Papers with Code 아카이브

Classification

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

하위 태스크
24
평가표
72
보존된 평가 기록
134

Classification is the task of categorizing a set of data into predefined classes or groups. The aim of classification is to train a model to correctly predict the class or group of new, unseen data. The model is trained on a labeled dataset where each instance is assigned a class label. The learning algorithm then builds a mapping between the features of the data and the class labels. This mapping is then used to predict the class label of new, unseen data points. The quality of the prediction is usually evaluated using metrics such as accuracy, precision, and recall.

데이터셋

출처에 같은 태스크 이름으로 연결된 데이터셋입니다. 동명 태스크가 포함될 수 있으므로, 비교할 때는 평가 결과의 조건도 확인하세요.

Autooral dataset

A multi-tasking oral ulcer dataset (Autooral dataset) is proposed. Autooral dataset contains two major tasks of disease segmentation and classification. The labeling of the Autooral dataset was done by three experienced dentists. At the end of the annotation, we formed 420 images of oral data with high quality after cropping and removal operations. We standardize the image size to 256*256. The original image is a 24-bit RGB image, the ground truth for the segmentation task is an 8-bit image, and there are five different disease types for the classification task (including cancerous ulcers, traumatic ulcers and traumatic blood blisters, herpes-like aphthous ulcers, mild aphthous ulcers, severe aphthous ulcers). The ratio of the 5 different ulcer types for the classification task was 9:9:15:18:22 (with a few exclusions). Further, by chi-square test, there were significant differences in gender (p=0.04) and age (p=0.01) of the patients among the 5 ulcer types. All-age coverage, a 13-year collection interval, and the presence of 12 underlying diseases, among other things, demonstrated that we had sufficient sample diversity.

Avicenna: Deductive Commonsense Reasoning

A syllogism is a common form of deductive reasoning that requires precisely two premises and one conclusion. The Avicenna corpus is a benchmark for syllogistic NLI and syllogistic NLG: - syllogistic NLI: Identifying the possibility of inferring between pairs of inputted sentences. - syllogistic NLG: Generating a conclusion sentence for two sentences with a syllogistic relation.

B-XAIC

B-XAIC consists of 50K small molecules represented as graphs and includes 7 graph classification tasks, each with ground truth labels and corresponding explanations.

BaitBuster-Bangla: A Comprehensive Dataset for Clickbait Detection in Bangla with Multi-Feature and Multi-Modal Analysis

