{"url":"/dataset/nemo-corpus","name":"NEMO-Corpus","full_name":"NEMO Hebrew NER and Morphology Corpus","description_markdown":"Named Entity (NER) annotations of the Hebrew Treebank (Haaretz newspaper) corpus, including: morpheme and token level NER labels, nested mentions, and more.  \r\nWe publish the NEMO corpus in the TACL paper [*\"Neural Modeling for Named Entities and Morphology (NEMO^2)\"*](https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00404/107206/Neural-Modeling-for-Named-Entities-and-Morphology) [1], where we use it in extensive experiments and analyses, showing the importance of morphological boundaries for neural modeling of NER in morphologically rich languages. Code for these models and experiments can be found in the [NEMO code repo](https://github.com/OnlpLab/NEMO).\r\n\r\n## Main features:\r\n1. Morpheme, token-single and token-multi sequence labels. Morpheme labels provide exact boundaries, token-multi provide partial sub-word morphological but no exact boundaries, token-single provides only token-level information. \r\n1. All annotations are in `BIOSE` format (`B`=Begin, `I`=Inside, `O`=Outside, `S`=Singleton, `E`=End).\r\n1. Widely-used OntoNotes entity category set: `GPE` (geo-political entity), `PER` (person), `LOC` (location), `ORG` (organization), `FAC` (facility), `EVE` (event), `WOA` (work-of-art), `ANG` (language), `DUC` (product).\r\n1. NEMO includes NER annotations for the two major versions of the Hebrew Treebank, UD (Universal Dependency) and SPMRL. These can be aligned to the other morphosyntactic information layers of the treebank using [bclm](https://github.com/OnlpLab/bclm)\r\n1. We provide nested mentions. Only the first, widest, layer is used in the NEMO^2 paper. We invite you to take on this challenge!\r\n1. Guidelines used for annotation are provided [here](https://github.com/OnlpLab/NEMO-Corpus/tree/main/guidelines).\r\n1. Corpus was annotated by two native Hebrew speakers of academic education, and curated by the project manager. We provide the original annotations made by the annotators as well to promote work on [learning with disagreements](https://sites.google.com/view/semeval2021-task12/home).\r\n1. Annotation was performed using [WebAnno](https://webanno.github.io/webanno/) (version 3.4.5)\r\n\r\n\r\n## Basic Corpus Statistics\r\n\r\n|                              | train        | dev          | test          |\r\n|------------------------------|           --:|           --:|            --:|\r\n|  Sentences                   |  4,937       |  500         |  706          |\r\n|  Tokens                      |  93,504      |  8,531       |  12,619       |\r\n|  Morphemes                   |  127,031     |  11,301      |  16,828       |\r\n|  All mentions                |  6,282       |  499         |  932          |\r\n|  Type: Person         (PER)  |  2,128       |  193         |  267          |\r\n|  Type: Organization   (ORG)  |  2,043       |  119         |  408          |\r\n|  Type: Geo-Political  (GPE)  |  1,377       |  121         |  195          |\r\n|  Type: Location       (LOC)  |  331         |  28          |  41           |\r\n|  Type: Facility       (FAC)  |  163         |  12          |  11           |\r\n|  Type: Work-of-Art    (WOA)  |  114         |  9           |  6            |\r\n|  Type: Event          (EVE)  |  57          |  12          |  0            |\r\n|  Type: Product        (DUC)  |  36          |  2           |  3            |\r\n|  Type: Language       (ANG)  |  33          |  3           |  1            |\r\n\r\n\r\n## Evaluation\r\nAn evaluation script is provided in the [NEMO code repo](https://github.com/OnlpLab/NEMO#evaluation) along with evaluation instructions.\r\n\r\n## Citation\r\n```\r\n@article{10.1162/tacl_a_00404,\r\n    author = {Bareket, Dan and Tsarfaty, Reut},\r\n    title = \"{Neural Modeling for Named Entities and Morphology (NEMO2)}\",\r\n    journal = {Transactions of the Association for Computational Linguistics},\r\n    volume = {9},\r\n    pages = {909-928},\r\n    year = {2021},\r\n    month = {09},\r\n    abstract = \"{Named Entity Recognition (NER) is a fundamental NLP task, commonly formulated as classification over a sequence of tokens. Morphologically rich languages (MRLs) pose a challenge to this basic formulation, as the boundaries of named entities do not necessarily coincide with token boundaries, rather, they respect morphological boundaries. To address NER in MRLs we then need to answer two fundamental questions, namely, what are the basic units to be labeled, and how can these units be detected and classified in realistic settings (i.e., where no gold morphology is available). We empirically investigate these questions on a novel NER benchmark, with parallel token- level and morpheme-level NER annotations, which we develop for Modern Hebrew, a morphologically rich-and-ambiguous language. Our results show that explicitly modeling morphological boundaries leads to improved NER performance, and that a novel hybrid architecture, in which NER precedes and prunes morphological decomposition, greatly outperforms the standard pipeline, where morphological decomposition strictly precedes NER, setting a new performance bar for both Hebrew NER and Hebrew morphological decomposition tasks.