Hello there!

This homepage is mainly a list of references to my academic contributions: papers, datasets, and code. Please feel free to drop me a line if you're interested in any of my work!

Resources and Tools

• nlp.ffzg.hr: free resources and tools for Croatian NLP (tagger, parser, NER, etc.)
• SETimes.HR+: Croatian dependency treebank (and other data)
• META-SHARE: more resources for Croatian
• irtg.googlecode.com: IRTG toolbox
• Universal Dependencies: cross-linguistically consistent treebank annotations
• LOWLANDS: papers datasets
• XNLI: test data for cross-lingual inference

Publications

2023

• A Perez, B Vasilev, Ž Agić, C Thrysøe, V Hargitai, M Dahlgaard, C Røssel. Accelerating and Scaling Data Products with Serverless. In Serverless Computing: Principles and Paradigms, pp. 149--173. link

2021

• Ž Agić. Sharing Dependency Parsers between Similar Languages. In Similar Languages, Varieties, and Dialects: A Computational Perspective, pp. 167--184. link

2020

• L Borgholt, J D Havtorn, Ž Agić, A Søgaard, L Maaløe, C Igel. Do End-to-End Speech Recognition Models Care About Context? In Proc. Interspeech. pdf
• J D Havtorn, J Latko, J Edin, L Maaløe, L Borgholt, L Belgrano, N F Jakobsen, R Sdun, Ž Agić. MultiQT: Multimodal Learning for Real-Time Question Tracking in Speech. In Proc. ACL, pp. 2370--2380. pdf
• R Litschko, I Vulić, Ž Agić, G Glavaš. Towards Instance-Level Parser Selection for Cross-Lingual Transfer of Dependency Parsers. In Proc. COLING. pdf

2019

• Ž Agić, I Vulić. JW300: A Wide-Coverage Parallel Corpus for Low-Resource Languages. In Proc. ACL, pp. 3204--3210. pdf poster data no longer available
• B Plank, S Klerke, Ž Agić. The Best of Both Worlds: Lexical Resources To Improve Low-Resource Part-of-Speech Tagging. arXiv. pdf

2018

• B Plank, Ž Agić. Distant supervision from disparate sources for low-resource part-of-speech tagging. In Proc. EMNLP, pp. 614--620. pdf code
• Ž Agić, N Schluter. Baselines and test data for cross-lingual inference. In Proc. LREC, pp. 3890--3894. pdf data
• J V Enghoff, S Harrison, Ž Agić. Low-resource named entity recognition via multi-source projection: Not quite there yet? In Proc. W-NUT at EMNLP, pp. 195--201. pdf
• A Vásquez, R E Aguirre, C Angulo, J Miller, C Villanueva, Ž Agić, R Zariquiey, A Oncevay. Toward Universal Dependencies for Shipibo-Konibo. In Proc. UDW at EMNLP. pdf
• N Ljubešić, Ž Agić, F Klubička, V Batanović, T Erjavec. hr500k -- A reference training corpus of Croatian. In Proc. JTDH. pdf data

2017

• Ž Agić, N Schluter. How (not) to train a dependency parser: The curious case of jackknifing part-of-speech taggers. In Proc. ACL. pdf poster
• Ž Agić, B Plank, A Søgaard. Cross-lingual tagger evaluation without test data. In Proc. EACL, pp. 248--253. pdf poster
• H Martínez Alonso, Ž Agić, B Plank, A Søgaard. Parsing Universal Dependencies without training. In Proc. EACL, pp. 229--239. pdf
• Ž Agić. Cross-lingual parser selection for low-resource languages. In Proc. UDW at NODALIDA, pp. 1--10. pdf
• N Schluter, Ž Agić. Empirically sampling Universal Dependencies. In Proc. UDW at NODALIDA, pp. 117--122. pdf
• T Samardžić, M Starović, Ž Agić, N Ljubešić. Universal Dependencies for Serbian in comparison with Croatian and other Slavic languages. In Proc. BSNLP at EACL, pp. 39--44. pdf

2016

• Ž Agić, A Johannsen, B Plank, H Martínez Alonso, N Schluter, A Søgaard. Multilingual projection for parsing truly low-resource languages. TACL 4:301--312. pdf slides video
• A Johannsen, Ž Agić, A Søgaard. Joint part-of-speech and dependency projection from multiple sources. In Proc. ACL, pp. 561--566. pdf slides video
• J Tiedemann, Ž Agić. Synthetic treebanking for cross-lingual dependency parsing. Journal of Artificial Intelligence Research 55:209--248. pdf
• N Ljubešić, F Klubička, Ž Agić, I-P Jazbec. New Inflectional lexicons for improved processing of Croatian and Serbian. In Proc. LREC, pp. 4264--4270. pdf
• Ž Agić, D Merkler. Wise serpents and stupid owls? Animacy and sentiment in two centuries of English similes. Society & Animals, accepted for publication.

