| dc.contributor.author | Znotiņš, Artūrs |
| dc.contributor.author | Paikens, Pēteris |
| dc.contributor.author | Grūzītis, Normunds |
| dc.date.accessioned | 2021-05-21T15:51:22Z |
| dc.date.available | 2021-05-21T15:51:22Z |
| dc.date.issued | 2020 |
| dc.identifier.uri | http://hdl.handle.net/20.500.12574/40 |
| dc.description | An automatically derived AMR annotation layer of the FullStack multi-layer text corpus of Latvian. First, Latvian UD Treebank (v2.5) sentences were translated to English using a state-of-the-art Latvian-English neural MT system (Hugo.lv). Second, a state-of-the-art AMR parser for English (AMREager) was applied to the MT-translated sentences. Additionally, alignment information is provided for both Latvian-English translations and English-AMR parses. |
| dc.language.iso | eng |
| dc.language.iso | lav |
| dc.publisher | AiLab IMCS UL |
| dc.relation.isreferencedby | http://www.lrec-conf.org/proceedings/lrec2018/pdf/935.pdf |
| dc.rights | Creative Commons - Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) |
| dc.rights.uri | http://creativecommons.org/licenses/by-sa/4.0/ |
| dc.rights.label | PUB |
| dc.source.uri | https://github.com/LUMII-AILab/FullStack |
| dc.subject | AMR |
| dc.subject | Abstract Meaning Representation |
| dc.title | Latvian AMR Sembank |
| dc.type | corpus |
| metashare.ResourceInfo#ContentInfo.detailedType | other |
| metashare.ResourceInfo#ContentInfo.mediaType | text |
| has.files | yes |
| branding | CLARIN Centre of Latvian language resources and tools |
| demo.uri | https://github.com/LUMII-AILab/FullStack/tree/master/AMR |
| contact.person | Normunds Grūzītis normundsg@ailab.lv IMCS UL |
| sponsor | European Regional Development Fund 1.1.1.1/16/A/219 Full Stack of Language Resources for Natural Language Understanding and Generation in Latvian nationalFunds |
| size.info | 12,691 sentences |
| files.size | 12370336 |
| files.count | 4 |
Files in this item
Download all files in item (11.8 MB)This item is
Creative Commons - Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)
Publicly Available
and licensed under:Creative Commons - Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)
- Name
- lv-amr-train.txt
- Size
- 9.14 MB
- Format
- Text file
- Description
- Training data
- MD5
- 1cfff686c00b5cae11340398518048e7
# ::id a-n136-p21s1
# ::snt -LRB- 10 -RRB- Directive 80/987/EEC should be amended accordingly .
# ::snt_lv (10) Attiecīgi ir jāgroza Direktīva 80/987/EEK.
# ::alignments 0-1|0.4 1-2|0.3 2-3|0.2 3-4|0.1.0 4-5|0.1 5-6|0 7-8|0.0 8-9|0.0.0
# ::alignments_translation 0-0 1-1 2-2 6-3 7-4 4-5 5-6 5-7 3-8 8-9
(v5 / recommend-01
:ARG1 (v6 / amend-01
:manner (v7 / accordingly))
:time (v4 / genericconcept
:mod (v3 / directive))
:ARG1 (v2 / -rrb-)
:li 10
:condition (v1 / -lrb-))
# ::id a-p5878-p1s1
# ::snt The family of David and Victoria Beckham live in Los Angeles , but it looks like they could move to England soon .
# ::snt_lv Deivida un Viktorijas Bekhemu ģimene mitinās Losandželosā, bet izskatās, ka drīzumā viņi varētu pārvākties uz Angliju.
# ::alignments 1-2|0.3 3-4|0.2.0+0.2.0.0+0.2.0.0.0+0.2.0.1 4-5|0.2 5-7|0.0.0+0.0.0.0+0.0.0.0.0+0.0.0.0.1 7-8|0.0 9-11|0.0.1+0.0.1.0+0.0.1.0.0+0.0.1.0.1+0.0.1.1 12-13|0 14-15|0.1 15-16|0.1.0 17-18|0.1.0.0 18-19|0.1.0.0 . . .
