Explosion builds **developer tools** for AI, Machine Learning and Natural Language Processing. ### Project - [spaCy](/content/_/project/spacy/index.html) - [Prodigy](/content/_/project/prodigy/index.html) - [Ellf](/content/_/project/ellf/index.html) - [Thinc](/content/_/project/thinc/index.html) - [Consulting](/content/_/project/consulting/index.html) - [Case Study](/content/_/project/case_study/index.html) ### Topics - [LLMs](/content/_/topic/llms/index.html) - [NLP Strategy](/content/_/topic/strategy/index.html) - [Annotation](/content/_/topic/annotation/index.html) - [Biomedical](/content/_/topic/biomedical/index.html) - [Finance](/content/_/topic/finance/index.html) - [Media](/content/_/topic/media/index.html) - [Legal](/content/_/topic/legal/index.html) - [Humanities](/content/_/topic/humanities/index.html) - [Computer Vision](/content/_/topic/computer-vision/index.html) ### Category - [Blog](/content/_/category/blog/index.html) - [Release](/content/_/category/release/index.html) - [Universe](/content/_/category/universe/index.html) - [Talk](/content/_/category/talk/index.html) - [Interview](/content/_/category/interview/index.html) - [Video](/content/_/category/video/index.html) - [Paper](/content/_/category/paper/index.html) - [Book](/content/_/category/book/index.html) ### Tasks Select...Code GenerationCoreference ResolutionDependency ParsingDistillationEmbeddings & VectorsEntity LinkingEvaluationImage ClassificationImage SegmentationLayout AnalysisLemmatizationNamed Entity RecognitionObject DetectionOptical Character Recognition (OCR)Part-of-Speech TaggingPII AnonymizationQuestion AnsweringRelation ExtractionRetrieval-Augmented Generation (RAG)Rule-Based MatchingSpan CategorizationText ClassificationText GenerationTokenizationWeak Supervision ### Authors Select...Adriane BoydÁkos KádárBasile DuraChung-Fan TsaiDamian RomeroDaniël de KokDuygu AltinokEdward SchmuhlHelena SteckmeisterIndia KerleInes MontaniKabir KhanLj MirandaMadeesh KannanMagdalena AniołMatthew HonnibalPaul O’Leary McCannPeter BaumgartnerPhilip VolletRaphael MitschRehan AhmedRichard HudsonRyan WesslenSofie Van LandeghemVictoria SlocumVincent D. WarmerdamVinit RavishankarWalter Henry ## [The Nesta Skills Extractor Library](https://www.escoe.ac.uk/the-skills-extractor-library/) [Economic Statistics Centre of Excellence](https://www.escoe.ac.uk/the-skills-extractor-library/) [A new library for extracting skills from job adverts and mapping them to a taxonomy of your choice, built on top of spaCy.](https://www.escoe.ac.uk/the-skills-extractor-library/) ## [AI/ML for the rest of us](https://www.linkedin.com/pulse/ai-rest-us-github/) [GitHub Newsletter](https://www.linkedin.com/pulse/ai-rest-us-github/) ## [Training spaCy NER Models with Prodigy](https://github.com/explosion/assets/blob/main/Prodigy/Prodigy_NER_flowchart_v2_0_0_light.pdf) [This handy flowchart contains our most common tips, tricks, and best practices for training and updating spaCy named entity recognition models with Prodigy.](https://github.com/explosion/assets/blob/main/Prodigy/Prodigy_NER_flowchart_v2_0_0_light.pdf) ## [Multi hash embeddings in spaCy](https://arxiv.org/abs/2212.09255) [Miranda, Kádár, Boyd, Van Landeghem, Søgaard, Honnibal (2022)](https://arxiv.org/abs/2212.09255) [In this technical report we lay out a bit of history and introduce the embedding methods in spaCy in detail. Second, we critically evaluate the hash embedding architecture with multi-embeddings on Named Entity Recognition datasets from a variety of domains and languages. The experiments validate most key design choices behind spaCy’s embedders, but we also uncover a few surprising results.](https://arxiv.org/abs/2212.09255) ## [The triangulation of ethical leader signals using qualitative, experimental, and data science methods](https://www.sciencedirect.com/science/article/abs/pii/S1048984322000613) [Banks, Ross, Toth, Tonidandel, Goloujeh, Dou, Wesslen (2022)](https://www.sciencedirect.com/science/article/abs/pii/S1048984322000613) [This additional text was labeled by the same coding team using Prodigy, \,...