Keeletehnoloogia laboratoorium
TalTech prioriteetne teadussuund
Klassifikaator (Frascati)
Keeled ja kirjandus 6.2
Uurimisrühma juht
Uurimisrühma liige
Rõbnikov, Daniil
Velve, Andra
Sildam, Tiia
Doktorant
Võtmesõna
kõnetehnoloogia
foneetika
kõnekorpused
Ülevaade
Keeletehnoloogia laboratoorium keskendub järgmisetele teemadele:• Kõnetuvastus• Kõneleja, kõneldava keele ja aktsendi identifitseerimine• Kõnekorpused• Foneetika (eesti keele prosoodia, L2 kõne)• Mitmesugused loomuliku keele töötluse alamteemadLabori üheks väljapaistvamaks tegevuseks on eesti keele kõnetuvastuse arendus ning avalikult kättesaadavate kõnetuvastusteenuste loomine. Kuigi labor keskendub arendustöös eesti keelele, on enamik laboris loodud meetodeid ja tehnoloogiaid keelest sõltumatud. Laboris välja töötatud tarkvara on saadaval vaba tarkvara litsentsi alusel.
Tähtsamad tulemused
2023. aasta tulemused:Uurimisrühma ühisprojekt EMTAga ooperilauljate hääle analüüsil on andnud huvitavaid tulemusi ning artikkel selles ilmus ühes mainekamas akustika-alases ajakirjas Journal of the Acosutic Society of America. Tegemist on maailma mastaabis uudse tööga, sest ooperilaulu arusaadavust pole foneetilise metodoloogiaga varem uuritud.Tanel Alumäe ja Daniil Rõbnikov osalesid konverentsi ASRU 2023 osana peetud võistlusel MADASR Challenge, kus mõõdeti kõnetuvastussüsteemide kvaliteeti kahe Indias kõneldava dialektirikka keele peal. Laboris välja töötatud mudelid andsid parimad tulemused üle kõikide osalenud tiimide. Lahendusest valmis ka artikkel, mis avaldati sama konverentsi kogumikus.
Alumäe, T.; Kalda, J.; Bode, K.; Kaitsa, M. (2023). Automatic closed captioning for Estonian live broadcasts. Proceedings of the 24th Nordic Conference on Computational Linguistics (NoDaLiDa), May 22 - 24, 2023, Tórshavn, Faroe Islands. University of Tartu Library, 492−499. (NEALT Proceedings Series; 52).
Vurma, A.; Meister, E.; Meister, L.; Ross, J.; Raju, M.; Kala, V.; Dede, T. (2023). The intensities of vowels and plosive bursts and their impact on text intelligibility in singing. The Journal of the Acoustical Society of America, 154 (4), 2653−2664. DOI: 10.1121/10.0021968
Alumäe, T.; Kukk, K.; Le, V.iet-B.; Barras, C.; Messaoudi, A.; Ben Khender, W. (2023). Exploring the impact of pretrained models and web-scraped data for the 2022 NIST Language Recognition Evaluation. INTERSPEECH 2023, 20-24 August 2023, Dublin, Ireland. ISCA, 516−520. DOI: 10.21437/Interspeech.2023-1790
Seotud projektid
Uurimisrühma tegevusperiood
2023
2022
2021
2020
2019
2018
2017
Seotud struktuuriüksus
Teadusgrupiga seotud publikatsioonid
- Kukk, K., Alumäe, T. Improving language identification of accented speech // Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, 2022, 18 September-22 September. : International Speech Communication Association, 2022. p. 1288-1292 : ill.
https://doi.org/10.21437/Interspeech.2022-10455 - Malmi, A., Lippus, P., Meister, E. Spectral and temporal properties of Estonian palatalization // Journal of the International Phonetic Association (2022), p. 1-26 : ill.
https://doi.org/10.1017/S0025100321000360 - Valk, J., Alumäe, T. VOXLINGUA107 : A dataset for spoken language recognition // IEEE Spoken Language Technology Workshop. : IEEE, 2021. p. 652-658.