The dataset contains a total of 253,070 records, with 18 features. The features are categorized into four different types: Metadata, Primary Data, Engagement Stats, and Label. Under the Metadata category contains basic information about the channel and video, such as their unique identifiers, date and time of publication, and thumbnail URLs. The Primary Data category contains information about the title and description of the video. The "Processed" columns refer to the cleaned data after denoising, deduplication and debiased for further analysis. The Engagement Stats category contains data on user engagement metrics for each video. The Label category contains predefined auto labels, human annotated labels, and AI generated pseudo labels. Auto labels are labels that are automatically derived based on a review of their titles, descriptions, and thumbnails over time. Channels with consistently misleading, exaggerated, or sensationalized content were labeled as clickbait. Those focusing on factual information delivery without emotional appeals were labeled non-clickbait. Human labels are labels that are manually derived by volunteer human annotators and AI labels are labels that are generated by a fine-tuned AI model. The following table presents a detailed overview and definitions of the features. | Feature Type | Feature Name | Data Type | Definition | |----------------------------|----------------------|---------------|--------------------------------------------------------------| | Metadata | channel_id | string | ID of the YouTube channel | | Metadata | channel_name | string | Name of the YouTube channel | | Metadata | channel_url | string | URL of the YouTube channel | | Metadata | video_id | string | ID of the video | | Metadata | publishedAt | datetime | Date and time when the video was published | | Primary Data | title | string | Title of the video | | Primary Data (Processed) | title_debiased | string | Debiased title of the video | | Primary Data | description | string | Debiased description of the video | | Primary Data (Processed) | description_debiased | string | Description of the YouTube video without bias | | Metadata | url | string | URL of the video | | Engagement Stats | viewCount | int | Number of views the video has received | | Engagement Stats | commentCount | int | Number of comments on the video | | Engagement Stats | likeCount | int | Number of likes on the video | | Engagement Stats | dislikeCount | int | Number of dislikes on the video | | Metadata | thumbnails | string | URL of the thumbnail for the video | | Label | auto_labeled | string | Automatically labeled using manual review | | Label (Processed) | human_labeled | string | Labeled by human | | Label (Processed) | ai_labeled | string | Labeled by an AI model fine-tuned on human labeled data | ## Paper * Data in Brief: https://doi.org/10.1016/j.dib.2024.110239 * arXiv Link: https://arxiv.org/abs/2310.11465 ## Dataset * Mendeley: https://data.mendeley.com/datasets/3c6ztw5nft/ ## Citation ### MLA Al Imran, Abdullah, Md Sakib Hossain Shovon, and M. F. Mridha. "BaitBuster-Bangla: A Comprehensive Dataset for Clickbait Detection in Bangla with Multi-Feature and Multi-Modal Analysis." Data in Brief (2024): 110239. ### BibText @article{IMRAN2024110239, title = {BaitBuster-Bangla: A Comprehensive Dataset for Clickbait Detection in Bangla with Multi-Feature and Multi-Modal Analysis}, journal = {Data in Brief}, pages = {110239}, year = {2024}, issn = {2352-3409}, doi = {https://doi.org/10.1016/j.dib.2024.110239}, url = {https://www.sciencedirect.com/science/article/pii/S2352340924002105}, author = {Abdullah Al Imran and Md Sakib Hossain Shovon and M.F. Mridha}, keywords = {Bangla clickbait dataset, YouTube clickbait, Multi-modal clickbait dataset, Multi-feature clickbait dataset, Bangla natural language processing, User behavior modeling, Social Media Analysis}, abstract = {This study presents a large multi-modal Bangla YouTube clickbait dataset consisting of 253,070 data points collected through an automated process using the YouTube API and Python web automation frameworks. The dataset contains 18 diverse features categorized into metadata, primary content, engagement statistics, and labels for individual videos from 58 Bangla YouTube channels. A rigorous preprocessing step has been applied to denoise, deduplicate, and remove bias from the features, ensuring unbiased and reliable analysis. As the largest and most robust clickbait corpus in Bangla to date, this dataset provides significant value for natural language processing and data science researchers seeking to advance modeling of clickbait phenomena in low-resource languages. Its multi-modal nature allows for comprehensive analyses of clickbait across content, user interactions, and linguistic dimensions to develop more sophisticated detection methods with cross-linguistic applications.} }

Bala-Copa

The Balanced Choice of Plausible Alternatives dataset is a benchmark for training machine learning models that are robust to superficial cues/spurious correlations. The dataset extends the COPA dataset(Roemmele et al. 2011) with mirrored instances that mitigate against token-level superficial cues in the original COPA answers. The superficial cues in the original COPA datasets result from an unbalanced token distribution between the correct and the incorrect answer choices, i.e., some tokens appear more in the correct choices than the incorrect ones. Balanced COPA equalizes the token distribution by adding mirrored instances with identical answer choices but different labels. The details about the creation of Balanced COPA and the implementation of the baselines are available in the paper.

