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.
9
datasets available to search
ShareScore release 0.9.0
Dataset results
9 results for “Graph Deep Learning”
Challenges in Migrating Imperative Deep Learning Programs to Graph Execution: An Empirical Study
<p>Efficiency is essential to support responsiveness w.r.t. ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code that supports symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development tends to produce DL code that is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, less error-prone imperative DL frameworks encouraging eager execution have emerged but at the expense of run-time performance. While hybrid approaches aim for the "best of both worlds," the challenges in applying them in the real world are largely unknown. We conduct a data-driven analysis of challenges—and resultant bugs—involved in writing reliable yet performant imperative DL code by studying 250 open-source projects, consisting of 19.7 MLOC, along with 470 and 446 manually examined code patches and bug reports, respectively. The results indicate that hybridization: (i) is prone to API misuse, (ii) can result in performance degradation—the opposite of its intention, and (iii) has limited application due to execution mode incompatibility. We put forth several recommendations, best practices, and anti-patterns for effectively hybridizing imperative DL code, potentially benefiting DL practitioners, API designers, tool developers, and educators.</p>
Speculative Automated Refactoring of Imperative Deep Learning Programs to Graph Execution
<p>Efficiency is essential to support ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code---supporting symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, imperative DL frameworks encouraging eager execution have emerged but at the expense of run-time performance. Though hybrid approaches aim for the "best of both worlds," using them effectively requires subtle considerations. Our key insight is that, while DL programs typically execute sequentially, hybridizing imperative DL code resembles parallelizing sequential code in traditional systems. Inspired by this, we present an automated refactoring approach that assists developers in determining which otherwise eagerly-executed imperative DL functions could be effectively and efficiently executed as graphs. The approach features novel static imperative tensor and side-effect analyses for Python. Due to its inherent dynamism, analyzing Python may be unsound; however, the conservative approach leverages a speculative (keyword-based) analysis for resolving difficult cases that informs developers of any assumptions made. The approach is: (i) implemented as a plug-in to the PyDev Eclipse IDE that integrates the WALA Ariadne analysis framework and (ii) evaluated on nineteen DL projects consisting of 132 KLOC. The results show that 326 of 766 candidate functions (42.56%) were refactorable, and an average relative speedup of 2.16x on performance tests was observed with negligible differences in model accuracy. The results indicate that the approach is useful in optimizing imperative DL code to its full potential.</p>
Deep learning models predicting gene functions and pathways using public DRKG knowledge graph and graph neural network
<p>The attached dataset contains pretrained link prediction models, as described in our paper 'Morphological Map of Under- and Over-Expression of Genes in Human Cells'.</p>
A supervised Graph-based deep learning algorithm to detect and quantify clustered particles
<p>In this data repository, we provide the necessary data for replicating results, including both simulated and biological datasets. Additionally, the repository includes trained models to infer from these datasets.</p>
Processed CODEX Datasets from - Graph deep learning for the characterization of tumour microenvironments from spatial protein profiles in tissue specimens
<p>This entry provides access to processed CODEX data files of three studies analyzed in the article "Graph deep learning for the characterization of tumour microenvironments from spatial protein profiles in tissue specimens". Details of datasets can be found in the Methods section of the article.</p> <p>For each dataset:</p> <ul> <li>A comma-separated values (CSV) file containing metadata of regions is included</li> <li>A zip file containing multiple CSV files is included: <ul> <li>`{region_id}.cell_data.csv`, a table containing three columns: "CELL_ID", "X", and "Y". This table provides centroid locations for all cells segmented in this region.</li> <li>`{region_id}.expression.csv`, a table containing multiple columns: "CELL_ID", "DAPI", "CD45", etc. This table provides detailed protein biomarker expression quantified and normalized for all cells in this region.