https://doi.org/10.1109/SLT48900.2021.9383459 - Paats, A., Alumäe, T., Meister, E., Fridolin, I. Retrospective analysis of clinical performance of an Estonian speech recognition system for radiology : effects of different acoustic and language models // Journal of digital imaging (2018) vol. 31, 5, p. 615–621 : ill.
https://doi.org/10.1007/s10278-018-0085-8 - Alumäe, T. Training speaker recognition models with recording-level labels // 2018 IEEE Workshop on Spoken Language Technology : SLT 2018 : Proceedings, December 18–21, 2018, Athens, Greece. Danvers : IEEE, 2018. p. 1066-1072.
http://doi.org/10.1109/SLT.2018.8639601 - Łańcucki, A., Chorowski, J., Sanchez, G., Marxer, R., Alumäe, T. et al. Robust training of vector quantized bottleneck models // 2020 International Joint Conference onNeural Networks (IJCNN), 19-24 July 2020, Glasgow, UK : proceedings. Danvers : IEEE, 2020. art. 163566 : 7 p.
https://doi.org/10.1109/IJCNN48605.2020.9207145 - Tena, A., Claria, F., Solsona, F., Meister, E., Povedano, M. Detection of bulbar involvement in patients with amyotrophic lateral sclerosis by machine learning voice analysis : diagnostic decision support development study // JMIR Medical Informatics (2021) vol. 9, 3, e21331, 18 p. : ill.
https://doi.org/10.2196/21331 - Olev, A., Alumäe, T. Estonian speech recognition and transcription editing service // Baltic journal of modern computing (2022) vol. 10, 3, p. 409-421.
https://doi.org/10.22364/bjmc.2022.10.3.14 - Bond, F., Morgado da Costa, L., Goodman, M.W., McCrae, J.P., Lohk, A. Some issues with building a multilingual wordnet // LREC 2020 Marseille : Twelfth International Conference on Language Resources and Evaluation, May 11-16, 2020, Marseille, France : conference proceedings. Paris : European Language Resources Association, 2020. p. 3189-3197.
http://www.lrec-conf.org/proceedings/lrec2020/LREC-2020.pdf - Leier, M., Riid, A., Alumäe, T., Reinsalu, U., Pihlak, R., Udal, A., Heinsar, R., Vainküla, S. Smart elevator with unsupervised learning for visitor profiling and personalised destination prediction // 2021 IEEE International Conference on Cognitive and Computational Aspects of Situation Management (CogSIMA) : Virtual Conference, 14-22 May 2021 : proceedings. Danvers : IEEE, 2021. p. 9-16.
https://doi.org/10.1109/CogSIMA51574.2021.9475921 - Alumäe, T., Valk, J. The TalTech systems for the short-duration speaker verification challenge 2020 // 21st Annual Conference of the International Speech Communication Association (INTERSPEECH 2020): Cognitive Intelligence for Speech Processing : Proceedings of a meeting held 25-29 October 2020, Shanghai, China. Red Hook : International Speech Communication Association, 2020. p. 746-750.
https://www.isca-speech.org/archive/interspeech_2020/alumae20_interspeech.html http://www.proceedings.com/56854.html - Paulsen, G., Tuulik, M., Lohk, A., Vainik, E. From verbal to adjectival : evaluating the lexicalization of participles in an Estonian corpus // Slovenščina 2.0 (2022) vol. 10, 1, p. 65-97.
https://doi.org/10.4312/slo2.0.2022.1.65-97 - Lohk, A., Orav, H., Vare, K., Bond, F., Vaik, R. New polysemy structures in Wordnets induced by vertical polysemy // Proceedings of the 10th Global WordNet Conference : GWC 2019, July 23–27, 2019, Wroclaw, Poland. Wroclaw : Oficyna Wydawnicza Politechniki Wroclawskiej, 2019. p. 394-403.
https://clarin-pl.eu/dspace/handle/11321/718 "scopus" - Tuulik, M., Vainik, E., Paulsen, G., Lohk, A. Kuidas ära tunda adjektiivi? Korpuskäitumise mustrite analüüs // Eesti Rakenduslingvistika Ühingu aastaraamat 2022 = Estonian papers in applied linguistics 2022. Tallinn : Eesti Keele Sihtasutus, 2022. lk. 279-302. (Eesti Rakenduslingvistika Ühingu aastaraamat ; 18).