}\",\r\n    issn = {2307-387X},\r\n    doi = {10.1162/tacl_a_00404},\r\n    url = {https://doi.org/10.1162/tacl\\_a\\_00404},\r\n    eprint = {https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl\\_a\\_00404/1962472/tacl\\_a\\_00404.pdf},\r\n}\r\n```","description_withheld":null,"homepage":"https://github.com/OnlpLab/NEMO-Corpus","introduced_date":"2020-07-30","introduced_date_note":null,"introduced_by":{"paper":"/paper/neural-modeling-for-named-entities-and","title":"Neural Modeling for Named Entities and Morphology (NEMO^2)","first_author":"Dan Bareket","url":null},"license":null,"modalities":[{"name":"Texts","url":"/datasets/modality/texts"}],"tasks":[{"name":"Named Entity Recognition (NER)","url":"/task/named-entity-recognition-ner","datasets_with_task":"/datasets/task/named-entity-recognition-ner"},{"name":"Nested Named Entity Recognition","url":"/task/nested-named-entity-recognition","datasets_with_task":"/datasets/task/nested-named-entity-recognition"},{"name":"Morphological Disambiguation","url":"/task/morphological-disambiguation","datasets_with_task":"/datasets/task/morphological-disambiguation"}],"languages":[{"name":"Hebrew","url":"/datasets/language/hebrew"}],"variants":["NEMO-Corpus","NEMO-Corpus (morph,dev)","NEMO-Corpus (morph,test)","NEMO-Corpus (token,dev)","NEMO-Corpus (token,test)"],"data_loaders":[],"num_papers_in_archive":3,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/named-entity-recognition-on-nemo-corpus-morph-1","task":"Named Entity Recognition (NER)","dataset_variant":"NEMO-Corpus (morph,test)","rows":3,"metrics":["F1"],"first_row_in_archive_order":{"model":"AlephBERT-base Pipeline","paper":"/paper/alephbert-a-hebrew-large-pre-trained-language","metrics":{"F1":"80.15"},"code_links":[{"title":"8400thehealthnetwork/hebsafeharbor","url":"https://github.com/8400thehealthnetwork/hebsafeharbor"},{"title":"OnlpLab/Hebrew-Sentiment-Data","url":"https://github.com/OnlpLab/Hebrew-Sentiment-Data"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/named-entity-recognition-on-nemo-corpus-token-1","task":"Named Entity Recognition (NER)","dataset_variant":"NEMO-Corpus (token,test)","rows":2,"metrics":["F1"],"first_row_in_archive_order":{"model":"AlephBERT-base","paper":"/paper/alephbert-a-hebrew-large-pre-trained-language","metrics":{"F1":"84.91"},"code_links":[{"title":"8400thehealthnetwork/hebsafeharbor","url":"https://github.com/8400thehealthnetwork/hebsafeharbor"},{"title":"OnlpLab/Hebrew-Sentiment-Data","url":"https://github.com/OnlpLab/Hebrew-Sentiment-Data"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/named-entity-recognition-ner-on-nemo-corpus","task":"Named Entity Recognition (NER)","dataset_variant":"NEMO-Corpus","rows":1,"metrics":["F1"],"first_row_in_archive_order":{"model":"AlephBERTGimmel-base MTL","paper":"/paper/large-pre-trained-models-with-extra-large","metrics":{"F1":"80.39"},"code_links":[]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/large-pre-trained-models-with-extra-large","title":"Large Pre-Trained Models with Extra-Large Vocabularies: A Contrastive Analysis of Hebrew BERT Models and a New One to Outperform Them All","date":"2022-11-28","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/alephbert-a-hebrew-large-pre-trained-language","title":"AlephBERT:A Hebrew Large Pre-Trained Language Model to Start-off your Hebrew NLP Application With","date":"2021-04-08","rows_on_this_dataset":3,"code_links":2,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":2,"samples_ran":2,"samples_unverified":0,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/neural-modeling-for-named-entities-and","title":"Neural Modeling for Named Entities and Morphology (NEMO^2)","date":"2020-07-30","rows_on_this_dataset":2,"code_links":4,"syntology":null}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":1,"samples_harvested":2,"samples_ran":2,"samples_unverified":0,"pointer_only_for_licence":0,"papers_with_no_sample_that_ran":0,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}