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.
284
datasets available to search
ShareScore release 0.9.0
Dataset results
284 results for “reasoning”
A Benchmark Suite for Systematically Evaluating Reasoning Shortcuts
<p><strong>Codebase</strong> [<a href="https://unitn-sml.github.io/rsbench/">Github</a>] | <strong>Dataset</strong> [<a href="doi.org/10.5281/zenodo.11612556">Zenodo</a>]</p> <p> </p> <p><strong>Abstract</strong></p> <p>The advent of powerful neural classifiers has increased interest in problems that require both learning and reasoning. These problems are critical for understanding important properties of models, such as trustworthiness, generalization, interpretability, and compliance to safety and structural constraints. However, recent research observed that tasks requiring both learning and reasoning on background knowledge often suffer from <em>reasoning shortcuts</em> (RSs): predictors can solve the downstream reasoning task without associating the correct concepts to the high-dimensional data. To address this issue, we introduce <strong>rsbench</strong>, a comprehensive benchmark suite designed to systematically evaluate the impact of RSs on models by providing easy access to highly customizable tasks affected by RSs. Furthermore, rsbench implements common metrics for evaluating concept quality and introduces novel formal verification procedures for assessing the presence of RSs in learning tasks. Using rsbench, we highlight that obtaining high quality concepts in both purely neural and neuro-symbolic models is a far-from-solved problem. rsbench is available on <a href="https://unitn-sml.github.io/rsbench">Github</a>.</p> <p> </p> <p><strong>Usage</strong></p> <p>We recommend visiting the official code <a href="https://unitn-sml.github.io/rsbench/">website</a> for instructions on how to use the dataset and accompaying software code.</p> <p> </p> <p><strong>License</strong></p> <p>All ready-made data sets and generated datasets are distributed under the <a href="https://creativecommons.org/licenses/by-sa/4.0/">CC-BY-SA 4.0</a> license, with the exception of <code>Kand-Logic</code>, which is derived from <code>Kandinsky-patterns</code> and as such is distributed under the <a href="https://www.gnu.org/licenses/gpl-3.0.en.html">GPL-3.0</a> license.</p> <p> </p> <p><strong>Datasets Overview</strong></p> <ul> <li><strong>CLIP-embeddings. </strong>This folder contains the saved activations from a pretrained CLIP model applied to the tested dataset. It includes embeddings that represent the dataset in a format suitable for further analysis and experimentation.</li> <li><strong>BDD_OIA-original-dataset</strong>. This directory holds the original files from the X-OIA project by Xu et al. [1]. These datasets have been made publicly available for ease of access and further research. If you are going to use it, please consider citing the original authors.</li> <li><strong>kand-logic-3k</strong>. This folder contains all images generated for the Kand-Logic project. Each image is accompanied by annotations for both concepts and labels.</li> <li><strong>bbox-kand-logic-3k</strong>. In this directory, you will find images from the Kand-Logic project that have undergone a preprocessing step. These images are extracted based on bounding boxes, rescaled, and include annotations for concepts and labels.</li> <li><strong>sdd-oia</strong>. This folder includes all images and labels generated using rsbench.</li> <li><strong>sdd-oia-embeddings</strong>. This directory contains 512-dimensional embeddings extracted from a pretrained ResNet18 model on ImageNet. The embeddings are derived from the sdd-oia`dataset.</li> <li><strong>BDD-OIA-preprocessed</strong>. Here you will find preprocessed data that follow the methodology outlined by Sawada and Nakamura [2]. The folder contains 2048-dimensional embeddings extracted from a pretrained Faster-RCNN model on the BDD-100k dataset.</li> </ul> <p>The original BDD datasets can be downloaded from the following Google Drive link: [<a href="https://drive.google.com/file/d/1WFiwRi_sMA_McZnkbEjh8Rnl-Im7_9Mk/view">Download BDD Dataset</a>].</p> <p> </p> <p><strong>References</strong></p> <p>[1] Xu et al., *Explainable Object-Induced Action Decision for Autonomous Vehicles*, CVPR 2020.</p> <p>[2] Sawada and Nakamura, *Concept Bottleneck Model With Additional Unsupervised Concepts*, IEEE 2022.</p> <p> </p>
