Papers with Code 아카이브

Text Summarization

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

하위 태스크
12
평가표
37
보존된 평가 기록
233

Text Summarization is a natural language processing (NLP) task that involves condensing a lengthy text document into a shorter, more compact version while still retaining the most important information and meaning. The goal is to produce a summary that accurately represents the content of the original text in a concise form.

There are different approaches to text summarization, including extractive methods that identify and extract important sentences or phrases from the text, and abstractive methods that generate new text based on the content of the original text.

데이터셋

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

PeerSum

PeerSum is a new MDS dataset using peer reviews of scientific publications. The dataset differs from the existing MDS datasets in that summaries (i.e., the meta-reviews) are highly abstractive and they are real summaries of the source documents. In PeerSum, we have reviews (with scores), comments and responses as the source documents and the meta-review (with an acceptance outcome) as the ground truth summary. Each sample of this dataset contains a summary, corresponding source documents and also other complementary information (e.g., review scores) for one paper. The second version of PeerSum (peersum_v2) has 16,308 samples, while there are 10,862 samples in the first version. The dataset is stored in the json format. For each sample, details are based on following keys with explanation: * paper_id: unique id for each sample * title: the title of the corresponding paper * abstract: paper abstract * score: final score of this paper (if there is not a final, it will be an average of review scores) * acceptance: acceptance of the paper (e.g., accept, reject or spotlight) * meta_review: meta-review of the paper and this is treated as the summary * reviews: [review_id, writer, content (rating, confidence, comment), replyto] review_id and replyto are for the conversation structure * label: train, val, test (8/1/1) For each review (i.e., official review, public comment, or author/reviewer response): * review_id: unique id of each review * writer: official_reviewer, public, author * content: (rating, confidence, comment) * replyto: connect to a review (review_id and replyto are for the conversation structure)

PoC

A dataset containing the documents, source and fusion sentences, and human annotations of points of correspondence between sentences. The dataset bridges the gap between coreference resolution and summarization. Source: Understanding Points of Correspondence between Sentences for Abstractive Summarization

SuMe

Can language models read biomedical texts and explain the biomedical mechanisms discussed? In this work we introduce a biomedical mechanism summarization task. Biomedical studies often investigate the mechanisms behind how one entity (e.g., a protein or a chemical) affects another in a biological context. The abstracts of these publications often include a focused set of sentences that present relevant supporting statements regarding such relationships, associated experimental evidence, and a concluding sentence that summarizes the mechanism underlying the relationship. We leverage this structure and create a summarization task, where the input is a collection of sentences and the main entities in an abstract, and the output includes the relationship and a sentence that summarizes the mechanism. Using a small amount of manually labeled mechanism sentences, we train a mechanism sentence classifier to filter a large biomedical abstract collection and create a summarization dataset with 22k instances. We also introduce conclusion sentence generation as a pretraining task with 611k instances. We benchmark the performance of large bio-domain language models. We find that while the pretraining task help improves performance, the best model produces acceptable mechanism outputs in only 32% of the instances, which shows the task presents significant challenges in biomedical language understanding and summarization.

VNDS

A single-document Vietnamese summarization dataset

AUTH at BioLaySumm 2024 - Bringing Scientific Content to Kids

Science Journal for Kids Data This repository contains a dataset of abstracts from the Science Journal for Kids website and the original academic papers. It includes metadata such as titles, URLs, reading levels, and links to the full academic papers. The dataset is designed to support research and analysis of educational content tailored for young learners. ## Data The dataset is a curated collection of 284 original scientific abstracts and their adapted abstracts for children. These summaries are sourced from the Science Journal for Kids (SJK), a platform dedicated to making scientific research accessible and engaging for younger audiences. The dataset covers a diverse range of subjects, including biology, chemistry, health, environmental science, and more. The CSV file is approximately 716 KB in size. ## Data Format The dataset includes the following columns: - Category: The category or categories to which the article belongs. This can include areas such as Biodiversity, Conservation, Biology, Chemistry, Health, and Environmental Science. - Title: The title of the article, providing a concise description of the main topic or finding of the research. - Kids Abstract: The abstract of the article aimed at a younger audience. - Abstract (Original academic paper): The abstract from the original academic paper. - URL (Original academic paper): The URL to access the original academic paper. - Reading Levels: The suggested reading levels for the lay summary, which may include Elementary school, Middle school, Lower high school, and Upper high school. ## Citing this Repository If you use this dataset in your research, please cite it using the following BibTeX entry: @inproceedings{BioLaySumm_2024, author = {Loukritia Stefanou and Tatiana Passali and Grigorios Tsoumakas}, title = {AUTH at BioLaySumm 2024: Bringing Scientific Content to Kids}, booktitle = {Proceedings of the ACL 2024 BioNLP Workshop}, year = {2024}, address = {Bangkok, Thailand}, note = {A paper presented at the BioLaySumm 2024 shared task on lay summarization of biomedical research articles.} }