https://doi.org/10.5128/ERYa18.16 https://www.ester.ee/record=b2033361*est - Viht, A., Lohk, A. Kvantitatiivne vaade Uue Testamendi 1630.-1730. aastate tõlgetele // Emakeele Seltsi aastaraamat (2022) vol. 67, 1, lk. 169-194.
https://doi.org/10.3176/esa67.09 - Härm, H., Alumäe, T. Abstractive summarization of broadcast news stories for Estonian // Baltic journal of modern computing (2022) vol. 10, 3, p. 511-524.
https://doi.org/10.22364/bjmc.2022.10.3.23 - Alumäe, T., Kong, J. Combining hybrid and end-to-end approaches for the OpenASR20 challenge // Interspeech 2021 : Brno, Czechia, 30 August - 3 September 2021 : Proceedings of the 22nd Annual Conference of the International Speech Communication Association, Brno 30 August 2021 through 3 September 2021. : International Speech Communication Association, 2021. p. 1585-1589.
https://doi.org/10.21437/Interspeech.2021-1086 - Talts, S., Alumäe, T. Analyzing candidate speaking time in Estonian Parliament election debates // DHN 2020 : Digital Humanities in the Nordic Countries : proceedings of the Digital Humanities in the Nordic Countries 5th Conference : Riga, Latvia, October 21-23, 2020. Aachen : CEUR-WS.org, 2020. p. 351–363. (CEUR workshop proceedings ; 2612).
http://ceur-ws.org/Vol-2612/short22.pdf - Tavi, L., Kinnunen, T., Meister, E., Gonzalez-Hautamäki, R., Malmi, A. Articulation during voice disguise: a pilot study // Speech and Computer : 23rd International Conference, SPECOM 2021, St. Petersburg, Russia, September 27-30, 2021 : proceedings. Cham : Springer Nature Switzerland AG, 2021. p. 680-691. (Lecture notes in artificial intelligence ; 12997).
https://doi.org/10.1007/978-3-030-87802-3_61 - Meister, E., Meister, L. Developmental changes of vowel acoustics in adolescents // Interspeech 2021 : Brno, Czechia, 30 August - 3 September 2021 : Proceedings of the 22nd Annual Conference of the International Speech Communication Association, Brno 30 August 2021 through 3 September 2021. : International Speech Communication Association, 2021. p. 1688-1692.
https://doi.org/10.21437/Interspeech.2021-1649 - Malmi, A., Lippus, P., Meister, E. Articulatory properties of Estonian palatalization by Russian L1 speakers // Eesti ja soome-ugri keeleteaduse ajakiri = Journal of Estonian and Finno-Ugric linguistics (2022) vol. 13, 2, p. 79-118 : ill.
https://doi.org/10.12697/jeful.2022.13.2.03 - Meister, E., Meister, L. Estonian elderly speech corpus - design, collection and preliminary acoustic analysis // Baltic journal of modern computing (2022) vol. 10, 3, p. 360-371.
https://doi.org/10.22364/bjmc.2022.10.3.09 - Piel, L.K., Alumäe, T. Speech-based identification of children’s gender and age with neural networks // Human Language Technologies - the Baltic Perspective : Proceedings of the Eighth International Conference, Baltic HLT 2018. Amsterdam : IOS Press, 2018. p. 104-111. (Frontiers in artificial intelligence and applications ; 307).
https://doi.org/10.3233/978-1-61499-912-6-104 - Lohk, A., Tombak, M., Vare, K. An experiment : using Google Translate and semantic mirrors to create synsets with many lexical units // Proceedings of the 9th Global WordNet Conference : GWC 2018, January 8-12, 2018, Singapore. [S.l.] : The Global Word Net Association, 2018. p. 328-332.