CausalBench A Comprehensive Benchmark for Evaluating Causal Reasoning Capabilities of Large Language Models
<p>CausalBench is a comprehensive benchmark dataset designed to evaluate the causal reasoning capabilities of large language models. The primary uses of this dataset include, but are not limited to:</p> <p>- Testing the performance of large language models on causal reasoning tasks</p> <p>- Serving as a benchmark dataset for causal reasoning research</p> <p>- Improving and developing new causal reasoning algorithms and models</p>
Data study "The Impact of Augmented Reality on Biodiversity Learning in a Pedagogical Scenario Based on Analogical Reasoning: An Experimental Study"
<p>This data was collected in 2023 as part of a study on the impact of location-based AR on biodiversity education. </p>
OWL Reasoner Collection for Evaluation
<p>Set of reasoners and auxiliary scripts for use with the evaluation benchmarking framework at <a href="https://github.com/sisinflab-swot/owl-reasoner-test-framework">https://github.com/sisinflab-swot/owl-reasoner-test-framework</a>. It includes:</p> <ul> <li> <p><a href="https://bitbucket.org/dtsarkov/factplusplus">Fact++</a> (version 1.6.5);</p> </li> <li> <p><a href="http://www.hermit-reasoner.com/">HermiT</a> (version 1.3.8);</p> </li> <li> <p><a href="http://derivo.de/en/products/konclude/">Konclude</a> (version 0.6.2-544);</p> </li> <li> <p><a href="http://trowl.org/">TrOWL</a> (version 1.5);</p> </li> <li> <p><a href="http://sisinflab.poliba.it/swottools/minime/">Mini-ME Java</a> (version 2.0);</p> </li> <li> <p><a href="http://sisinflab.poliba.it/swottools/minime-swift/">Mini-ME Swift</a> (version 1.0).</p> </li> </ul>
ALN Reasoner Evaluation Dataset
<p>Dataset for evaluation of Description Logics reasoners in the Attributive Language with unqualified Number restrictions (ALN). The dataset consists in two sections: a) for standard reasoning tasks, 1398 knowledge bases with expressivity up to ALN, extracted from the 2014 OWL Reasoner Evaluation Workshop competition dataset; b) for non-standard reasoning tasks, 4 knowledge bases, each with a different number of "request" and "resource" individuals.</p>
Towards a critique of digital reason
<p>Opening Lecture of the 10th European Summer University in Digital Humanities "Culture & Technology" 2019</p>
Fig. 3 in MAX WEBER AN ARTIFICIAL SOCIOLOGY. PROPOSAL FOR A NEURAL NETWORK WITH WEBERIAN REASONING
Fig. 3. The ornithopod dinosaur Gasparinisaura cincosaltensis Coria and Salgado, 1996 from the Late Cretaceous Anacleto Formation of Patagonia, MCSPv 111 (A) and MCSPv 112 (B). A1, nearly complete postcranial skeleton in left lateral view; A2, cluster of gastroliths in the abdominal cavity; B1, nearly complete skeleton in dorsal view; B2, the largest cluster of gastroliths below the last dorsal vertebra.
Fig. 2 in MAX WEBER AN ARTIFICIAL SOCIOLOGY. PROPOSAL FOR A NEURAL NETWORK WITH WEBERIAN REASONING
Fig. 2. The ornithopod dinosaur Gasparinisaura cincosaltensis Coria and Salgado, 1996, MUCPv 213 from the Late Cretaceous Anacleto Formation of Patagonia. A. Forelimb bones and ribs. B. The longest gastrolith in contact with two right dorsal ribs. C. Cluster of gastroliths associated with the ribs. D. Scanning electron microphotograph of a gastrolith from metamorphic rock.
Fig. 1 in MAX WEBER AN ARTIFICIAL SOCIOLOGY. PROPOSAL FOR A NEURAL NETWORK WITH WEBERIAN REASONING
Fig. 1. The surroundings of Cinco Saltos City where MCSPv 111, MCSPV 112 (Site 1) and MUCPv 213 (Site 2) were found (modified from Andreis et al. 1974).