2015

• Ž Agić, D Hovy, A Søgaard. If all you have is a bit of the Bible: Learning POS taggers for truly low-resource languages. In Proc. ACL-IJCNLP, pp. 268--272. pdf slides video
• A Søgaard, Ž Agić, H Martínez Alonso, B Plank, B Bohnet, A Johannsen. Inverted indexing for cross-lingual NLP. In Proc. ACL-IJCNLP, pp. 1713--1722. pdf video
• B Plank, H Martínez Alonso, Ž Agić, D Merkler, A Søgaard. Do dependency parsing metrics correlate with human judgments? In Proc. CoNLL, pp. 315--320. pdf poster video
• Ž Agić, N Ljubešić. Universal Dependencies for Croatian (that work for Serbian, too). In Proc. BSNLP at RANLP, pp. 1--8. pdf slides
• Ž Agić, A Koller, S Oepen. Semantic dependency graph parsing using tree approximations. In Proc. IWCS, pp. 217--227. pdf slides
• M Barrett, Ž Agić, A Søgaard. The Dundee treebank. In Proc. TLT, pp. 242--248. pdf

2014

• Ž Agić, J Tiedemann, K Dobrovoljc, S Krek, D Merkler, S Može. Cross-lingual dependency parsing of related languages with rich morphosyntactic tagsets. In Proc. LT4CloseLang at EMNLP, pp. 12--23. pdf slides
• Ž Agić, A Koller. Potsdam: Semantic dependency parsing by bidirectional graph-tree transformations and syntactic parsing. In Proc. SemEval, pp. 465--470. pdf poster
• J Tiedemann, Ž Agić, J Nivre. Treebank translation for cross-lingual parser induction. In Proc. CoNLL, pp. 130--140. pdf poster slides
• X Carreras, L Padró, L Zhang, A Rettinger, Z Li, E García-Cuesta, Ž Agić, B Bekavac, B Fortuna, T Štajner. XLike project language analysis services. In Proc. EACL, pp. 9--12. pdf
• Ž Agić, N Ljubešić. The SETimes.HR linguistically annotated corpus of Croatian. In Proc. LREC, pp. 1724--1727. pdf
• Ž Agić, D Berović, D Merkler, M Tadić. Croatian dependency treebank 2.0: New annotation guidelines for improved parsing. In Proc. LREC, pp. 2313--2319. pdf poster
• L Padró, Ž Agić, X Carreras, B Fortuna, E García-Cuesta, Z Li, T Štajner, M Tadić. Language processing infrastructure in the XLike project. In Proc. LREC, pp. 3811--3816. pdf
• D Merkler, Ž Agić, M Tadić. Automatic enrichment of Croatian morphological lexicon Using large corpora and web search. In Proc. CILC.

2013

• J Šnajder, S Padó, Ž Agić. Building and evaluating a distributional memory for Croatian. In Proc. ACL, pp. 784--789. pdf slides
• Ž Agić, D Merkler, D Berović. Parsing Croatian and Serbian by using Croatian dependency treebanks. In Proc. SPMRL at EMNLP, pp. 22--33. pdf slides
• Ž Agić, D Merkler. Three syntactic formalisms for data-driven dependency parsing of Croatian. In LNCS, 8082:560--567. pdf slides
• Ž Agić, N Ljubešić, D Merkler. Lemmatization and morphosyntactic tagging of Croatian and Serbian. In Proc. BSNLP at ACL, pp. 48--57. pdf slides
• N Ljubešić, M Stupar, T Jurić, Ž Agić. Combining available datasets for building named entity recognition models of Croatian and Slovene. Slovenščina 2.0, 1(2):35--57. pdf
• Ž Agić, B Bekavac. Domain-aware evaluation of named entity recognition systems for Croatian. Journal of Computing and Information Technology, 21(3):195--209. pdf
• D Merkler, Ž Agić, A Agić. Babel treebank of public messages in Croatian. Procedia -- Social and Behavioral Sciences, 95:490--497. pdf poster
• Ž Agić, B Bekavac. Domain dependence of statistical named entity recognition and classification in Croatian texts. In Proc. ITI, pp. 277--283. pdf
• D Merkler, Ž Agić. Sentiscope: A system for sentiment analysis in daily horoscopes. In Proc. NooJ, pp. 173--181. poster

