External Lexical Information for Multilingual Part-of-Speech Tagging

Benoît Sagot 1
1 ALPAGE - Analyse Linguistique Profonde à Grande Echelle ; Large-scale deep linguistic processing
Inria Paris-Rocquencourt, UPD7 - Université Paris Diderot - Paris 7
Abstract : Morphosyntactic lexicons and word vector representations have both proven useful for improving the accuracy of statistical part-of-speech taggers. Here we compare the performances of four systems on datasets covering 16 languages, two of these systems being feature-based (MEMMs and CRFs) and two of them being neural-based (bi-LSTMs). We show that, on average, all four approaches perform similarly and reach state-of-the-art results. Yet better performances are obtained with our feature-based models on lexically richer datasets (e.g. for morphologically rich languages), whereas neural-based results are higher on datasets with less lexical variability (e.g. for English). These conclusions hold in particular for the MEMM models relying on our system MElt, which benefited from newly designed features. This shows that, under certain conditions, feature-based approaches enriched with morphosyntactic lexicons are competitive with respect to neural methods.
Document type :
Reports
Complete list of metadatas

https://hal.inria.fr/hal-01330301
Contributor : Benoît Sagot <>
Submitted on : Saturday, August 6, 2016 - 3:54:53 PM
Last modification on : Friday, January 4, 2019 - 5:33:24 PM

Files

RR-8924.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01330301, version 3
  • ARXIV : 1606.03676

Citation

Benoît Sagot. External Lexical Information for Multilingual Part-of-Speech Tagging. [Research Report] RR-8924, Inria Paris. 2016. ⟨hal-01330301v3⟩

Share

Metrics

Record views

236

Files downloads

242