OWL reasoner evaluation results - mobile and edge
<p>Results of an evaluation campaign concerning OWL ontology classification on mobile and edge devices.</p> <p>Experiments were conducted using the evOWLuator framework.</p> <h2>Mobile</h2> <p><strong>Device:</strong> Samsung Galaxy Tab S6 Lite</p> <p><strong>Dataset:</strong> <a href="https://doi.org/10.5281/zenodo.10791">ORE 2014 competition corpus</a> (full dataset)</p> <p><strong>Reasoners:</strong></p> <ul> <li><a href="http://www.hermit-reasoner.com">HermiT</a> (v1.3.8)</li> <li><a href="https://github.com/stardog-union/pellet">Pellet</a> (v2.3.1)</li> <li><a href="https://jfact.sourceforge.net">JFact</a> (v1.2.1)</li> <li><a href="https://github.com/TrOWL/core">TrOWL</a> (v1.5)</li> <li><a href="https://github.com/liveontologies/elk-reasoner">ELK</a> (v0.4.3)</li> <li><a href="http://julianmendez.github.io/jcel">jcel</a> (v0.24.1)</li> </ul> <p><strong>Metrics</strong></p> <ul> <li>Correctness</li> <li>Parsing and reasoning time</li> <li>Peak memory usage</li> </ul> <h2>Edge</h2> <p><strong>Device:</strong> Raspberry Pi 4 Model B+</p> <p><strong>Dataset:</strong> <a href="https://doi.org/10.5281/zenodo.10791">ORE 2014 competition corpus</a> (subset)</p> <p><strong>Reasoners:</strong></p> <ul> <li><a href="https://www.derivo.de/products/konclude">Konclude</a> (v0.7.0)</li> <li><a href="https://github.com/TrOWL/core">TrOWL</a> (v1.5)</li> <li><a href="http://owl.cs.manchester.ac.uk/tools/fact">FaCT++</a> (v1.6.3.1)</li> </ul> <p><strong>Metrics:</strong></p> <ul> <li>Correctness</li> <li>Parsing and reasoning time</li> <li>Peak memory usage</li> <li>Energy usage (<a href="https://github.com/fenrus75/powertop">PowerTOP</a>)</li> <li>Energy usage (<a href="https://www.msoon.com/high-voltage-power-monitor">Monsoon High Voltage Power Monitor</a>)</li> </ul>
Experimental Repository for "Certifying Without Loss of Generality Reasoning in Solution-Improving Maximum Satisfiability"
<h2>Experimental repository for "Certifying Without Loss of Generality Reasoning in Solution-Improving Maximum Satisfiability".</h2> <h3>The directory is structured as follows:</h3> <ul> <li><code>data</code>: Data that has been processed into CSVs and the scripts to analyse the experiments.</li> <li><code>plots</code>: The plots generated from the data that are used in the paper.</li> <li><code>raw_data</code>: The raw data logs for the experiments and scripts to extract relevant data from the logs.</li> <li><code>source_code</code>: Source code for the checker <code>VeriPB</code> and the MaxSAT solver Pacose in the different variants used for the experiments. The Pacose version in <code>PacoseMaxSATSolver-baseline</code> is Pacose without proof logging, the version in <code>PacoseMaxSATSolver-certified</code> is Pacose with proof logging. Proof logging using only assumptions for the coarse convergence can be enabled via the option <code>--WithAssumptions</code>.</li> </ul> <h3>Install Requirements</h3> <p>To install and run the MaxSAT solver Pacose and the pseudo-Boolean proof checker VeriPB your need to have the following components installed:</p> <ul> <li>Python 3.6.9 or higher with pip and setuptools installed</li> <li>g++ 7.5.0 or higher</li> <li>libgmp</li> </ul> <p>These can be installed in Ubuntu / Debian via</p> <pre><code>sudo apt-get update && apt-get install \ python3 \ python3-pip \ python3-dev \ g++ \ libgmp-dev pip3 install --user \ setuptools</code></pre> <h3>How to Run?</h3> <p>The MaxSAT solver Pacose can be compiled using the install script in <code>PacoseMaxSATSolver-certified</code>:</p> <p><code>./install</code></p> <p>To run Pacose with proof logging, where the proof should be written to <code>proof.pbp</code>, run the following command in the <code>PacoseMaxSATSolver-certified</code> directory:</p> <p><code>./bin/Pacose --proofFile proof.pbp instance.wcnf</code></p> <p>The proof can be checked with VeriPB. To compile VeriPB run the following in the <code>VeriPB</code> directory:</p> <p><code>pip install .</code></p> <p>To check the proof with VeriPB, run the following inside the <code>PacoseMaxSATSolver-certified</code> directory:</p> <p><code>veripb --wcnf instance.wcnf proof.pbp</code></p>
Analogical Reasoning (video recording)
<p>Video recording of the lecture "Analogical Reasoning" given on 2021-10-06 as Module 6 of <a href="https://redwood.berkeley.edu/courses/computing-with-high-dimensional-vectors">Neuroscience 299: Computing with High-Dimensional Vectors</a> at the <a href="https://redwood.berkeley.edu/">Redwood Center for Theoretical Neuroscience</a>, University of California, Berkeley.</p>