</li> <li>`{region_id}.cell_types.csv`, a table containing two columns: "CELL_ID" and "CELL_TYPE". This table provides cell type annotations for all cells in this region.</li> <li>`{region_id}.cell_features.csv`, a table containing two columns: "CELL_ID" and "SIZE". This table provides morphology descriptors (only containing cell size for these studies) for all cells in this region.</li> </ul> </li> </ul> <p>These data files are also available through the Enable Medicine Public Study page: <a href="https://app.enablemedicine.com/portal/atlas-library/studies/92394a9f-6b48-4897-87de-999614952d94?sid=1168">https://app.enablemedicine.com/portal/atlas-library/studies/92394a9f-6b48-4897-87de-999614952d94?sid=1168</a>. Raw multiplexed immunofluorescence images will be accessible through the visualizer app of Enable Medicine Portal.</p> <p>Codes for this study are stored in <a href="https://gitlab.com/enable-medicine-public/space-gm">https://gitlab.com/enable-medicine-public/space-gm</a>. Please direct all further questions and/or issues to the gitlab repository or lead contact (A.E.T.).</p>
HAPPY: a deep learning pipeline for mapping cell-to-tissue graphs across placenta histology whole slide images
<p>These two zipped folders contain all data necessary to train, validate and reproduce results from the paper.</p> <p>Unzipping the files will create 6 folders. Data from folders with the same name across both zips should be combined into one folder. The 'annotations' folder contains all ground truth annotations for training all three deep learning models. The 'datasets' folder contains images for training the nuclei localisation and cell classification models. The 'embeddings' folder contains cell embedding vectors and nuclei coordinates from two slides used to create nodes to train the graph tissue classification model. The 'graph_splits' folder contains regions defining the validation and test splits for the graph model. The 'slides' folder contains a sample region of a whole slide image as a .tiff file for running the inference demo. The 'trained_models' folder contains trained weights for each of the three models.</p> <p>Further instructions for dataset use and creation of custom datasets are available in the GitHub readme: https://github.com/Nellaker-group/happy.</p>
Data for paper "Graph Deep Learning Model for Mapping Mineral Prospectivity"
<p>Four prospecting information, namely, NE- and NW- trending faults, Agno Batholithic pluton margins, and porphyry intrusive contacts for mineral prospectivity mapping in Baguio district, Philippines.</p>
Deep learning and knowledge graph powered drug combination discovery against infectious diseases
<p>The Datasets and source codes for paper "<strong>Deep learning and knowledge graph powered drug combination discovery against infectious diseases</strong>".</p>
Spatially resolved transcriptomics and graph-based deep-learning improve accuracy of routine CNS tumor diagnostics
<p><span>The diagnostic landscape of brain tumors has recently evolved to integrate comprehensive molecular markers alongside traditional histopathological evaluation. Foremost, genome-wide DNA methylation profiling and next generation sequencing (NGS) has become a cornerstone in classifying Central Nervous System (CNS) tumors, as recognized by its inclusion into the 2021 WHO classification. Despite its diagnostic precision, a limiting requirement for NGS and methylation profiling is sufficient DNA quality and quantity which restricts its feasibility, especially in cases with small biopsy samples or low tumor cell content, both frequent challenges in specimen of diffusely growing CNS lesions. Addressing these challenges, we demonstrate a application, namely <strong>NePSTA </strong>(<strong>Ne</strong>uro<strong>P</strong>athology <strong><em>S</em></strong><em>patial <strong>T</strong>ranscriptomic<strong> A</strong>nalysis</em>), which is capable of generating comprehensive morphological and molecular neuropathological diagnostics from single 5 µm tissue sections. Our framework employs 10x Visium spatial transcriptomics with graph neural networks for automated histological and molecular evaluations. Trained and evaluated across 130 patients with CNS malignancies and healthy donors across four medical centers, NePSTA<strong> </strong>integrates spatial gene expression data and inferred CNAs to predict tissue histology and methylation-based subclasses with high accuracy. Further, we demonstrate the ability to reconstruct immunohistochemistry and genotype profiling on single thin slides of minute tissue biopsies. Our approach has minimal tissue requirements, often inadequate for conventional molecular diagnostics, demonstrating the potential to transform neuropathological diagnostics and enhance tumor subtype identification with implications for fast and precise diagnostic work-up.</span></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.