2012

• Ž Agić. K-best spanning tree dependency parsing with verb valency lexicon reranking. In Proc. COLING, pp. 1--12. pdf poster
• Ž Agić, D Merkler. Rule-based sentiment analysis in narrow domain: Detecting sentiment in daily horoscopes using Sentiscope. In Proc. SAAIP at COLING, pp. 115--124. pdf slides
• D Berović, Ž Agić, M Tadić. Croatian dependency treebank: Recent development and initial experiments. In Proc. LREC, pp. 1902--1906. pdf
• Ž Agić, D Merkler, D Berović. Slovene-Croatian treebank transfer using bilingual lexicon improves Croatian dependency parsing. In Proc. IS-LTC, pp. 5--9. pdf slides
• D Katunar, K Šojat, Ž Agić, Ž Stojanović. Visualization of the Croatian WordNet. In Applied Linguistics Today: Research and Perspectives, pp. 199--210.
• D Merkler, D Berović, Ž Agić. Disambiguation of homographic adjective and adverb forms in Croatian texts. In Proc. NooJ, pp. 142--151.
• Ž Agić. Pristupi ovisnosnom parsanju hrvatskih tekstova. Doctoral thesis, Faculty of Humanities and Social Sciences, University of Zagreb, 2012. pdf slides (in Croatian)

2011

• Ž Agić, D Merkler, D Berović, M Tadić. Development and applications of the Croatian 1984 corpus for the MULTEXT-East resources. In Proc. SlaviCorp.
• Ž Agić, N Ljubešić, M Tadić. NooJ module for sentiment identification in Croatian financial texts. In Proc. NooJ.

2010

• Ž Agić, N Ljubešić, M Tadić. Towards sentiment analysis of financial texts in Croatian. In Proc. LREC, pp. 1164--1167. pdf
• K Vučković, Ž Agić, M Tadić. Improving chunking accuracy on Croatian texts by morphosyntactic tagging. In Proc. LREC, pp. 1944--1949. pdf
• S Seljan, M Tadić, Ž Agić, J Šnajder, B Dalbelo Bašić, V Osmann. Corpus aligner (CorAl) evaluation on English-Croatian parallel corpora. In Proc. LREC, pp. 3481--3484. pdf
• Ž Agić, K Šojat, M Tadić. An experiment in verb valency frame extraction from Croatian dependency dreebank. In Proc. ITI, pp. 55--60. pdf
• Ž Agić, M Tadić, Z Dovedan. Tagger voting improves morphosyntactic tagging accuracy on Croatian texts. In Proc. ITI, pp. 61--66. (best student paper award) pdf
• K Šojat, Ž Agić, M Tadić. Verb valency frame extraction using morphological and syntactic features of Croatian. In Proc. FASSBL, pp. 119--126. pdf
• K Vučković, Ž Agić, M Tadić. Sentence classification and clause cetection for Croatian. In Proc. FASSBL, pp. 131--138. pdf
• B Bekavac, Ž Agić, M Tadić. Interacting Croatian NERC system and Intex/NooJ environment. In Proc. NooJ, pp. 21--29.

2009

• Ž Agić, M Tadić, Z Dovedan. Evaluating full lemmatization of Croatian texts. In Recent Advances in Intelligent Information Systems, pp. 175--184. pdf
• Ž Agić, M Tadić, Z Dovedan. Error analysis in Croatian morphosyntactic tagging. In Proc. ITI, pp. 521--526. pdf
• Ž Agić, M Tadić, Z Dovedan. Tagset reductions in morphosyntactic tagging of Croatian texts. In Proc. InFuture, pp. 289--298. pdf
• B Bekavac, Ž Agić, K Šojat, M Tadić. Detecting measurement expressions using NooJ. In Proc. NooJ, pp. 121--128. pdf

2008

• Ž Agić, M Tadić, Z Dovedan. Improving part-of-speech tagging accuracy for Croatian by morphological analysis. Informatica 32(4):445--451. pdf
• Ž Agić, M Tadić, Z Dovedan. Combining part-of-speech tagger and inflectional lexicon for Croatian. In Proc. IS-LTC, pp. 116--121. pdf
• I Raffaelli, M Tadić, B Bekavac, Ž Agić. Building Croatian WordNet. In Proc. GWC, pp. 349--359. pdf
• Ž Agić, M Tadić, Z Dovedan. Investigating language independence in HMM PoS/MSD-tagging. In Proc. ITI, pp. 657--662. pdf
• N Ljubešić, Ž Agić, N Bakarić. Document representation methods for news event detection in Croatian. In Proc. FASSBL, pp. 79--84. pdf
• S Seljan, Ž Agić, M Tadić. Evaluating sentence alignment on Croatian-English parallel corpora. In Proc. FASSBL, pp. 101--108. pdf

older

• Ž Agić. 2007. Jezične tehnologije. Sustavi 1(1):17--21.
• Ž Agić, M Tadić. 2006. Evaluating morphosyntactic tagging of Croatian texts. In Proc. LREC. pdf
• Ž Agić. 2005. Morfosintaktičko označavanje hrvatskoga jezika. Master's thesis, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of Split, 2005. (in Croatian)