Evaluating Datalog Tools for Meta-reasoning over OWL 2 QL
<p>There has been increasing interest in enriching ontologies with meta-modeling and meta-querying for the past few years. Unfortunately, the Direct Semantics for OWL2 and SPARQL does not support meta-constructs in a satisfactory way: While meta-axioms (involving identifiers used both as a class and an individual) can be syntactically expressed using punning, different occurrences of the same identifier will be treated as different entities. For example, GoldenEagle used as an instance of EndangeredSpecies and also as a class containing individuals, will be treated as two separate, unrelated entities. Meta-queries (for example, asking for classes that also occur as individuals) are not allowed at all in SPARQL under the Direct Semantics Entailment Regime. To overcome this, a new semantic flavour for SPARQL, called Meta-modeling Semantics Entailment Regime (MSER), has been introduced. In previous work, Cima et al. have proposed a reduction from OWL 2 QL (a light-weight profile of OWL 2) and associated meta-queries in SPARQL to query answering over Datalog rules. In this paper, we experiment with various logic programming tools that support Datalog querying to determine their suitability as back-ends to MSER query answering. These tools stem from different logic programming paradigms (Prolog, pure Datalog, Answer Set Programming). Our work shows that the Datalog approach to MSER querying is practical also for sizeable ontologies.</p>
A database of contexts with polypredicative reason constructions in the New Testament (База данных по контекстам с полипредикативными причинными конструкциями в Новом Завете)
<p>The database accumulates contexts with polypredicative reason constructions in the text of the New Testament.</p> <p>The initial sample was based on the distribution of reason conjunctions (used as keywords) in the translations of the New Testament into three languages: Russian (<em>poskol'ku</em>, <em>tak kak</em>, <em>potomu čto</em>, <em>potomu, čto</em>, <em>ottogo, čto</em>, <em>iz-za togo, čto</em>), English (<em>because</em>, <em>since</em>, <em>for</em>), and French (<em>puisqu</em>(<em>e</em>), <em>parce qu</em>(<em>e</em>), <em>comme</em>, <em>car</em>). The raw sample obtained by automated extraction was then manually edited to remove the contexts containing irrelevant (non-causal) uses of the keywords and to split the verses containing more than one reason context.</p> <p>The resulting database contains 1630 reason contexts (in 1504 verses), which are likely to be translated with polypredicative reason constructions in any language. The database is intended as a tool for extracting relevant material for studies on individual languages and for typological studies of reason constructions.</p> <p>In addition to the main database, a smaller subsample of 86 contexts has been annotated for semantic/pragmatic types of reason meanings and for the information structure of the reason construction.</p> <p>The database was created at the Institute for Linguistic Studies, RAS with support from the Russian Science Foundation grant # 18-18-00472 “Causal Constructions in World Languages (Semantics and Typology)”.</p> <p> </p> <p>База данных объединяет контексты с полипредикативными причинными конструкциями в тексте Нового Завета.</p> <p>Исходная выборка была основана на распределении причинных союзов (послуживших ключевыми словами) в переводах Нового Завета на три языка: русский (<em>поскольку</em>, <em>так как</em>, <em>потому что</em>, <em>потому,</em><em> что</em>, <em>оттого, что</em>, <em>из-за того, что</em>), английский (<em>because</em>, <em>since</em>, <em>for</em>) и французский (<em>puisqu</em>(<em>e</em>), <em>parce qu</em>(<em>e</em>), <em>comme</em>, <em>car</em>). Эта выборка, полученная при помощи автоматического извлечения подходящих стихов, затем была вручную обработана: из неё были удалены контексты с нерелевантными (непричинными) употреблениями ключевых слов, а также были разделены на несколько вхождений те стихи, которые содержали более одного причинного контекста.</p> <p>Получившаяся база данных содержит 1630 причинных контекстов (из 1504 стихов), в которых в переводе на любой язык можно с высокой вероятностью ожидать появления полипредикативных причинных конструкций. База может быть использована для извлечения релевантного материала по таким конструкциям для конкретно-языковых и типологических исследований.</p> <p>В дополнение к основной базе данных для небольшой подвыборки (86 контекстов) произведена пилотная разметка по различным семантико-прагматическим типам причинных контекстов и по типам информационной структуры в них.</p> <p>База данных создана в ИЛИ РАН при поддержке гранта РНФ № 18-18-00472 «Причинные конструкции в языках мира: семантика и типология»</p>