BanglaBook

This repository contains the code, data, and models of the paper titled "BᴀɴɢʟᴀBᴏᴏᴋ: A Large-scale Bangla Dataset for Sentiment Analysis from Book Reviews" published in the Findings of the Association for Computational Linguistics: ACL 2023. arXiv anthology GoogleScholar ResearchGate GitHub HuggingFace PDF Slides Video License: Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license ## Data Format Each row consists of a book review sample. The table below describes what each column signifies. Column Title | Description ------------ | ------------- id | The unique identification number of the sample Book_Name | The title of the book that has been evaluated by the review Writer_Name | The name of the book's author Category | The genre to which the book belongs Rating | A numerical value $r$ such that $1\leq r \leq 5$<br>A score reflecting the reviewer's subjective assessment of the book's quality Review | The review text written by the reviewer Site | The name of the online bookshop sentiment | The conveyed sentiment and class label of the review<br>For a review sample $i$ with rating $r_i$, the sentiment label $S_i$ is,<br>$S_i =\begin{cases}Negative, & \text{if $r_i \leq 2$}\\\Neutral, & \text{if $r_i = 3$}\\\Positive, & \text{if $r_i \geq 4$}\end{cases}$ label | The numerical representation of the sentiment label<br>For a review sample $i$ with sentiment label $S_i$, the numerical label is,<br>$label_i =\begin{cases}0, & \text{if $S_i = Negative$}\\\1, & \text{if $S_i = Neutral$}\\\2, & \text{if $S_i = Positive$}\end{cases}$ ## Data Construction ### Data Collection Process For the data collection and preparation process of the BᴀɴɢʟᴀBᴏᴏᴋ dataset, we first compile a list of URLs for authors from online bookstores. From there, we procure URLs for the books. We meticulously scrape information such as book titles, author names, book categories, review texts, reviewer names, review dates, and ratings by utilizing these book URLs. <img src="https://github.com/mohsinulkabir14/BanglaBook/raw/main/images/banglabookgithub1.png" alt="drawing" style="width:1000px;"/> ### Labeling, Translation, and Validation of the Curated Samples If a review does not have a rating, we deem it unannotated. Reviews with a rating of 1 or 2 are classified as negative, a rating of 3 is considered neutral, and a rating of 4 or 5 is classified as positive. After discarding the unannotated reviews, we curate a final dataset of 158,065 annotated reviews. Of these, 89,371 are written entirely in Bangla. The remaining 68,694 reviews were written in Romanized Bangla, English, or a mix of languages. They are translated into Bangla with Google Translator and a custom Python program using the googletrans library. The translations are subsequently subjected to manual review and scrutiny to confirm their accuracy. <img src="https://github.com/mohsinulkabir14/BanglaBook/raw/main/images/banglabookgithub2.png" alt="drawing" style="width:1000px;"/> ## Results <img src="https://github.com/mohsinulkabir14/BanglaBook/raw/main/images/banglabookgithub3.png" alt="drawing" style="width:1000px;"/> ## Citation If you find this work useful, please cite our paper: bib @inproceedings{kabir-etal-2023-banglabook, title = "{B}angla{B}ook: A Large-scale {B}angla Dataset for Sentiment Analysis from Book Reviews", author = "Kabir, Mohsinul and Bin Mahfuz, Obayed and Raiyan, Syed Rifat and Mahmud, Hasan and Hasan, Md Kamrul", booktitle = "Findings of the Association for Computational Linguistics: ACL 2023", month = jul, year = "2023", address = "Toronto, Canada", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2023.findings-acl.80", pages = "1237--1247", abstract = "The analysis of consumer sentiment, as expressed through reviews, can provide a wealth of insight regarding the quality of a product. While the study of sentiment analysis has been widely explored in many popular languages, relatively less attention has been given to the Bangla language, mostly due to a lack of relevant data and cross-domain adaptability. To address this limitation, we present BanglaBook, a large-scale dataset of Bangla book reviews consisting of 158,065 samples classified into three broad categories: positive, negative, and neutral. We provide a detailed statistical analysis of the dataset and employ a range of machine learning models to establish baselines including SVM, LSTM, and Bangla-BERT. Our findings demonstrate a substantial performance advantage of pre-trained models over models that rely on manually crafted features, emphasizing the necessity for additional training resources in this domain. Additionally, we conduct an in-depth error analysis by examining sentiment unigrams, which may provide insight into common classification errors in under-resourced languages like Bangla. Our codes and data are publicly available at https://github.com/mohsinulkabir14/BanglaBook.", }

BanglaEmotion

BanglaEmotion is a manually annotated Bangla Emotion corpus, which incorporates the diversity of fine-grained emotion expressions in social-media text. More fine-grained emotion labels are considered such as Sadness, Happiness, Disgust, Surprise, Fear and Anger - which are, according to Paul Ekman (1999), the six basic emotion categories. For this task, a large amount of raw text data are collected from the user’s comments on two different Facebook groups (Ekattor TV and Airport Magistrates) and from the public post of a popular blogger and activist Dr. Imran H Sarker. These comments are mostly reactions to ongoing socio-political issues and towards the economic success and failure of Bangladesh. A total of 32923 comments are scraped from the three sources aforementioned above. Out of these, a total of 6314 comments were annotated into the six categories. The distribution of the annotated corpus is as follows: sad = 1341 happy = 1908 disgust = 703 surprise = 562 fear = 384 angry = 1416 A balanced set is also provided from the above data and split the dataset into training and test set of equal ratio. A proportion of 5:1 is used for training and evaluation purposes. More information on the dataset and the experiments on it could be found in our paper (related links below).

BASIR

Government fiscal policies, particularly annual union budgets, exert significant influence on financial markets. However, real-time analysis of budgetary impacts on sector-specific equity performance remains methodologically challenging and largely unexplored. This study proposes a framework to systematically identify and rank sectors poised to benefit from India's Union Budget announcements. The framework addresses two core tasks: (1) multi-label classification of excerpts from budget transcripts into 81 predefined economic sectors, and (2) performance ranking of these sectors. Leveraging a comprehensive corpus of Indian Union Budget transcripts from 1947 to 2025, we introduce BASIR (Budget-Assisted Sectoral Impact Ranking), an annotated dataset mapping excerpts from budgetary transcripts to sectoral impacts.