- Name
- lv-amr-dev.txt
- Size
- 1.43 MB
- Format
- Text file
- Description
- Development data
- MD5
- b0676cffe4f613578c91a3b33f8d70bc
# ::id a-p3754-p5s1
# ::snt The company started the project in 2001 .
# ::snt_lv Uzņēmums projekta realizāciju sāka 2001.gadā.
# ::alignments 1-2|0.0 2-3|0 4-5|0.1 6-7|0.1.0+0.1.0.0
# ::alignments_translation 0-0 0-1 3-2 2-3 1-4 5-5 4-6 6-6 7-7
(v2 / start-01
:ARG0 (v1 / company)
:ARG1 (v3 / project
:time (v4 / date-entity
:year 2001)))
# ::id a-p3754-p5s2
# ::snt Since then , more than seven million lats have been invested in landfill development .
# ::snt_lv Kopš tā laika poligona attīstībā ieguldīti vairāk nekā septiņi miljoni latu.
# ::alignments 0-1|0.1 3-4|0.0.0 7-8|0.0 10-11|0 12-13|0.2.0 13-14|0.2
# ::alignments_translation 0-0 1-1 2-1 6-2 6-3 7-4 8-5 9-6 10-7 5-8 5-9 5-10 4-11 3-12 4-13 11-14
(v4 / invest-01
:ARG1 (v3 / lat
:quant (v2 / more))
:op1 (v1 / since)
:ARG2 (v6 / develop-02
:ARG1 (v5 / landfill)
:ARG0 v3))
# ::id a-d62-p38s1
# ::snt `` We wo n't be let in there , we do n't have such a fine key ! ''
# . . .
- Name
- lv-amr-test.txt
- Size
- 1.22 MB
- Format
- Text file
- Description
- Test data
- MD5
- 118bfbf48eb1d1536091a94c0cb790e0
# ::id a-p3753-p48s1
# ::snt FOR INFORMATION
# ::snt_lv INFORMĀCIJAI
# ::alignments 1-2|0
# ::alignments_translation 0-0 0-1
(v1 / information)
# ::id a-p15196-p5s1
# ::snt However , he summarized that unpopular decisions in the education sector would have the `` back '' of the Prime Minister if they were rational .
# ::snt_lv Tomēr viņš rezumēja, ka nepopulāriem lēmumiem izglītības nozarē būs premjera "aizmugure", ja tie būs racionāli.
# ::alignments 3-4|0 5-6|0.0.0+0.0.0.0+0.0.0.0.0 6-7|0.0 9-10|0.0.1.0 10-11|0.0.1 14-15|0.0.2 15-16|0.0.2.0 16-17|0.0.3 19-20|0.0.3.0.2 20-21|0.0.3.0+0.0.3.0.0 24-25|0.0.3.0.1
# ::alignments_translation 0-0 1-1 1-2 2-3 3-4 4-4 5-5 6-6 8-7 7-8 7-9 8-10 9-11 9-12 10-13 11-14 12-15 13-16 10-17 10-18 10-19 10-20 14-21 15-21 16-22 17-23 18-24 19-25
(v1 / summarize-01
:ARG1 (v4 / decide-01
:ARG1 (v2 / thing
:degree-of (v3 / popular-02
:polarity -))
:location (v6 / sector
:mod (v5 / educate-01)) . . .
- Name
- README.txt
- Size
- 1.64 KB
- Format
- Text file
- Description
- Data format description of automatically derived AMR annotation layer of the FullStack-LV multi-layer text corpus. See also: https://github.com/LUMII-AILab/FullStack/tree/master/AMR.
- MD5
- ab87e5b03a6721f7e6e82beffb577290
DATA FRORMAT
METADATA
id - the sentence ID according to the Latvian UD Treebank data.
snt_lv - the original sentence (Latvian).
snt - the machine-translated sentence (English).
alignments_translation - token-level alignment between the Latvian and English sentences.
alignments - alignment between tokens (English) and nodes of the AMR graph (see the alignment format: https://github.com/lil-lab/amr/blob/07b581e7d160fa2625eeefa86ae5e9fe5c589be2/utils/jamr/docs/Alignment_Format.md).
AMR GRAPH
The English translation encoded as an AMR graph in the PENMAN notation (see the language specification: https://amr.isi.edu/language.html).
Each record in the dataset is separated by an empty line. The dataset is split into subsets for training, development and testing using the same split as for the Latvian UD Treebank (v2.5).
EXAMPLE
# ::id a-p3315-p3s6
# ::snt Sister Ilze is also a model .
# ::snt_lv Māsa Ilze arī ir modelētāja.
# ::alignments 0-1|0.2+0.2.0+0.2.0.0 1-2|0 . . .
