crafted for the specific task. In response to these challenges, here we propose a transfer learning framework that is valid for a generic scenario. In this framework, generated images help to improve the performances of an object detector in a few-real data regime. This is achieved through a diffusion-based generative model that was pretrained on large generic datasets. With respect to the state-of-the-art, we find that it is not necessary to fine tune the generative model on the specific domain of interest. We believe that this is an important advance because it mitigates the labor-intensive task of manual labeling the images in object detection tasks. We validate our approach focusing on fishes in an underwater environment, and on the more common domain of cars in an urban setting. Our method achieves detection performance comparable to models trained on thousands of images, using only a few hundreds of input data. Our results pave the way for new generative AI-based protocols for machine learning applications in various domains, for instance ranging from geophysics to biology and medicine. The provided datasets are built with the help of Gligen and the already existing NuImages, Ozfish and Deepfish datasets. The file "CarGenerated.zip" contains images generated with Gligen and with provided bounding boxes around cars in an urban environment. The file "fishes_on_bkg.zip" provides fish images generated with fishes from Deepfish inpainted with Gligen on generated backgrounds. The file "fish_text.zip" contains images completely generated with Gligen containing fishes with annotated bounding boxes. Finally, the file "oz_masked_512.zip" contains a simpler dataset of copy paste images of Deepfish fishes on Ozfish backrounds. All the files contains the images saved in different folders for training and validation, plus an index file called gt_fish.csv for the bounding boxes.</p>
Figure 8. Prototype Front-End-Generative Learning Objects Instantiated with Random Numbers Based Expressions
<p>We modeled classes for each AGLO section: scenario, theory, questions, and<br> feedbacks. We modeled domain specific classes for the learned concepts like trees and graphs<br> having generative methods controllable through parameters.<br> Figure 8 depicts our prototype front-end.</p>
Figure 7. Generated Tree Example-Generative Learning Objects Instantiated with Random Numbers Based Expressions
<p>In this subsection we give an example of a model instantiation. We describe the instantiation<br> process for all the symbols in the AGLO model. The nNoOfElements symbol was instantiated with<br> the generated value of 10. The tabKeys symbol was instantiated with the [A,B,C,D,E,F,G,H,I]<br> generated array. The tree symbol was instantiated with a ”Tree” object built with arguments equal<br> with previously instantiated symbols. Class ”Tree” is part of the auxiliary JavaScript library.The treeSVG symbol was assigned to the tree SVG string representation obtained by a method call and<br> is depicted in figure 7.</p>
Figure 5. AGLO Scenario Symbols-Generative Learning Objects Instantiated with Random Numbers Based Expressions
<p>analyzed AGLO that is displayed to the learner for localization and selection purposes.<br> The second XML element is the scenario element containing a text description of the AGLO<br> and a set of symbols. The description is expressed in natural language and we can notice that it<br> contains four main steps:<br> i) random tree generation;<br> ii) index computation for presentation;<br> iii) parent index computation for answer validation;<br> iv) access to the first two keys for particular feedback generation.<br> In the scenario section depicted in figure 5 several symbols are defined with the following<br> semantics.</p>
Figure 6. AGLO Presentation Sections-Generative Learning Objects Instantiated with Random Numbers Based Expressions
<p>The seventh and eighth symbols named node1 and node2, both of type character, are<br> initialized with the first two values from the keys array. These symbols will be used later in the<br> feedback section. Their initialization is based on accessing the keys using the array indexing<br> operator.<br> Next, we discuss about the presentation sections of the AGLO depicted in figure 6.</p>
Figure 3. Autogenerative Learning Object Model Definition (2)Generative Learning Objects Instantiated with Random Numbers Based Expressions
<p>Formally, the theory section (see line 11) contains only static data. Here the educator places<br> theoretical information about the learned concept in HTML format.<br> The question, answers and feedbacks sections (see lines 12-13, 14-17, 18-21) contain a mix<br> of static data and dynamic values referring the previously computed symbol values. The semantic of<br> these sections is to create a dynamic content to be presented to the learner. Formalisms for specific<br> details are also present, like:<br> i) correctness - because we need to know which is the correct answer if they are multiple;<br> ii) feedbacks answer identifier list - because feedbacks may be related to certain answers and<br> showed to the learner only in certain conditions.</p>
Figure 2. Autogenerative Learning Object Model Definition (1)-Generative Learning Objects Instantiated with Random Numbers Based Expressions
<p>The proposed model is defined using the EBNF meta-language resulting in the concise<br> grammar from figures 2 and 3. We will not present the full XML Schema because of clarity and<br> space reasons.<br> Structurally, the AGLO model has six sections: i) name; ii) scenario; iii) theory; iv) question;<br> v) answers; vi) feedbacks. Next, we will explain each section in details.</p>