Bengali Ekman's Six Basic Emotions Corpus

The dataset contains 36000 Bangla data based on Ekman's six basic emotions. This data was first introduced in the paper Alternative non-BERT model choices for the textual classification in low-resource languages and environments. The whole dataset is balanced and evenly distributed among all the six classes.

BFN

This database is a database of backdoored neural networks intended for face recognition. The networks are of the FaceNet architecture and are trained on Casia-WebFace, with and without additional samples (which are the source of the backdoor). More information regarding backdoors and the project within which this fits can be found in the public release of the source code : https://gitlab.idiap.ch/bob/bob.paper.backdoored_facenets.biosig2022. There are two sets of backdoored networks. A first one with backdoors with varying triggers (the triggers dataset) and a second one with backdoors with varying trigger placement strategies (the locations dataset). A third set of networks is also provided, just regular networks without any backdoor, referred to as the clean dataset. Configuration yaml files are provided to replicate the backdoored networks using the repository content linked above, in addition to pickle files containing validation scores on all validation datasets. The purpose of this dataset is to allow for evaluation of backdoored network detection work on face recognition networks. The backdoored networks here are trained from a clean checkpoint and finetuned on poisoned data. The characteristics of each network is provided in the yaml file colocated with the checkpoint. Multiple triggers (organic and synthetic) are explored, in addition to multiple placement strategies (systematically random, static, in context etc). The finetuning is done exclusively on the layers reported in the yaml file.

BFRD

This is a binary dataset used for Bengali fake review detection in the paper "Bengali Fake Reviews: A Benchmark Dataset and Detection System" accepted in Neurocomputing, a journal published by Elsevier. - Annotated by 4 native Bangla speakers with more than 90% trustworthiness score. - Fleiss' Kappa Score: 0.83 ## Number of Total Data - Fake - 1339 - Non-fake - 7710 ## Class wise statistics of BFRD dataset | Statistics | Fake | Non-fake | |--------------------------|---------|-------------| | Total words | 1,55,789 | 9,27,902 | | Total unique words | 17,739 | 51,200 | | Max Review length | 693 | 1,614 | | Avg number of words | 116.35 | 120.35 | | Avg number of unique words | 84.99 | 88.42 | ## Citation If you use the datasets, please cite the following paper: @article{SHAHARIAR2024127732, title = {Bengali fake reviews: A benchmark dataset and detection system}, journal = {Neurocomputing}, pages = {127732}, year = {2024}, issn = {0925-2312}, doi = {https://doi.org/10.1016/j.neucom.2024.127732}, url = {https://www.sciencedirect.com/science/article/pii/S0925231224005034}, author = {G.M. Shahariar and Md. Tanvir Rouf Shawon and Faisal Muhammad Shah and Mohammad Shafiul Alam and Md. Shahriar Mahbub}, keywords = {Bengali fake reviews detection, Ensemble learning, Transformers, Deep learning, Augmentation, Transliteration}, abstract = {The proliferation of fake reviews on various online platforms has created a major concern for both consumers and businesses. Such reviews can deceive customers and cause damage to the reputation of products or services, making it crucial to identify them. Although the detection of fake reviews has been extensively studied in English language, detecting fake reviews in non-English languages such as Bengali is still a relatively unexplored research area. The novelty of the study unfolds on three fronts: (i) a new publicly available dataset called Bengali Fake Review Detection (BFRD) dataset is introduced, (ii) a unique pipeline has been proposed that translates English words to their corresponding Bengali meaning and also back transliterates Romanized Bengali to Bengali, (iii) a weighted ensemble model that combines four pre-trained transformers model is proposed. The developed dataset consists of 7710 non-fake and 1339 fake food-related reviews collected from social media posts. Rigorous experiments have been conducted to compare multiple deep learning and pre-trained transformer language models and our proposed model to identify the best-performing model. According to the experimental results, the proposed ensemble model attained a weighted F1-score of 0.9843 on a dataset of 13,390 reviews, comprising 1339 actual fake reviews, 5,356 augmented fake reviews, and 6695 reviews randomly selected from the 7710 non-fake instances.} }

BGG dataset

We recorded gun sounds by changing the type and position of guns to diversify distances and angles in the PUBG environment. The BGG dataset consists of 2,195 samples with 37 different types of guns and five directions, including a silence in which there is no gunfire, but noises exist. The distance from the firearms ranged from 0 meters to 600 meters. The audio was recorded in stereo (i.e., two-channel audio), and each sample contains various environmental noises (e.g., water splashing, walking, and bullet friction).

