The PubMed dataset consists of 19717 scientific publications from PubMed database pertaining to diabetes classified into one of three classes. The citation network consists of 44338 links. Each publication in the dataset is described by a TF/IDF weighted word vector from a dictionary which consists of 500 unique words. Source: https://linqs.soe.ucsc.edu/data

SciCite is a dataset of citation intents that addresses multiple scientific domains and is more than five times larger than ACL-ARC. Source: Structural Scaffolds for Citation Intent Classification in Scientific Publications Image Source: https://arxiv.org/pdf/1904.01608v2.pdf

PubMed 200k RCT is new dataset based on PubMed for sequential sentence classification. The dataset consists of approximately 200,000 abstracts of randomized controlled trials, totaling 2.3 million sentences. Each sentence of each abstract is labeled with their role in the abstract using one of the following classes: background, objective, method, result, or conclusion. The purpose of releasing this dataset is twofold. First, the majority of datasets for sequential short-text classification (i.e., classification of short texts that appear in sequences) are small: the authors hope that releasing a new large dataset will help develop more accurate algorithms for this task. Second, from an application perspective, researchers need better tools to efficiently skim through the literature. Automatically classifying each sentence in an abstract would help researchers read abstracts more efficiently, especially in fields where abstracts may be long, such as the medical field. Source: GitHub Image Source: https://arxiv.org/pdf/1710.06071.pdf

ACL Anthology Reference Corpus (ACL ARC) is a collection of 10,920 academic papers from the ACL Anthology. ACL ARC is cleaned to remove: - files that look like not full papers, paper fragments, foreign-language papers (e.g., French), or pure junk. - headers (title and author information; NOT abstract). - footers ("References" line and the actual references). - some bad characters (spurious characters). - some page numbers (i.e., a single number appearing on a line, with nothing else attached to it). - significant foreign-language (e.g., French) content in an otherwise English paper. The cleaned corpus has 10,628 documents. Source: ACL ARC

A SemEval shared task in which participants must extract definitions from free text using a term-definition pair corpus that reflects the complex reality of definitions in natural language. Source: SemEval-2020 Task 6: Definition extraction from free text with the DEFT corpus

The IndoNLU benchmark is a collection of resources for training, evaluating, and analyzing natural language understanding systems for Bahasa Indonesia. It is a joint venture from many Indonesia NLP enthusiasts from different institutions such as Gojek, Institut Teknologi Bandung, HKUST, Universitas Multimedia Nusantara, Prosa.ai, and Universitas Indonesia. Source: IndoNLU: Benchmark and Resources for Evaluating Indonesian Natural Language Understanding

Although deep face recognition has achieved impressive results in recent years, there is increasing controversy regarding racial and gender bias of the models, questioning their trustworthiness and deployment into sensitive scenarios. DemogPairs is a validation set with 10.8K facial images and 58.3M identity verification pairs, distributed in demographically-balanced folds of Asian, Black and White females and males. We also propose a benchmark of experiments using DemogPairs over state-of-the-art deep face recognition models in order to analyze their cross-demographic behavior and potential demographic biases (see figure below).

CSPubSum is a dataset for summarisation of computer science publications, created by exploiting a large resource of author provided summaries and show straightforward ways of extending it further. Source: A Supervised Approach to Extractive Summarisation of Scientific Papers

PcMSP is a dataset annotated from 305 open access scientific articles for material science information extraction that simultaneously contains the synthesis sentences extracted from the experimental paragraphs, as well as the entity mentions and intra-sentence relations. Source: PcMSP: A Dataset for Scientific Action Graphs Extraction from Polycrystalline Materials Synthesis Procedure Text Image Source: https://arxiv.org/pdf/2210.12401v1.pdf

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

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.

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

E2E Refined is a dataset for sentence classification. It consists of 40,560 examples for training, 4,489 for validation, and 4,555 for test. It is a refined version of the well-known MR-to-text E2E dataset where many deletion/insertion/substitution errors has been fixed.

A Racial Fairness Benchmark Dataset for Face Forgery Detection.

The first annotated corpus for multilingual analysis of potentially unfair clauses in online Terms of Service. The data set comprises a total of 100 contracts, obtained from 25 documents annotated in four different languages: English, German, Italian, and Polish. For each contract, potentially unfair clauses for the consumer are annotated, for nine different unfairness categories.

Paper Field is built from the Microsoft Academic Graph and maps paper titles to one of 7 fields of study. Each field of study - geography, politics, economics, business, sociology, medicine, and psychology - has approximately 12K training examples.

Press Briefing Claim Dataset The dataset contains a total of 53 press briefings from a time span of over four years (2017-2021). While, on average, one press briefing per month is held, the distribution is highly skewed towards recent years. The Press briefings can be categorized into five main thematic categories: Climate, Energy and Mobility, Medicine, Technology and Science. The press briefings are unevenly distributed among the five categories with a major focus on medical press briefings, which also reflects the COVID-19 pandemic. In total, 177 speakers, excluding the hosts, were detected in the dataset. Most of the time the speakers are invited experts. Excluding the top 10 percentile press briefings as outliers, a press briefing had four guests on average. The press briefings consist of 3066 passages, 1719 from guests and journalists and 1294 from hosts. Regarding the sentences, 25040 sentences were collected in total, 5955 from hosts and 18942 from guests.