Figure 4. AGLO Name and Description-Generative Learning Objects Instantiated with Random Numbers Based Expressions
<p>In this subsection we describe an AGLO model example. Figure 4 presents in details all the<br> XML elements containing AGLO metadata.</p>
Figure 1. The AGLO Workflow-Generative Learning Objects Instantiated with Random Numbers Based Expressions
<p>E-learning is a key area of research with a great influence on the developments of several<br> industries. For example, the nowadays ITC industry is in a continuous growth because of its<br> applications in almost all industrial domains. Companies tend to lack qualified human resources and<br> because of that they reject high economical value projects. In response to this lack of human<br> resource problem, universities started to develop several alternative study programs, many of them<br> are based on e-learning technology and namely on electronic learning materials. Learning objects<br> (LO) are considered to be digital resources that support learning and can be delivered across<br> networks in large or small sizes (Wiley, 2000). In order to increase the reusability and<br> interoperability of LOs, standards were developed by several organizations (IEEE Learning<br> Technology Standards Committee) (e.g. LOM http://ltsc.ieee.org/doc/wg12/LOMv4.1.htm.)</p>
BRAIN Journal-Auto-generative Learning Objects in Online Assessment of Data Structures Disciplines-Figure 3. Symbols definition for a graph-based test
<p>In this scenario, we intend to generate a random graph and compute a deep first-search node list. The first defined random symbol is n, namely the number of nodes in the graph as an integer from 5 to 9. The next symbol is named g and denotes the graph object created randomly using 3 parameters: the number of nodes, the minimum, and the maximum value for the weight. For the number of nodes, we used the previously computed value of n, whereas for the weights, we used two constants 0 and 1 since the graph is not weighted</p>
BRAIN Journal-Auto-generative Learning Objects in Online Assessment of Data Structures Disciplines-Figure 2. Auto-generative Learning Object Model Definition
<p>In this section, we will present the structure of AGLOs in the context of our approach. The AGLO meta-model is structured in XML as in Figure 2,a refinement from Chirila, Ciocarlie, and Stoicu (2015). The AGLO definition contains several sections like name, scenario, theory, question, answers, and feedback (line 01). The name element contains the name of the AGLO, possibly a small description in the human language (line 02). The section of the scenario (line 03) contains a comment (line 04) followed by a set of symbol definitions. The comment should describe the imagined scenario in details and it has the same role as code comments. The symbol is the central element of the AGLO model. The symbol has a name and is very similar to programming language variables. Symbols may be called also parameters since they control the content of the AGLO content in the process of instantiation. </p>
BRAIN Journal-Auto-generative Learning Objects in Online Assessment of Data Structures Disciplines-Figure 1. The AGLO online assessment approach
<p>In Figure 1, we present the lifetime of AGLOs in the context of online student assessment following a set of steps. In the backend, the tutor develops an AGLO model respecting a predefined meta-model. The model is intuitive, it has a few sections where symbols are defined using formulas and random numbers and then used in a section of a presentation for the student. When such models are created they are stored in a storage facility like a database to be selected by the student through the web application frontend. In the frontend, the students access the web application using a web browser from a workstation, tablet or smartphone. In the assessment process, the student will access several AGLOs. At this step, the accessed AGLOs are instantiated with random numbers, formulas are evaluated to fulfill the designed learning or testing scenario and to create the presentation content for the student. Nevertheless, the instantiated symbols will be used for the automatic assessment of the answers correctness</p>
BRAIN Journal-Auto-generative Learning Objects in Online Assessment of Data Structures Disciplines-Figure 4. Online test assessment example
<p>Thus, applying these restrictions the computed solution is C, E, G, J, L, H, I and is unique. Node C is the starting node since it is the first from the lexicographical point of view. The first step CE is the only choice coping with the restrictions from the [CE, CG, and CJ] edges. Next, EG is the first edge in the list of [EG, EJ]. The next step is GJ which is the only choice. Edge JL is another unique choice. Edge LH is the next step from the list [LH, LI]. Finally, the last edge is obtained by backtracking to node L and then taking edge LI. These restrictions allow us to drive the student to build only one solution from the possible set of solutions. This will determine an easier way of comparing the student’s answer with the answer of the computer. Another more general solution is to use validation functions which require implementation in domain libraries written in JavaScript. </p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.