BRISC

BRISC is a high-quality, expert-annotated MRI dataset curated for brain tumor segmentation and classification. It addresses common limitations in existing datasets (e.g., BraTS, Figshare), including class imbalance, narrow tumor focus, and annotation inconsistencies.

BTS

The Building TimeSeries (BTS) dataset covers three buildings over a three-year period, comprising more than ten thousand timeseries data points with hundreds of unique ontologies. Moreover, the metadata is standardised in the formed of knowledge graph using the Brick schema. https://github.com/cruiseresearchgroup/DIEF_BTS

BugClassify

Dataset of 5,591 labeled issue tickets. Originally created by Herzig et al. in : "It’s Not a Bug, It’s a Feature: How Misclassification Impacts Bug Prediction" (paper)

CAER-Dynamic

13,201 clips from 79 TV shows. Each video clip was manually annotated with six emotion categories, including “anger”, “disgust”, “fear”, “happy”, “sad”, and “surprise“, as well as “neutral”.

Cards Against Humanity

A dataset of games played in the card game "Cards Against Humanity" (CAH), by human players, derived from the online CAH labs. Each round includes the cards presented to users - a "black" prompt with a blank or question and 10 "white" punchlines as possible responses, and which punchline was picked by a player each round, along with text and metadata. An example prompt is “TSA guidelines now prohibit ___ on airplanes”. Candidate punchlines are “Goblins”, “BATMAN!!!”, “Poor people”, and “The right amount of cocaine”. Importantly, many cards are offensive or politically incorrect. Used to explore human humor preferences. Available upon request from CAH labs: mail@cardsagainsthumanity.com Train/test splits and data processing available in the paper/code: "Cards Against AI: Predicting Humor in a Fill-in-the-blank Party Game": https://github.com/ddofer/CAH

CareerCoach 2022

The CareerCoach 2022 gold standard is available for download in the NIF and JSON format, and draws upon documents from a corpus of over 99,000 education courses which have been retrieved from 488 different education providers. The corpus contains two partitions.: * Partition (P1) supports the content extraction (i.e., text segmentation and text segment classification) tasks and comprises 169 documents and gold standard annotations for page segments * Partition (P2) contains 75 documents with a significantly richer set of annotations that consider content extraction, entities and slots. It supports benchmarking knowledge extraction tasks such as entity recognition, entity classification, entity linking, and slot filling on top of the content extraction task.

Caselaw4

Caselaw4 is a dataset of 350k common law judicial decisions from the U.S. Caselaw Access Project, of which 250k have been automatically annotated with binary outcome labels of AFFIRM and REVERSE. The court case reports used in the dataset are from New Mexico, North Carolina, Illinois, and Arkansas Courts of Appeal. These Courts hear appeals exclusively from lower courts within their respective states, on matters of domestic state law, and the data for these jurisdictions are freely available. Since each case in Caselaw4 appeals some lower court ruling, the possible outcomes of each case are as follows: - the previous ruling is kept as is (AFFIRM); - the previous ruling is changed/annulled (REVERSE); - some parts of the previous ruling are kept and some are changed (MIXED); - the appeal is dismissed (a type of AFFIRM). The data in Caselaw4 are stored in JSON format. In addition to the original metadata about the case name, date, court, judges, cases cited etc., we (a) automatically annotated a subset of 250k cases with the AFFIRM or REVERSE outcome label (with weighted average precision of 95.45%), and (b) manually annotated 500 cases from the New Mexico Court of Appeals with the AFFIRM, REVERSE, or MIXED outcome label as well as with the outcome sentences.

Chinese AI and Law (CAIL) 2018

Large-scale Chinese legal dataset for judgment prediction. \dataset contains more than 2.6 million criminal cases published by the Supreme People's Court of China, which are several times larger than other datasets in existing works on judgment prediction. Source: CAIL2018: A Large-Scale Legal Dataset for Judgment Prediction

CHIP-CTC

CHIP Clinical Trial Classification, a dataset aimed at classifying clinical trials eligibility criteria, which are fundamental guidelines of clinical trials defined to identify whether a subject meets a clinical trial or not, is used for the CHIP-CTC task. All text data are collected from the website of the Chinese Clinical Trial Registry (ChiCTR) , and a total of 44 categories are defined. The task is like text classification; although it is not a new task, studies and corpus for the Chinese clinical trial criterion are still limited, and we hope to promote future researches for social benefits.