\] ## [Finetuning and Bulk Labelling Images with Prodigy](https://www.youtube.com/watch?v=DmH3JmX3w2I) [In this video, we’ll show how you might be able to improve the annotation experience by using bulk labelling for image classification.](https://www.youtube.com/watch?v=DmH3JmX3w2I) ## [End-to-end Neural Coreference Resolution in spaCy](/content/blog/coref/index.html) [Coreference resolution is the problem of resolving entities in texts to references such as pronouns. Even if you've never heard of it, it's something we all do constantly every day, and is a key to understanding natural language. We recently added an experimental implementation of an end-to-end neural coreference component to spaCy. This post explains the architecture of our model in detail.](/content/blog/coref/index.html) ## [Introducing spaCy v3.4](/content/blog/spacy-v3-4/index.html) [spaCy v3.4 brings typing and speed improvements along with new vectors for English CNN pipelines and new trained pipelines for Croatian.](/content/blog/spacy-v3-4/index.html) ## Other Articles - [Finding Bad Labels for Text Classification with Jupyter and Prodigy](https://www.youtube.com/watch?v=khZ5-AN-n2Y) - [Finding Bad Image Data using UMAP and Prodigy](https://www.youtube.com/watch?v=s0Y45xscE-0) - [Applied Language Technology](https://applied-language-technology.mooc.fi/html/index.html) - [Healthsea: an end-to-end spaCy pipeline for exploring health supplement effects](/content/blog/healthsea/index.html) - [Fiscal data in text: Information extraction from audit reports using Natural Language Processing](https://www.cambridge.org/core/journals/data-and-policy/article/fiscal-data-in-text-information-extraction-from-audit-reports-using-natural-language-processing/F4CAA159BD8C5C71873D85FCF1E4AA96) - [Introducing spaCy v3.5](/content/blog/spacy-v3-5/index.html) - [Data is the new coffee](https://www.youtube.com/watch?v=GrQcVU-eapc) - [Extracting Structured Information from Greek Legislation Data](https://repository.ihu.edu.gr/xmlui/handle/11544/30135) - [Fast transformer inference with Metal Performance Shaders](/content/blog/metal-performance-shaders/index.html) ## [spaCy Cheat Sheet](https://github.com/explosion/assets/blob/main/spaCy/spaCy-cheat-sheet.pdf) [Everything you need to know about spaCy as a handy two-page PDF.](https://github.com/explosion/assets/blob/main/spaCy/spaCy-cheat-sheet.pdf) ## [floret: lightweight, robust word vectors](/content/blog/floret-vectors/index.html) [An exploration of floret vectors: lightweight vectors for noisy data, novel words, rich morphology and more.](/content/blog/floret-vectors/index.html) ## [Introducing Span Categorization in Prodigy and spaCy](https://www.youtube.com/watch?v=xgV3Rlj49lQ) [In this video, we’ll show you how to use Prodigy for spaCy’s Span Categorizer. We’ll be annotating food recipes and looking into ways to help with consistent annotations and speed up the process with patterns and temporary models.](https://www.youtube.com/watch?v=xgV3Rlj49lQ) ## [Solutions for Advanced NLP for Diverse Languages](https://speakerdeck.com/inesmontani/advanced-nlp-for-diverse-languages) [New Languages for NLP Keynote](https://speakerdeck.com/inesmontani/advanced-nlp-for-diverse-languages) [This talk discusses spaCy’s philosophy for modern NLP, its extensible design and new recent features to enable the development of advanced natural language processing pipelines for typologically diverse languages.](https://speakerdeck.com/inesmontani/advanced-nlp-for-diverse-languages) ## [Automated Identification of Clinical Procedures in Free-Text Electronic Clinical Records with a Low-Code Named Entity Recognition Workflow](https://www.thieme-connect.com/products/ejournals/abstract/10.1055/s-0042-1749358) [Macri, Teoh, Bacchi, Sun, Selva, Casson, Chan (2022), Methods of Information in Medicine](https://www.thieme-connect.com/products/ejournals/abstract/10.1055/s-0042-1749358) ## [When Women Make Headlines](https://pudding.cool/2022/02/women-in-headlines/) [The Pudding](https://pudding.cool/2022/02/women-in-headlines/) [Using spaCy and other packages from the NLP ecosystem for analyzing more than 382,000 headlines to see how women are represented (or misrepresented) in the news.](https://pudding.cool/2022/02/women-in-headlines/) ## [How the Guardian uses AI to analyse articles](https://www.journalismaifestival.com/) [JournalismAI Festival](https://www.journalismaifestival.com/) ## [Tools to Improve Training Data](https://www.youtube.com/watch?v=KRQJDLyc1uM) [Talking Language AI - Cohere](https://www.youtube.com/watch?v=KRQJDLyc1uM) ## [How the Guardian approaches quote extraction with NLP](/content/blog/guardian/index.html) [A case study of the Guardian's spaCy-Prodigy workflow to modularize quote extraction for content creation. This study includes iterative annotation guidelines and custom interface functionality.](/content/blog/guardian/index.html) ### [More Updates and Articles](/content/blog/index.html)