AUTH at BioLaySumm 2024: Bringing Scientific Content to Kids

Science Journal for Kids Data This repository contains a dataset of abstracts from the Science Journal for Kids website and the original academic papers. It includes metadata such as titles, URLs, reading levels, and links to the full academic papers. The dataset is designed to support research and analysis of educational content tailored for young learners. ## Data The dataset is a curated collection of 284 original scientific abstracts and their adapted abstracts for children. These summaries are sourced from the Science Journal for Kids (SJK), a platform dedicated to making scientific research accessible and engaging for younger audiences. The dataset covers a diverse range of subjects, including biology, chemistry, health, environmental science, and more. The CSV file is approximately 716 KB in size. ## Data Format The dataset includes the following columns: - Category: The category or categories to which the article belongs. This can include areas such as Biodiversity, Conservation, Biology, Chemistry, Health, and Environmental Science. - Title: The title of the article, providing a concise description of the main topic or finding of the research. - Kids Abstract: The abstract of the article aimed at a younger audience. - Abstract (Original academic paper): The abstract from the original academic paper. - URL (Original academic paper): The URL to access the original academic paper. - Reading Levels: The suggested reading levels for the lay summary, which may include Elementary school, Middle school, Lower high school, and Upper high school. ## Citing this Repository If you use this dataset in your research, please cite it using the following BibTeX entry: @inproceedings{BioLaySumm_2024, author = {Loukritia Stefanou and Tatiana Passali and Grigorios Tsoumakas}, title = {AUTH at BioLaySumm 2024: Bringing Scientific Content to Kids}, booktitle = {Proceedings of the ACL 2024 BioNLP Workshop}, year = {2024}, address = {Bangkok, Thailand}, note = {A paper presented at the BioLaySumm 2024 shared task on lay summarization of biomedical research articles.} }

Bengali Curated News Summary Dataset

Click to add a brief description of the dataset (Markdown and LaTeX enabled). Provide: * a high-level explanation of the dataset characteristics * explain motivations and summary of its content * potential use cases of the dataset

BrWac2Wiki

This is a dataset for multi-document summarization in Portuguese, what means that it has examples of multiple documents (input) related to human-written summaries (output). In particular, it has entries of multiple related texts from Brazilian websites about a subject, and the summary is the Portuguese Wikipedia lead section on the same subject (lead: the first section, i.e., summary, of any Wipedia article). Input texts were extracted from BrWac corpus, and the output from Brazilian Wikipedia dumps page. BrWac2Wiki contains 114.652 examples of (documents, wikipedia) pairs! So it is suitable for training and validating AI models for multi-document summarization in Portuguese. More information on the paper "PLSUM: Generating PT-BR Wikipedia by Summarizing Websites", by André Seidel Oliveira¹ and Anna Helena Reali Costa¹, that is going to be presented at ENIAC 2021. Our work is inspired by WikiSum, a similar dataset for the English language.

ComSum

ComSum is a data set of 7 million commit messages for text summarization. When documenting commits, software code changes, both a message and its summary are posted. These messages are gathered and filtered to curate developers' work summarization data set.

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" }

DUC 2006

There is currently much interest and activity aimed at building powerful multi-purpose information systems. The agencies involved include DARPA, ARDA and NIST. Their programmes, for example DARPA's TIDES (Translingual Information Detection Extraction and Summarization) programme, ARDA's Advanced Question & Answering Program and NIST's TREC (Text Retrieval Conferences) programme cover a range of subprogrammes. These focus on different tasks requiring their own evaluation designs. Within TIDES and among other researchers interested in document understanding, a group grew up which has been focusing on summarization and the evaluation of summarization systems. Part of the initial evaluation for TIDES called for a workshop to be held in the fall of 2000 to explore different ways of summarizing a common set of documents. Additionally a road mapping effort was started in March of 2000 to lay plans for a long-term evaluation effort in summarization. Out of the initial workshop and the roadmapping effort has grown a continuing evaluation in the area of text summarization called the Document Understanding Conferences (DUC). Sponsored by the Advanced Research and Development Activity (ARDA), the conference series is run by the National Institute of Standards and Technology (NIST) to further progress in summarization and enable researchers to participate in large-scale experiments.