Coil100-Augmented

This dataset derives from Coil100. There are more than 1,1M images of 100 objects. Each object was turned on a turnable through 360 degrees to vary object pose with respect to a fixed color camera. Images of the objects were taken at pose intervals of 5 degrees. This corresponds to 72 poses per object. Then planar rotation (9 angles) and 18 scaling factors has been applied. Objects have a wide variety of complex geometric and reflectance characteristics. This augmented version of Coil100 has been design for Disentangled Representation Learning for real images, the Factors of Variations are: | Factors | # values | |----------|----------| | Object | 100 | | 3D Pose | 72 | | Rotation | 9 | | Scale | 18 | The binarized version is also available.

Color-connectivity

Synthetic graph classification datasets with the task of recognizing the connectivity of same-colored nodes in 4 graphs of varying topology. * The four Color-connectivity datasets were created by taking a graph and randomly coloring half of its nodes one color, e.g., red, and the other nodes blue, such that the red nodes either form a single connected island or two disjoint islands. The binary classification task is then distinguishing between these two cases. The node colorings were sampled by running two red-coloring random walks starting from two random nodes. * For the underlying graph topology we used: 1) 16x16 2D grid, 2) 32x32 2D grid, 3) Euroroad road network (Šubelj et al. 2011), and 4) Minnesota road network. * We sampled a balanced set of 15,000 coloring examples for each graph, except for Minnesota network for which we generated 6,000 examples due to memory constraints. * The Color-connectivity task requires combination of local and long-range graph information processing to which most existing message-passing Graph Neural Networks (GNNs) do not scale. These datasets can serve as a common-sense validation for new and more powerful GNN methods. These testbed datasets can still be improved, as the node features are minimal (only a binary color) and recognition of particular topological patterns (e.g., rings or other subgraphs) is not needed to solve the task.

Colors

A large dataset of color names and their respective RGB values stores in CSV.

CORBEL

Dataset included measuring static tension under 2 kg load in different points of the CB and measurements in dynamic conditions. The latter conditions presumed the range of the linear belt speeds between nu_1 = 0.5 and nu_max = 1.7 m/s. 400 Hz unified sampling frequency for the experiments. It corresponded with 140 samples. Signal is divided into fixed-length (points P) signals of 0.2 s (80 points), 0.4 s (160 points), 0.8 s (320 points), 1.6 s (640 points), 3.2 s (1280 points), and 5.0 s (2000 points). In such for each length equivalent data points are avaible: Signal length (s), time points (#), Samples (#) 0.2, 80, 384200 0.4, 160, 368200 0.8, 320, 336200 1.6, 640, 272200 3.2, 1280, 144200 5.0, 2000, 200 The binary classification into loaded/no load task is investigated.

COVID-19 Tweets with Motivation and Topics

The dataset contains Tweet IDs along with the location and tweet timestamp. The tweets are labeled based on motivating/demotivating status, stance towards the COVID-19 vaccine, and topic in the tweet text. To comply with Twitter guidelines, we removed the tweet texts and author information.

CPCXR

The COVID-19 Posteroanterior Chest X-Ray fused (CPCXR) dataset is generated by the fusion of three publicly available datasets: COVID-19 cxr image, Radiological Society of North America (RSNA), and U.S. national library of medicine (USNLM) collected Montgomery country - NLM(MC). The dataset consists of samples of diseases labeled as COVID-19, Tuberculosis, Other pneumonia (SARS, MERS, etc.), and Normal. The dataset can be utilized to train an evaulate deep learning and machine learning models as binary and multi-class classification problem. Source: https://github.com/nspunn1993/COVID-19-PA-CXR-fused-dataset

CRCDX

Histological images of colorectal cancer, derived from the TCGA database

CSAbstruct Dataset

CSAbstruct is a new dataset of annotated computer science abstracts with sentence labels according to their rhetorical roles. The key difference between this dataset and PUBMED-RCT is that PubMed abstracts are written according to a predefined structure, whereas computer science papers are free-form. Therefore, there is more variety in writing styles in CSABSTRUCT. CSABSTRUCT is collected from the Semantic Scholar corpus (Ammar et al., 2018). Each sentence is annotated by 5 workers on the Figure-eight platform,6 with one of 5 categories {BACKGROUND, OBJECTIVE, METHOD, RESULT, OTHER}. Source: Pretrained Language Models for Sequential Sentence Classification

CSL (Chinese Scientific Literature)