http://doi.org/10.1109/EmpiRE.2018.00012 - Ullah, A., Alumäe, T. Data augmentation and teacher-student training for LF-MMI based robust speech recognition // Text, Speech, and Dialogue : 21st International Conference, TSD 2018, Brno, Czech Republic, September 11-14, 2018 : proceedings. Cham : Springer International Publishing AG, 2018. p. 403-410. (Lecture notes in computer science ; 11107).
https://doi.org/10.1007/978-3-030-00794-2_43 - Meister, E., Meister, L. Production of Estonian vowels by Finnish speakers // Kõneuurimise suundi II = Aspects of speech studies II. Tartu : University of Tartu Press, 2019. p. 129–143 : ill. (Eesti ja soome-ugri keeleteaduse ajakiri = Journal of Estonian and Finno-Ugric linguistics ; vol. 10, 1).
https://doi.org/10.12697/jeful.2019.10.1.07 - Lohk, A., Ross, K. Joachim Rossihniuse ja Heinrich Stahli perikoopide võrdlus, A comparison of the pericopes of Joachim Rossihnius and Heinrich Stahl // Emakeele Seltsi aastaraamat. Tallinn : Teaduste Akadeemia Kirjastus, 2019. lk. 65–110.
https://doi.org/10.3176/esa64.03 https://kirj.ee/the-yearbook-of-the-estonian-mother-tongue-society-publications/?filter[year]=2019&filter[issue]=370&filter[publication]=2952 - Karu, M., Alumäe, T. Weakly supervised training of speaker identification models // Odyssey 2018 : The Speaker and Language Recognition Workshop, 26-29 June 2018, Les Sables d'Olonne, France : proceedings. San Francisco : International Speech Communication Association, 2018. p. 24-30 : ill.
https://www.isca-speech.org/archive/Odyssey_2018/pdfs/41.pdf - Paulsen, G., Vainik, E., Tuulik, M., Lohk, A. The lexicographer's voice : word classes in the digital era // Electronic lexicography in the 21st century: Smart lexicography : proceedings of the eLex 2019 conference. Brno : Lexical Computing CZ s.r.o., 2019. p. 319-337 : ill.
https://elex.link/elex2019/wp-content/uploads/2019/10/eLex-2019_Proceedings.pdf - Tavi, L., Alumäe, T., Werner, S. Recognition of creaky voice from emergency calls // Interspeech 2019 : 15-19 September 2019, Graz. [S.l.] : International Speech Communication Association, 2019. p. 1990-1994 : ill.
https://doi.org/10.21437/Interspeech.2019-1253 - Tavi, L., Alumäe, T., Werner, S. Recognition of creaky voice from emergency calls // INTERSPEECH 2019 : "Crossroads of speech and language" : Grac-Austria, September 15th-19th 2019. [S.l.] : International Speech Communication Association, 2019. p. 191.
https://www.isca-speech.org/archive/Interspeech_2019/booklet.pdf - Alumäe, T., Tilk, O., Ullah, A. Advanced rich transcription system for Estonian speech // Human Language Technologies - the Baltic Perspective : Proceedings of the Eighth International Conference, Baltic HLT 2018. Amsterdam : IOS Press, 2018. p. 1-8. (Frontiers in artificial intelligence and applications ; 307).
https://doi.org/10.3233/978-1-61499-912-6-1 - Vainik, E., Lohk, A., Paulsen, G. The Distribution Index Calculator for Estonian // Electronic lexicography in the 21st century : post-editing lexicography : proceedings of the eLex 2021 conference : virtual, 5–7 July 2021. Brno : Lexical Computing CZ s.r.o, 2021. p. 121-138.
https://elex.link/elex2021/proceedings-download/ - Paulsen, G., Vainik, E., Lohk, A., Tuulik, M. Catching lexemes. The case of Estonian noun-based ambiforms // Electronic lexicography in the 21st century : post-editing lexicography : proceedings of the eLex 2021 conference : virtual, 5–7 July 2021. Brno : Lexical Computing CZ s.r.o, 2021. p. 288-311 : ill.
https://elex.link/elex2021/proceedings-download/ - Klavan, J., Alumäe, T., Tavast, A. Eesti keele väliskohakäänete kasutus poolspontaanses kõnes automaatse transkriptsiooni põhjal // Keel ja Kirjandus (2020) 8-9, lk. 757-774 : tab.