Defining Categorical Reasoning of Numerical Feature Models with Feature-Wise and Variant-Wise Quality Attributes
<p><strong>To watch it in Youtube:</strong></p> <p><a href="https://youtu.be/Uq2qtb4_K2U">https://youtu.be/Uq2qtb4_K2U</a></p> <p><strong>This is a pre-print, please access and cite the published version:</strong></p> <p><a href="https://doi.org/10.1145/3503229.3547057">https://doi.org/10.1145/3503229.3547057</a></p> <p>Automatic analysis of variability is an important stage of <em>Software Product Line</em> (SPL) engineering. Incorporating quality information into this stage poses a significant challenge. However, quality-aware automated analysis tools are rare, mainly because in existing solutions variability and quality information are not unified under the same model.</p> <p>In this paper, we make use of the <em>Quality Variability Model</em> (QVM), based on <em>Category Theory</em> (CT), to redefine reasoning operations. We start defining and composing the six most common operations in SPL, but now as quality-based queries, which tend to be unavailable in other approaches. Consequently, QVM supports interactions between variant-wise and feature-wise quality attributes. As a proof of concept, we present, implement and execute the operations as lambda reasoning for CQL IDE -- the state-of-the-art CT tool.</p>
Are there good ethical reasons why for profit publishers should no longer exist under the conditions of digital infrastructures? And what does this have to do with ethics as a reflexive discipline?
<p>Talk at the <a href="https://www.digital-philosophy.org/">Philosophy [in:of:for:and] Digital Knowledge Infrastructures</a> online workshop (08/09/2022).</p>
Dataset for exploring the reasons of diverging views in taxonomy
<p>Dataset for a vignette study to explore whether, and for what reasons, taxonomists disagree about ranking decisions.</p>
Dataset corresponding to the « Reasons for concern » about climate change impacts from all IPCC reports (TAR to AR6)
<p>This data corresponds to all the "burning embers" diagrams for the "Reasons for Concern" published in IPCC reports (and the related paper Smith et al. 2009 for AR4) until AR6 (thus including TAR, AR4, AR5, SR1.5 and AR6). For TAR to SR1.5, the data is the result of extracting information from the original figures, as presented in the related technical document <a href="https://doi.org/10.5281/zenodo.3992856">10.5281/zenodo.3992856</a>. As also explained in the Supplementary Information of Zommers et al. (2020), the data does not come directly from the IPCC, although it is based on the assessment provided in the IPCC reports listed in the references. For IPCC AR6, the source is the supplementary material of chapter 16. Details regarding specific values provided in the dataset are explained alongside the values in the main file: "RFCs-ALL-2023_05_12.xlsx". </p> <p>The main file includes the parameters needed to produce a diagram that supplements figure 3 from Zommers et al. (2020) with AR6 data and the confidence levels from previous reports when available. The Excel files in RFCs-2023-UsageExamples.zip contain the same data with different parameters, so that uploading these files to the Ember Factory (<a href="https://climrisk.org/emberfactory">https://climrisk.org/emberfactory</a>) produces different figures - including a comparison between AR5 and AR6 (as in IPCC AR6 Synthesis Report, but with AR5 confidence levels included). The resulting diagrams are also provided.</p>
ThoughtSource: A central hub for large language model reasoning data (code snapshot)
<p><strong>ThoughtSource is a meta-dataset and software library for chain-of-thought reasoning in large language models (LLMs). This repository contains a snapshot of the associated GitHub repository.</strong></p>
Fig. 4. A in Good Reasons and Guidance for Mapping Planktonic Protist Distributions
Fig. 4. A variogram for the ciliate Pleuronema sp. (inset) abundance. The best fit to the data (points) provided a pure nugget model; i.e. the distribution is random at the measured scale (40 m), with no observed patchiness.
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.