We present CSL, a large-scale Chinese Scientific Literature dataset, which contains the titles, abstracts, keywords and academic fields of 396,209 papers. To our knowledge, CSL is the first scientific document dataset in Chinese. Paper | Code and data ## Dataset We obtain the paper's meta-information from the National Engineering Research Center for Science and Technology Resources Sharing Service (NSTR) dated from 2010 to 2020. Then, we filter data by the Catalogue of Chinese Core Journals. According to the Catalogue and collected data, we divide academic fields into 13 first-level categories (e.g., Engineering, Science) and 67 second-level disciplines (e.g., Mechanics, Mathematics). In total, we collect 396,209 instances for the CSL dataset, represented as tuples <T, A, K, c, d>, where T is the title, A is the abstract, K is a list of keywords, c is the category label and d is the discipline label. The paper distribution over categories and the examples of disciplines are shown in below: | Category | #d | len(T) | len(A) | num(K) | #Samples | Discipline Examples | |-----------------|-------------:|-------:|-------:|-------:|----------:|---------------------------------------| | Engineering | 27 | 19.1 | 210.9 | 4.4 | 177,600 | Mechanics,Architecture,Electrical Science | | Science | 9 | 20.7 | 254.4 | 4.3 | 35,766 | Mathematics,Physics,Astronomy,Geography | | Agriculture | 7 | 17.1 | 177.1 | 7.1 | 39,560 | Crop Science,Horticulture,Forestry | | Medicine | 5 | 20.7 | 269.5 | 4.7 | 36,783 | Clinical Medicine,Dental Medicine,Pharmacy | | Management | 4 | 18.7 | 157.7 | 6.2 | 23,630 | Business Management,Public Administration | | Jurisprudence | 4 | 18.9 | 174.4 | 6.1 | 21,554 | Legal Science,Political Science,Sociology | | Pedagogy | 3 | 17.7 | 179.4 | 4.3 | 16,720 | Pedagogy,Psychology,Physical Education | | Economics | 2 | 19.5 | 177.2 | 4.5 | 11,558 | Theoretical Economics,Applied Economics | | Literature | 2 | 18.8 | 158.2 | 8.3 | 10,501 | Chinese Literature,Journalism | | Art | 1 | 17.8 | 170.8 | 5.4 | 5,201 | Art | | History | 1 | 17.6 | 181.0 | 6.0 | 6,270 | History | | Strategics | 1 | 17.5 | 169.3 | 4.0 | 3,555 | Military Science | | Philosophy | 1 | 18.0 | 176.5 | 8.0 | 7,511 | Philosophy | | All | 67 | | | | 396,209 | ## Evaluation Tasks We build a benchmark to facilitate the development of Chinese scientific literature NLP. It contains diverse tasks, ranging from classification to text generation, representing many practical scenarios. We randomly select 100k samples and split the datasets into the training set, validation set and test set according to the ratio, 0.8 : 0.1 : 0.1. This split is shared across different tasks, which allows multitask training and evaluation. Datasets are presented in text2text format. #### 1.Text Summarization (Title Prediction) Predict the paper title from the abstract. Data examples: { "prompt": "to title", "text_a": "多个相邻场景同时进行干涉参数外定标的过程称为联合定标,联合定标能够 \ 保证相邻场景的高程衔接性,能够实现无控制点场景的干涉定标.该文提出了 \ 一种适用于机载InSAR系统的联合定标算法...", "text_b": "基于加权最优化模型的机载InSAR联合定标算法" } #### 2.Keyword Generation Predict a list of keywords from a given paper title and abstract. Data examples: { "prompt": "to keywords", "text_a": "通过对72个圆心角为120°的双跨偏心支承弯箱梁桥模型的计算分析,以梁 \ 格系法为基础编制的3D-BSA软件系统为结构计算工具,用统计分析的方法建 \ 立双跨偏心支承弯箱梁桥结构反应在使用极限状态及承载能力极限状态下与 \ 桥梁跨长... 偏心支承对120°圆心角双跨弯箱梁桥的影响", "text_b": "曲线桥_箱形梁_偏心支承_设计_经验公式" } #### 3.Category Classification Predict the category with the paper title (13 classes). Data examples: { "prompt": "to category", "text_a": "基于模糊C均值聚类的流动单元划分方法——以克拉玛依油田五3中区克下组为例", "text_b": "工学" }, { "prompt": "to category", "text_a": "正畸牵引联合牙槽外科矫治上颌尖牙埋伏阻生的临床观察", "text_b": "医学" } #### 4.Discipline Classification Predict the discipline with the paper abstract (67 classes). Data examples: { "prompt": "to discipline", "text_a": "某铁矿选矿厂所产铁精矿含硫超过0.3%,而现场为了今后发展的需要,要 \ 求将含硫量降到0.1%以下.为此,针对该铁精矿中硫化物主要以磁黄铁矿 \ 形式存在、硫化物多与铁矿物连生且氧化程度较高的特点...", "text_b": "矿业工程" }, { "prompt": "to discipline", "text_a": "为了校正广角镜头的桶形畸变,提出一种新的桶形畸变数字校正方法.它 \ 使用点阵样板校正的方法,根据畸变图和理想图中圆点的位置关系,得出 \ 畸变图像素在X轴和Y轴方向上的偏移量曲面...", "text_b": "计算机科学与技术" }