https://keeljakirjandus.ee/ee/uncategorized/eesti-keele-valiskohakaanete-kasutus-poolspontaanses-kones-automaatse-transkriptsiooni-pohjal/ https://www.ester.ee/record=b1072340*est https://doi.org/10.54013/kk754a8 - Kelli, A., Vider, K., Kull, I., Siil, T., Linden, K., Tavast, A., Värv, A., Ginter, C., Meister, E. Keeleressursside loomise ja kasutamisega seonduvaid isikuandmete kaitse küsimusi // Eesti Rakenduslingvistika Ühingu aastaraamat 14 = Estonian papers in applied linguistics 14. Tallinn : Eesti Rakenduslingvistika Ühing, 2018. lk. 77-94. (Eesti Rakenduslingvistika Ühingu aastaraamat ; 14).
http://www.ester.ee/record=b2033361*est https://doi.org/10.5128/ERYa14.05 - Meister, E., Meister, L. Eesti laste kõne II. Vokaalide akustiline analüüs // Keel ja Kirjandus (2019) 62, 4, lk. 282–295 : ill.
https://dea.digar.ee/article/AKkeeljakirjandus/2019/04/0/8 - Vainik, E., Paulsen, G., Lohk, A., Tuulik, M. Towards the morphosyntactic corpus profile of prototypical adjectives in Estonian // Eesti Rakenduslingvistika Ühingu aastaraamat 2023 = Estonian papers in applied linguistics 2023. Tallinn : Eesti Rakenduslingvistika Ühing, 2023. p. 225-244 : ill. (Eesti Rakenduslingvistika Ühingu aastaraamat ; 19).
https://doi.org/10.5128/ERYa19.13 https://www.ester.ee/record=b2033361*est - Vainik, E., Paulsen, G., Lohk, A. Käändevormist sõnaks : mida näitab sagedus? // Eesti Rakenduslingvistika Ühingu aastaraamat 2021 = Estonian Papers in Applied Linguistics 2021. Tallinn : Eesti Rakenduslingvistika Ühing, 2021. lk. 285-307. (Eesti Rakenduslingvistika Ühingu aastaraamat ; 17).
https://doi.org/10.5128/ERYA17.16 https://www.ester.ee/record=b2033361*est - Meister, E., Meister, L. Eesti laste kõne III. Kõnetempo ja silbikestuste analüüs // Keel ja Kirjandus (2022) 3, lk. 226-244.
https://doi.org/10.54013/kk771a3 - Vurma, A., Dede, T., Kala, V., Meister, E., Meister, L., Raju, M., Ross, J. Vokaalide ja klusiilide intensiivsussuhted laulmisel teksti arusaadavuse mõjutajana // Eesti-uuringute Tippkeskuse kokkuvõttekonverents "Dialoogid Eestiga. Uus algus" : 15.-16. veebruar 2023 Tartus : kava, teesid. Tartu : EKM Teaduskirjastus, 2023. p. 43−44.
https://www.folklore.ee/CEES/2023/finaal/media/teesid.pdf - Lohk, A., Vainik, E., Paulsen, G., Rebane, M., Bond, F. Extended clusters of vertical polysemy : an explorative study of eleven wordnets // Eesti Rakenduslingvistika Ühingu aastaraamat 2021 = Estonian Papers in Applied Linguistics 2021. Tallinn : Eesti Rakenduslingvistika Ühing, 2021. p. 193-210 : ill. (Eesti Rakenduslingvistika Ühingu aastaraamat ; 17).
https://doi.org/10.5128/ERYa17.11 https://www.ester.ee/record=b2033361*est - Paulsen, G., Lohk, A., Tuulik, M., Vainik, E. From experiments to an application : the first prototype of an adjective detector for Estonian // Electronic lexicography in the 21st century (eLex 2023) : Invisible Lexicography : proceedings of the eLex 2023 conference. Brno : LexicalComputingCZ s.r.o., 2023. p. 476-500.
https://elex.link/elex2023/wp-content/uploads/elex2023_proceedings.pdf