Famous Keyword Twitter Replies

The "Famous Keyword Twitter Replies Dataset" is a comprehensive collection of Twitter data that focuses on popular keywords and their associated replies. This dataset contains five essential columns that provide valuable insights into the Twitter conversation dynamics: 1. Keyword: This column represents the specific keyword or topic of interest that generated the original tweet. It helps identify the context or subject matter around which the conversation revolves. 2. Main_tweet: The main_tweet column contains the original tweet related to the keyword. It serves as the starting point or focal point of the conversation and often provides essential information or opinions on the given topic. 3. Main_likes: This column provides the number of likes received by the main_tweet. Likes serve as a measure of engagement and indicate the level of popularity or resonance of the original tweet within the Twitter community. 4. Reply: The reply column consists of the replies or responses to the main_tweet. These replies may include comments, opinions, additional information, or discussions related to the keyword or the original tweet itself. The replies help capture the diverse perspectives and conversations that emerge in response to the main_tweet. 5. Reply_likes: This column records the number of likes received by each reply. Similar to the main_likes column, the reply_likes column measures the level of engagement and popularity of individual replies. It enables the identification of particularly noteworthy or well-received replies within the dataset. By analyzing this "Famous Keyword Twitter Replies Dataset," researchers, analysts, and data scientists can gain valuable insights into how popular keywords spark discussions on Twitter and how these discussions evolve through replies. The dataset's information on likes allows for the evaluation of tweet and reply popularity, helping to identify influential or impactful content. This dataset serves as a valuable resource for various applications, including sentiment analysis, trend identification, opinion mining, and understanding social media dynamics. > Number of tweets for each pairs of tweet and reply Total has 17255 pairs of tweet/reply

GameWikiSum

GameWikiSum is a domain-specific (video game) dataset for multi-document summarization, which is one hundred times larger than commonly used datasets, and in another domain than news. Input documents consist of long professional video game reviews as well as references of their gameplay sections in Wikipedia pages. Source: https://github.com/Diego999/GameWikiSum

HowSumm

HowSumm is a large-scale query-focused multi-document summarization dataset. It is focused on summarization of various sources to create HowTo guides. It is derived from wikiHow articles. HowSumm is partitioned into HowSumm-Step where target summary is relatively short (avg. 90 words) and HowSumm-Method where target summary is a concatenation of several steps an therefore longer (avg. 500 words). HowSumm-Method and HowSumm-Step contain 11,121 and 84348 instances, respectively. Description from: HowSumm Image source: HowSumm

ICLR Database

A maintained database tracks ICLR submissions and reviews, augmented with author profiles and higher-level textual features.

Inshorts News

Inshorts News dataset Inshorts provides a news summary in 60 words or less. Inshorts is a news service that offers short summaries of news from around the web. This dataset contains headlines and a summary of news items and their source.

M3LS

Significant developments in techniques such as encoder-decoder models have enabled us to represent information comprising multiple modalities. This information can further enhance many downstream tasks in the field of information retrieval and natural language processing; however, improvements in multi-modal techniques and their performance evaluation require large-scale multi-modal data which offers sufficient diversity. Multi-lingual modeling for a variety of tasks like multi-modal summarization, text generation, and translation leverages information derived from high-quality multi-lingual annotated data. In this work, we present the current largest multi-lingual multi-modal summarization dataset (M3LS), and it consists of over a million instances of document-image pairs along with a professionally annotated multi-modal summary for each pair. It is derived from news articles published by British Broadcasting Corporation(BBC) over a decade and spans 20 languages, targeting diversity across five language roots, it is also the largest summarization dataset for 13 languages and consists of cross-lingual summarization data for 2 languages. We formally define the multi-lingual multi-modal summarization task utilizing our dataset and report baseline scores from various state-of-the-art summarization techniques in a multi-lingual setting. We also compare it with many similar datasets to analyze the uniqueness and difficulty of M3LS.

MLSum-it

The MLSum-it dataset is the translated version (Helsinki-NLP/opus-mt-es-it) of the spanish portion of MLSum, containing news articles taken from BBC/mundo. There are two features: * source: Input news article. * target: Summary of the article.

MOS Dataset

This dataset was used in the paper 'Template-based Abstractive Microblog Opinion Summarisation' (to be published at TACL, 2022). The data is structured as follows: each file represents a cluster of tweets which contains the tweet IDs and a summary of the tweets written by journalists. The gold standard summary follows a template structure and depending on its opinion content, it contains a main story, majority opinion (if any) and/or minority opinions (if any).