CUCO Database

Many research articles have explored the impact of surgical interventions on voice and speech evaluations, but advances are limited by the lack of publicly accessible datasets. To address this, a comprehensive corpus of 107 Spanish Castilian speakers was recorded, including control speakers and patients who underwent upper airway surgeries such as Tonsillectomy, Functional Endoscopic Sinus Surgery, and Septoplasty. The dataset contains 3,800 audio files, averaging 35.51 ± 5.91 recordings per patient. This resource enables systematic investigation of the effects of upper respiratory tract surgery on voice and speech. Previous studies using this corpus have shown no relevant changes in key acoustic parameters for sustained vowel phonation, consistent with initial hypotheses. However, the analysis of speech recordings, particularly nasalised segments, remains open for further research. Additionally, this dataset facilitates the study of the impact of upper airway surgery on speaker recognition and identification methods, and testing of anti-spoofing methodologies for improved robustness.

CVE

CVE stands for Common Vulnerabilities and Exposures. CVE is a glossary that classifies vulnerabilities. The glossary analyzes vulnerabilities and then uses the Common Vulnerability Scoring System (CVSS) to evaluate the threat level of a vulnerability. A CVE score is often used for prioritizing the security of vulnerabilities.

Daily and Sports Activities

The dataset comprises motion sensor data of 19 daily and sports activities each performed by 8 subjects in their own style for 5 minutes. Five Xsens MTx units are used on the torso, arms, and legs. Dataset Characteristics: Multivariate, Time-Series Associated Tasks: Classification, Clustering Feature Type: Real # Instances: 9120 # Features: 5625

DAPFAM

Dataset DAPFAM See the accompanying paper: Ayaou et al., 2025 — “DAPFAM: A Domain‑Aware Patent Retrieval Dataset Aggregated at the Family Level” (arXiv:2506.22141). ## Summary DAPFAM provides 1 247 balanced query patent families and 45 336 target families with forward/backward‑citation relevance labels (≈ 50 K pairs). Each relevant link is explicitly marked in‑domain or out‑of‑domain according to IPC 3‑char overlap, enabling rigorous cross‑domain evaluation. * Full text (title · abstract · claims · description) plus rich metadata for every family. * Multi‑jurisdictional, English‑only text (families may originate in US, JP, EP, CN, …). * Parquet qrel file: qrels_all.parquet. ## Dataset Structure corpus.parquet # 45 336 rows, targets – every original column from the paper queries.parquet # 1 247 rows, queries – same columns + abstract_keywords qrels_all.parquet # (all | in | out) four‑column tables → query_id · relevant_id · relevance_score · domain_rel ## Citation @misc{ayaou2025dapfam, title={DAPFAM: A Domain-Aware Patent Retrieval Dataset Aggregated at the Family Level}, author={Iliass Ayaou and Denis Cavallucci and Hicham Chibane}, year={2025}, eprint={2506.22141}, archivePrefix={arXiv}, primaryClass={cs.CL} } ## Quick Stats * Queries: 1,247 * Corpus (targets): 45,336 * Qrels (all): 49,869 * Qrels (in): 19,736 * Qrels (out): 5,193

Dataset of Propaganda Techniques of the State-Sponsored Information Operation of the People's Republic of China

This data is for the Mis2-KDD 2021 under review paper: Dataset of Propaganda Techniques of the State-Sponsored Information Operation of the People’s Republic of China We present our dataset that focuses on propaganda techniques in Mandarin based on a state-linked information operations dataset from the PRC released by Twitter in July 2019. The dataset consists of multi-label propaganda techniques of the sampled tweets. In total, we have 9,950 labeled tweets with 21 different propaganda techniques. The tweets are the state-linked information operations dataset from the PRC released by Twitter.

출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.