--- title: "Nymble: a High-Performance Learning Name-finder — Bikel 1997" description: "Bikel and colleagues present Nymble, a trainable system for the MUC-6 named-entity (NE) task. 8 references, cited by 131 papers here." authors: ["Daniel M. Bikel", "Scott Miller", "Richard Schwartz", "Ralph Weischedel"] year: 1997 venue: "Fifth Conference on Applied Natural Language Processing" source: "https://aclanthology.org/A97-1029/" doi: "10.3115/974557.974586" licence: "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)" licence_url: "https://creativecommons.org/licenses/by-nc-sa/3.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/bikel-1997-nymble-high-performance-learning-name/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/bikel-1997-nymble-high-performance-learning-name/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [1997](https://md.research.elamur.ai/free/en/papers/1997/) / Nymble: a High-Performance Learning Name-finder # Nymble: a High-Performance Learning Name-finder **Authors:** [Daniel M. Bikel](https://md.research.elamur.ai/free/en/authors/daniel-m-bikel/), [Scott Miller](https://md.research.elamur.ai/free/en/authors/scott-miller/), [Richard Schwartz](https://md.research.elamur.ai/free/en/authors/richard-schwartz/) and [Ralph Weischedel](https://md.research.elamur.ai/free/en/authors/ralph-weischedel/) · **Year:** 1997 · **Venue:** Fifth Conference on Applied Natural Language Processing · **Licence:** [CC BY-NC-SA 3.0 (ACL Anthology, before 2016)](https://creativecommons.org/licenses/by-nc-sa/3.0/) **Links:** [Publisher's page](https://aclanthology.org/A97-1029/) · [PDF](https://aclanthology.org/A97-1029.pdf) · [DOI 10.3115/974557.974586](https://doi.org/10.3115/974557.974586) · [OpenAlex](https://openalex.org/W2163918411) **Citations:** 726 citations (OpenAlex, 2026-10-14) · [131](https://md.research.elamur.ai/free/en/papers/bikel-1997-nymble-high-performance-learning-name/cited-by/) in our library, of the 140,884 papers whose bibliographies we read ## Summary Bikel and colleagues present Nymble, a trainable system for the MUC-6 named-entity (NE) task. The task covers person, organization and location names, plus dates, times, percentages and money amounts. Earlier systems mostly used hand-written finite-state patterns, which have to be rewritten for each new language or entity type. Nymble instead learns from annotated text with a modified hidden Markov model. Each name class acts as its own small bigram language model, and the Viterbi algorithm picks the most probable class sequence for a sentence. Words carry a simple deterministic feature, such as number shape or capitalization type. Sparse statistics are handled by a separately trained unknown-word model and a chain of back-off estimates. The system was tested on English MUC-6 data from WSJ and on Spanish MET data from AFP. It reached F-measures of about 90 or higher, which the authors describe as the best published result for a learned name-finder and close to hand-built systems. It needed only about 100,000 words of training data. This shows that statistical NE recognition can be ported across languages cheaply. ## Key points - **Model structure:** an ergodic HMM with eight name-class states (including NOT-A-NAME) and special START-OF-SENTENCE and END-OF-SENTENCE states. Each class has its own internal bigram model. - **Word features:** 14 deterministic features (e.g., twoDigitNum, containsDigitAndDash, allCaps, capPeriod, firstWord, initCap, lowercase). Feature computation is the only language-dependent part, at roughly ten lines of code. - **Test data:** English was evaluated on the MUC-6 test set of 30 WSJ documents and Spanish on the MET test set of AFP articles. Only blind-test scores are reported. - **Scores:** Table 4.1 is garbled in the extracted text. It appears to show Nymble at about 93 F on mixed-case English, 91 on upper-case English and 90 on Spanish. The corresponding best reported scores appear to be 96, 89 and 93. - **Training-data size:** - English had 450,000 training words. Halving this barely hurt performance, and dropping to about 100,000 words cost only 1–2 points. At 50,000 words the drop was larger. - Spanish (223,000 words) showed almost no change when cut to about 100,000 words. ## Interesting ideas - **Name class as a mini language model.** Each entity type gets its own bigram model. The class-switching probability is conditioned on the previous class and the previous word. This captures cues like a title before a person name, and is a compact design for any sequence-labeling task. - **Ordered, deterministic word features.** Overlapping feature classes are resolved by a fixed precedence order. A dedicated firstWord flag records that capitalization at sentence start is uninformative. This is a cheap way to add shape information for low-resource languages, though the authors found Spanish lower-case organization and location words to be a pitfall. - **Swapped halves for unknown words.