19,997 machine learning datasets
19,997 dataset results
Mathematical dataset based on 71 famous mathematical identities. Each entry consists of a name of the identity (name), a representation of that identity (formula), a label whether the representation belongs to the identity (label), and an id of the mathematical identity (formula_name_id). The false pairs are intentionally challenging, e.g., a^2+2^b=c^2as falsified version of the Pythagorean Theorem. All entries have been generated by using data.json as starting point and applying the randomizing and falsifying algorithms from MathMutator (MAMUT). The formulas in the dataset are not just pure mathematical, but contain also textual descriptions of the mathematical identity. At most 400000 versions are generated per identity. There are ten times more falsified versions than true ones, such that the dataset can be used for a training with changing false examples every epoch.
Mathematical dataset based on 71 famous mathematical identities. Each entry consists of two identities (in formula or textual form), together with a label, whether the two versions describe the same mathematical identity. The false pairs are not randomly chosen, but intentionally hard by modifying equivalent representations (see ddrg/named_math_formulas for more information). At most 400000 versions are generated per identity. There are ten times more falsified versions than true ones, such that the dataset can be used for a training with changing false examples every epoch.
The WORC database consists in total of 930 patients composed of six datasets gathered at the Erasmus MC, consisting of patients with: 1) well-differentiated liposarcoma or lipoma (115 patients); 2) desmoid-type fibromatosis or extremity soft-tissue sarcomas (203 patients); 3) primary solid liver tumors, either malignant (hepatocellular carcinoma or intrahepatic cholangiocarcinoma) or benign (hepatocellular adenoma or focal nodular hyperplasia) (186 patients); 4) gastrointestinal stromal tumors (GISTs) and intra-abdominal gastrointestinal tumors radiologically resembling GISTs (246 patients); 5) colorectal liver metastases (77 patients); and 6) lung metastases of metastatic melanoma (103 patients). For each patient, either a magnetic resonance imaging (MRI) or computed tomography (CT) scan, collected from routine clinical care, one or multiple (semi-)automatic lesion segmentations, and ground truth labels from a gold standard (e.g., pathologically proven) are available. All datasets are
This collection contains 156 cases of MAASTRO Head and Neck images and RTStruct contours. Where PET exists, the frame of reference UID of the CT has been transferred to the PET. This collection may not be used for commercial purposes. This collection is freely available to browse, download, and use for scientific and educational purposes as outlined in the Attribution-NonCommercial 3.0 Unported (CC BY-NC 3.0) https://creativecommons.org/licenses/by-nc/3.0/. Please be sure to include the following citation if you make use of this data set : "Aerts, H. J. W. L., Velazquez, E. R., Leijenaar, R. T. H., Parmar, C., Grossmann, P., Carvalho, S., … Lambin, P. (2014, June 3). Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nature Communications. Nature Publishing Group. http://doi.org/10.1038/ncomms5006."
Purpose: Patients with 1p/19q codeleted low-grade glioma (LGG) have longer overall survival and better treatment response than patients with 1p/19q intact tumors. Therefore, it is relevant to know the 1p/19q status. To investigate whether the 1p/19q status can be assessed prior to tumor resection, we developed a machine learning algorithm to predict the 1p/19q status of presumed LGG based on preoperative MRI.
All Conference Alert is a tech startup dedicated to helping organizers and attendees of academic conferences, seminars, workshops, and webinars. Our platform allows organizers to list and promote their events, covering over 150 academic subjects across 135 countries.
A benchmark for legal question answering. The data only contains test set and got two splits: NitiBench-CCL representing Thai corporate and commercial law, and NitiBench-Tax containing official tax ruling in Thai scraped from official Revenue Department website.
The WangchanX-Legal-ThaiCCL-RAG dataset supports the development Retrieval-Augmented Generation (RAG) for Thai Legal question answering. This dataset is allows developers to finetune both retrieval model - to better retrieve relevant law section, and Large Language Model (LLM) - for instruction tuning. Our dataset supports Corporate and Commercial Law (thus ThaiCCL name)
LLM-Based Vulnerability Classification in Police Narratives This repository contains datasets used in our research on applying large language models (LLMs) to identify indicators of vulnerability in police incident narratives. These resources support the replication of findings in our paper: "Using Instruction-Tuned Large Language Models to Identify Indicators of Vulnerability in Police Incident Narratives."