MultiSum

MultiSum is a dataset for multimodal summarization (MSMO). It consists of 17 categories and 170 subcategories to encapsulate a diverse array of real-world scenarios. The dataset features: 1)Human-validated summaries for both video and textual content, providing superior human instruction and labels for multimodal learning. 2) Comprehensively and meticulously arranged categorization, spanning 17 principal categories and 170 subcategories to encapsulate a diverse array of real-world scenarios. 3) Benchmark tests performed on the proposed dataset to assess varied tasks and methods, including video temporal segmentation, video summarization, text summarization, and multimodal summarization. Source: MultiSum: A Dataset for Multimodal Summarization and Thumbnail Generation of Videos Image Source: MultiSum: A Dataset for Multimodal Summarization and Thumbnail Generation of Videos

OpenDebateEvidence

We introduce OpenDebateEvidence, a comprehensive dataset for argument mining and summarization sourced from the American Competitive Debate community. This dataset includes over 3.5 million documents with rich metadata, making it one of the most extensive collections of debate evidence. OpenDebateEvidence captures the complexity of arguments in high school and college debates, pro- viding valuable resources for training and evaluation. Our extensive experiments demonstrate the efficacy of fine-tuning state-of-the-art large language models for argumentative abstractive summarization across various methods, models, and datasets. By providing this comprehensive resource, we aim to advance com- putational argumentation and support practical applications for debaters, edu- cators, and researchers. OpenDebateEvidence is publicly available to support further research and innovation in computational argumentation. Access it here: https://huggingface.co/datasets/Yusuf5/OpenCaselist

PlainFact

PlainFact is a high-quality human-annotated dataset with fine-grained explanation (i.e., added information) annotations.

prompt-opin-summ

Click to add a brief description of the dataset (Markdown and LaTeX enabled). Provide: * a high-level explanation of the dataset characteristics * explain motivations and summary of its content * potential use cases of the dataset

Proto Summ

This is a large-scale court judgment dataset, where each judgment is a summary of the case description with a patternized style. It contains 2,003,390 court judgment documents. The case description is used as the input, and the court judgment as the summary. The average lengths of the input documents and summaries are 595.15 words and 273.57 words respectively. Source: https://arxiv.org/pdf/1909.08837.pdf

public_meetings

The public_meetings corpus contains meetings, made of pairs of automatic transcriptions from audio recordings and meeting reports written by a professional. 22 aligned meetings are provided in total. Source: public_meetings

QBSUM

A high-quality large-scale dataset consisting of 49,000+ data samples for the task of Chinese query-based document summarization. Source: QBSUM: a Large-Scale Query-Based Document Summarization Dataset from Real-world Applications

ShopTC-100K

ShopTC-100K Dataset The ShopTC-100K dataset is collected using TermMiner, an open-source data collection and topic modeling pipeline introduced in the paper: Harmful Terms and Where to Find Them: Measuring and Modeling Unfavorable Financial Terms and Conditions in Shopping Websites at Scale If you find this dataset or the related paper useful for your research, please cite our paper: @inproceedings{tsai2025harmful, author = {Elisa Tsai and Neal Mangaokar and Boyuan Zheng and Haizhong Zheng and Atul Prakash}, title = {Harmful Terms and Where to Find Them: Measuring and Modeling Unfavorable Financial Terms and Conditions in Shopping Websites at Scale}, booktitle = {Proceedings of the ACM Web Conference 2025 (WWW ’25)}, year = {2025}, location = {Sydney, NSW, Australia}, publisher = {ACM}, address = {New York, NY, USA}, pages = {14}, month = {April 28-May 2}, doi = {10.1145/3696410.3714573} } ## Dataset Description The dataset consists of sanitized terms extracted from e-commerce websites with English terms and conditions. The websites were sourced from the Tranco list (as of April 2024).