** The vocabulary is built on one half of the training data and unknown-word statistics are gathered on the other half, then the roles are reversed. All the data thus contributes to estimating unseen-word behavior, which helps when the corpus is small. - **Explicit end-of-class token.** A special +end+ token lets the model score where an entity stops. When predicting the next class, it falls back to the last real word of the previous class. This is a tidy boundary-modeling trick for segmenting tasks. - **Learning curves by repeated halving.** Training on successive halves of the data gives a quick, honest estimate of how much annotation a new language really needs. It helps decide whether to annotate more data or to build richer models. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* Fields: [Natural language processing](https://md.research.elamur.ai/free/en/shelves/nlp/) ## Cite Cite as: Daniel M. Bikel, Scott Miller, Richard Schwartz and Ralph Weischedel. 1997. Nymble: a High-Performance Learning Name-finder. Fifth Conference on Applied Natural Language Processing. The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{bikel1997nymble, title = {{Nymble: a High-Performance Learning Name-finder}}, author = {Daniel M. Bikel, Scott Miller, Richard Schwartz and Ralph Weischedel}, booktitle = {Fifth Conference on Applied Natural Language Processing}, year = {1997}, doi = {10.3115/974557.974586}, url = {https://doi.org/10.3115/974557.974586} } ``` - [← previous Using Lexical Chains for Text Summarization](https://md.research.elamur.ai/free/en/papers/barzilay-1997-lexical-chains-text-summarization/) - [next → PARADISE: A Framework for Evaluating Spoken Dialogue Agents](https://md.research.elamur.ai/free/en/papers/walker-1997-paradise-framework-evaluating-spoken-dialogue/) - [↑ back to papers of 1997](https://md.research.elamur.ai/free/en/papers/1997/) ## References 8 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Aberdeen, J., Burger, J., Day, D., Hirschman, L., Robinson, P. and Vilain, M. (1995) In Proceedings of the Sixth Message Understanding Conference (MUC-6)Morgan Kaufmann Publishers, Inc., Columbia, Maryland, pp. 141-155. [find](https://scholar.google.com/scholar?q=Aberdeen%2C+J%2E%2C+Burger%2C+J%2E%2C+Day%2C+D%2E%2C+Hirschman%2C+L%2E%2C+Robinson%2C+P%2E+and+Vilain%2C+M%2E+%281995%29+In+Proceedings+of+the+Sixth+Message+Understanding+Conference+%28MUC%2D6%29Morgan+Kaufmann+Publishers%2C+Inc%2E%2C+Columbia%2C+Mar) 2. Appelt, D. E., Jerry R. Hobbs, Bear, J., Israel, D., Kameyama, M., Kehler, A., Martin, D., Myers, K. and Tyson, M. (1995) In Proceedings of the Sixth Message Understanding Conference (MUC-6)Morgan Kaufmann Publishers, Inc., Columbia, Maryland, pp. 237-248. [find](https://scholar.google.com/scholar?q=Appelt%2C+D%2E+E%2E%2C+Jerry+R%2E+Hobbs%2C+Bear%2C+J%2E%2C+Israel%2C+D%2E%2C+Kameyama%2C+M%2E%2C+Kehler%2C+A%2E%2C+Martin%2C+D%2E%2C+Myers%2C+K%2E+and+Tyson%2C+M%2E+%281995%29+In+Proceedings+of+the+Sixth+Message+Understanding+Conference+%28MUC%2D6%29Morgan+Kau) 3. Church, K. (1988) In Second Conference on Applied Natural Language Processing, Austin, Texas. [find](https://scholar.google.com/scholar?q=Church%2C+K%2E+%281988%29+In+Second+Conference+on+Applied+Natural+Language+Processing%2C+Austin%2C+Texas%2E) 4. Cover, T. and Thomas, J. A. (1991) Elements of Information Theory, John Wiley & Sons, Inc., New York. [find](https://scholar.google.com/scholar?q=Cover%2C+T%2E+and+Thomas%2C+J%2E+A%2E+%281991%29+Elements+of+Information+Theory%2C+John+Wiley+%26+Sons%2C+Inc%2E%2C+New+York%2E) 5. Miller, S., Bobrow, R., Schwartz, R. and Ingria, R. (1994) In Human Language Technology Workshop, Morgan Kaufmann Publishers, Plainsboro, New Jersey, pp. 278-282. [find](https://scholar.google.com/scholar?q=Miller%2C+S%2E%2C+Bobrow%2C+R%2E%2C+Schwartz%2C+R%2E+and+Ingria%2C+R%2E+%281994%29+In+Human+Language+Technology+Workshop%2C+Morgan+Kaufmann+Publishers%2C+Plainsboro%2C+New+Jersey%2C+pp%2E+278%2D282%2E) 6. Viterbi, A. J. (1967) IEEE Transactions on Information Theory, IT-13(2), 260-269. [find](https://scholar.google.com/scholar?q=Viterbi%2C+A%2E+J%2E+%281967%29+IEEE+Transactions+on+Information+Theory%2C+IT%2D13%282%29%2C+260%2D269%2E) 7. Weischedel, R. (1995) In Proceedings of the Sixth Message Understanding Conference (MUC-6)Morgan Kaufmann Publishers, Inc., Columbia, Maryland, pp. 55-69. [find](https://scholar.google.com/scholar?q=Weischedel%2C+R%2E+%281995%29+In+Proceedings+of+the+Sixth+Message+Understanding+Conference+%28MUC%2D6%29Morgan+Kaufmann+Publishers%2C+Inc%2E%2C+Columbia%2C+Maryland%2C+pp%2E+55%2D69%2E) 8. Weischedel, R., Meteer, M., Schwartz, R., Ramshaw, L. and Palmucci, J. (1993) Computational Linguistics, 19(2), 359-382. 8. Acknowledqements The work reported here was supported in part by the Defense Advanced Research Projects Agency; a technical agent for part of the work was Fort Huachucha under… [find](https://scholar.google.com/scholar?q=Weischedel%2C+R%2E%2C+Meteer%2C+M%2E%2C+Schwartz%2C+R%2E%2C+Ramshaw%2C+L%2E+and+Palmucci%2C+J%2E+%281993%29+Computational+Linguistics%2C+19%282%29%2C+359%2D382%2E+8%2E+Acknowledqements+The+work+reported+here+was+supported+in+part+by+the+Defense)