This dataset comprises over 9,000 images captured in the AI2-THOR simulation environment, featuring 69 distinct object classes. It includes variations of certain objects, such as raw bread and cooked bread, to enhance diversity and realism. The images were collected using AI2-THOR's built-in tools and subsequently preprocessed and formatted for compatibility with YOLOv5, making it suitable for object detection tasks in simulated environments.
These datasets, ComCo and SimCo, designed for evaluating multi-object representation in Vision-Language Models (VLMs). These datasets provide controlled environments for analyzing model biases, object recognition, and compositionality in multi-object scenarios.
Simulations and hardware experiments of different coverage tasks saved as robot swarm objects. The coverage runs can be loaded and visualized. Additionally, the anomaly detection method can be applied to each run, as shown in example.ipynb.
Recognizing events and their coreferential men- tions in a document is essential for understand- ing semantic meanings of text. The existing re- search on event coreference resolution is mostly limited to news articles. In this paper, we present the first dataset for the legal domain, LegalCore, which has been annotated with comprehensive event and event coreference in- formation. The legal contract documents we an- notated in this dataset are several times longer than news articles, with an average length of around 25k tokens per document. The anno- tations show that legal documents have dense event mentions and feature both short-distance and super long-distance coreference links be- tween event mentions. We further benchmark mainstream Large Language Models (LLMs) on this dataset for both event identification and event coreference resolution tasks, and find that this dataset poses significant challenges for both open-source and proprietary LLMs, which all perform significantly worse
MCiteBench is a benchmark to evaluate multimodal citation text generation in Multimodal Large Language Models (MLLMs).
Street-View images captured at different timestamps often undergo geometric transformations. To make the VL-CMU-CD dataset more challenging and closer to real-world applications, we generate Unaligned VL-CMU-CD by pairing images with their adjacent neighbors within the same sequence. In this dataset, the adjacent neighbor distance is set to 2 to set to 2 to ensure a distinct difference from the original VL-CMU-CD.
This dataset comprises extensive multi-modal data related to the experimental study of ultrasonically excited pulsating fluid jets used for bone cement removal. Conducted at the Institute of Geonics, Ostrava, Czech Republic, the study explores the effect of varying standoff distances on erosion profiles, under controlled parameters including a fixed nozzle diameter, sonotrode frequency, supply pressure, and robot arm velocity. The dataset includes numerical data representing ablation profiles, captured as a large CSV file, and audio recordings captured using a high-resolution microphone. Ablation profiles are measured at standoff distances that span several discrete lengths, ensuring diverse sample conditions. The audio data derives from high-fidelity recordings at a 38.4 kHz sampling rate in .wav format, capturing the fluid jet's impact sounds as it interacts initially with a metal plate and subsequently with the bone cement. This audio information was processed into Mel Spectrograms,
Dataset of NP-Hard Job Shop Scheduling Problem (JSSP), specifically designed for LLM fine-tuning.
SOTIF-PCOD is a dataset generated using the CARLA simulator, specifically designed for Safety of the Intended Functionality (SOTIF) research. It consists of 547 frames of LiDAR point cloud data formatted in the KITTI standard, representing a single SOTIF-related use case.
Precision Liming Soil Datasets (LimeSoDa) is a collection of 31 datasets from a field- and farm-scale soil mapping context. These datasets are "ready-to-use" for modeling purposes, as they include target soil properties and features in a tidy tabular format. Three target soil properties are present in every dataset: (1) soil organic matter (SOM) or soil organic carbon (SOC), (2) pH, and (3) clay content, while the features for modeling are dataset-specific. The primary goal of LimeSoDa is to enable more reliable benchmarking of machine learning methods in digital soil mapping and pedometrics. All the associated materials and data from LimeSoDa can be downloaded in Zenodo data repository or using the R or Python package implementations. However, for a more in-depth analysis, we refer to the published paper "LimeSoDa: A Dataset Collection for Benchmarking of Machine Learning Regressors in Digital Soil Mapping" by Schmidinger et al. (2025).
Click to add a brief description of the dataset (Markdown and LaTeX enabled).