SubSumE

SubSumE Dataset This repository contains the SubSumE dataset for subjective document summarization. See the paper and the talk for details on dataset creation. Also check out our work SuDocu on example-based document summarization. ## Dataset Files Download the dataset from here. The dataset contains : * Simplified text from 48 Wikipedia pages of the states in the US. Additionally, all the sentences in these documents are put together in a single file processed_state_sentences.csv and are assigned a unique sentence id that is used in summary json files. * Intent-based summaries created by human annotators. Each datapoint file in the directory user_summary_jsons contains a json containing summaries of Wikipedia pages of eight states with following keys: * intent : Summarization intent provided to human annotators for generating the summary * summaries: List of summary jsons for eight states assigned to the annotator. Each json in the list contains following keys: * state_name: Name of the state * sentence_ids: Global ids of sentences (wrt processed_state_sentences.csv) present in the summary * sentences: List of sentences present in the summary * use_keywords: Keywords used by the annotator to search the document when creating summaries

TQBA++

Ultra-lightweight, multilingual QA eval dataset for rapid testing LLMs. Dataset Characteristics: - Multilingual: Includes packs for Arabic, Chinese, French, German, Japanese, Korean, Portuguese, Russian, Spanish, and Turkish. - Compact: Contains a curated English gold-standard set of 52 QA pairs (under 20kB), enabling immediate and resource-friendly evaluations. - Synthetic Generation: Features a LiteLLM-powered synthetic data generator see tinyqabenchmarkpp, allowing quick creation of custom evaluation sets tailored to specific domains or languages. - Metadata Support: Provided in Croissant-compatible formats, ready for seamless integration with modern evaluation harnesses and CI tools. Motivation and Content Summary: The primary motivation behind TQB++ is to enable rapid iteration and continuous integration (CI) of language models. Existing evaluation benchmarks typically involve significant computational overhead and slow feedback loops. In contrast, TQB++ offers near-instantaneous assessments of model performance and prompt stability across multiple languages. It is particularly sensitive to issues such as prompt-template regressions, tokenizer drift, and fine-tuning side effects. Potential Use Cases: - Continuous Integration (CI): Immediate detection of breaking changes or regressions in LLM pipelines. - Multilingual Model Validation: Quickly assess model accuracy and performance across multiple languages without large compute costs. - Prompt Optimization and Testing: Ideal for iterative prompt refinement workflows, enabling fast feedback loops and effective tuning. - Teaching and Prototyping: Educational use in courses or workshops, showcasing multilingual LLM evaluation in real-time scenarios.

UKIL-DB-EN

Bangladesh's legal system struggles with major challenges like delays, complexity, high costs, and millions of unresolved cases, which deter many from pursuing legal action due to lack of knowledge or financial constraints. This research seeks to develop a specialized Large Language Model (LLM) to assist in the Bangladeshi legal system. We created UKIL-DB-EN, an English corpus of Bangladeshi legal documents, by collecting and scraping data on various legal acts. We fine-tuned the GPT-2 model on this dataset to develop GPT2-UKIL-EN, an LLM focused on providing legal assistance in English. The model was rigorously evaluated using semantic assessments, including case studies supported by expert opinions. The evaluation provided promising results, demonstrating the potential for the model to assist in legal matters within Bangladesh. Our work represents the first structured effort toward building an AI-based legal assistant for Bangladesh. While the results are encouraging, further refinements are necessary to improve the model's accuracy, credibility, and safety.

VietMed-Sum

In doctor-patient conversations, identifying medically relevant information is crucial, posing the need for conversation summarization. In this work, we propose the first deployable real-time speech summarization system for real-world applications in industry, which generates a local summary after every N speech utterances within a conversation and a global summary after the end of a conversation. Our system could enhance user experience from a business standpoint, while also reducing computational costs from a technical perspective. Secondly, we present VietMed-Sum which, to our knowledge, is the first speech summarization dataset for medical conversations. Thirdly, we are the first to utilize LLM and human annotators collaboratively to create gold standard and synthetic summaries for medical conversation summarization.

Webis-TLDR-17 Corpus

This corpus contains preprocessed posts from the Reddit dataset, suitable for abstractive summarization using deep learning. The format is a json file where each line is a JSON object representing a post. The schema of each post is shown below: - author: string (nullable = true) - body: string (nullable = true) - normalizedBody: string (nullable = true) - content: string (nullable = true) - content_len: long (nullable = true) - summary: string (nullable = true) - summary_len: long (nullable = true) - id: string (nullable = true) - subreddit: string (nullable = true) - subreddit_id: string (nullable = true) - title: string (nullable = true) Specifically, the content and summary fields can be directly used as inputs to a deep learning model (e.g. Sequence to Sequence model ). The dataset consists of 3,848,330 posts with an average length of 270 words for content, and 28 words for the summary. The dataset is a combination of both the Submissions and Comments merged on the common schema. As a result, most of the comments which do not belong to any submission have null as their title. Note : This corpus does not contain a separate test set. Thus it is up to the users to divide the corpus into appropriate training, validation and test sets.

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