My research¶

Projects¶

I am currently working on the following funded research projects:

• NWO VI.Veni.212.228 “Rethinking Natural Language Generation: Bridging the Gap Between Discrete and Continuous Representations”
• Horizon Europe 101070631 “UTTER: Unified Transcription and Translation for Extended Reality”
• Trustworthy AI for Media (TAIM)

Publications¶

(Google Scholar, Semantic Scholar, ACL Anthology links.)

Book-like things:

• Vlad Niculae, Caio F. Corro, Nikita Nangia, Tsvetomila Mihaylova, André F. T. Martins. Discrete Latent Structure in Neural Networks. 2023. [ arXiv preprint ]

• Vlad Niculae. Learning Deep Models with Linguistically-Inspired Structure. Doctoral dissertation, Cornell University. 2018. [ open access doi ]

Papers:

• Evgeniia Tokarchuk, Vlad Niculae. The unreasonable effectiveness of random target embeddings for continuous-output neural machine translation. NAACL, 2024 (to appear). [ arXiv preprint ]

• Wafaa Mohammed, Vlad Niculae. On measuring context utilization in document-level MT systems. EACL Findings, 2024. [ anthology ] [ arXiv ]

• Sergey Troshin, Vlad Niculae. Wrapped β-Gaussians with compact support for exact probabilistic modeling on manifolds. TMLR 2023. [ pdf ] [ code ] [ openreview ]

• Vlad Niculae. Two derivations of Principal Component Analysis on datasets of distributions. 2023. [ arXiv preprint ]

• Ali Araabi, Vlad Niculae, Christof Monz. Joint Dropout: Improving generalizability in low-resource neural machine translation through phrase pair variables. In: MT Summit 2023. [ anthology ] [ arXiv ]

• David Stap, Vlad Niculae, Christof Monz. Viewing knowledge transfer in multilingual machine translation through a representational lens. Findings of the ACL, 2023. [ arXiv ]

• Valentina Zantedeschi, Luca Franceschi, Jean Kaddour, Matt J Kusner, Vlad Niculae. DAG learning on the Permutahedron. In: ICLR 2023. [ arXiv ]

• André F. T. Martins, Marcos Treviso, António Farinhas, Pedro M. Q. Aguiar, Mário A. T. Figueiredo, Mathieu Blondel, Vlad Niculae. Sparse continuous distributions and Fenchel-Young losses. JMLR 2022. [ jmlr abs ] [ arXiv preprint ] [ code ]

• Ali Araabi, Christof Monz, Vlad Niculae. How effective is Byte Pair Encoding for out-of-vocabulary words in Neural Machine Translation? In: AMTA 2022. [ anthology ]

• Evgeniia Tokarchuk, Vlad Niculae. On target representation in continuous-output Neural Machine Translation. In: Repl4NLP 2022: Workshop on Representation Learning for NLP. [ anthology ]

• Valentina Zantedeschi, Jean Kaddour, Luca Franceschi, Matt Kusner, Vlad Niculae. DAG learning on the Permutahedron. In: ICLR 2022 Workshop on the Elements of Reasoning: Objects, Structure and Causality. [ pdf ]

• Tsvetomila Mihaylova, Vlad Niculae, André F. T. Martins. Modeling structure with Undirected Neural Networks. In: Proc. of ICML 2022. [ arXiv ]

• António Farinhas, Wilker Aziz, Vlad Niculae, André F. T. Martins. Sparse communication via mixed distributions. In: Proc. of ICLR 2022. [ arXiv ]

• Valentina Zantedeschi, Matt J. Kusner, Vlad Niculae. Learning binary trees by argmin differentiation. In: Proc. of ICML 2021 [ arXiv ] [ code ]

• Pedro Henrique Martins, Vlad Niculae, Zita Marinho, André F. T. Martins. Sparse and structured visual attention. In: Proc. of ICIP 2021, IEEE [ arXiv ]

• André F. T. Martins, Marcos Treviso, António Farinhas, Vlad Niculae, Mário A. T. Figueiredo, Pedro M. Q. Aguiar. Sparse and Continuous Attention Mechanisms. In: Proc. of NeurIPS 2020 [ arXiv ]

• Mathieu Blondel, André F. T. Martins, Vlad Niculae. Learning with Fenchel-Young losses. JMLR 2020. [ arXiv ] [ code ]

• Tsvetomila Mihaylova, Vlad Niculae, André F. T. Martins. Understanding the mechanics of SPIGOT: Surrogate gradients for latent structure learning. In: Proc. of EMNLP 2020. [ anthology ]

• Gonçalo M. Correia, Vlad Niculae, Wilker Aziz, André F. T. Martins. Efficient Marginalization of Discrete and Structured Latent Variables via Sparsity. In: Proc. of NeurIPS 2020. [ arXiv ] [ code ]

• Vlad Niculae and André F. T. Martins. LP-SparseMAP: Differentiable relaxed optimization for sparse structured prediction. In: Proc. of ICML 2020 [ arXiv ] [ code ]

• Gonçalo M. Correia, Vlad Niculae, André F. T. Martins. Adaptively sparse transformers. In: Proc. of EMNLP-IJCNLP 2019 [ arXiv ] [ code ]

• Ben Peters, Vlad Niculae, André F. T. Martins. Sparse sequence-to-sequence models. In: Proc. of ACL 2019. [ anthology ] [ arXiv ] [ code ]

• Mathieu Blondel, André F. T. Martins, Vlad Niculae. Learning classifiers with Fenchel-Young losses: Generalized entropies, margins, and algorithms. In: Proc. of AISTATS 2019. [ arXiv ] [ code ]

• Vlad Niculae, André F. T. Martins, Mathieu Blondel, Claire Cardie. SparseMAP: Differentiable sparse structured inference. In: Proc. of ICML 2018. [ arXiv ] [ code ] [ slides ] [ video ]

• Vlad Niculae, André F. T. Martins, Claire Cardie. Towards dynamic computation graphs via sparse latent structure. In: Proc. of EMNLP 2018. [ anthology ] [ arXiv ] [ code ] [ slides ] [ video ]

• Vlad Niculae and Mathieu Blondel. A regularized framework for sparse and structured neural attention. In: Proc. of NeurIPS 2017. [ arXiv ] [ code ]

• Mathieu Blondel, Vlad Niculae, Takuma Otsuka, Naonori Ueda. Multi-output polynomial networks and factorization machines. In: Proc. of NeurIPS 2017. [ arXiv ]

• Vlad Niculae, Joonsuk Park, Claire Cardie. Argument mining with structured SVMs and RNNs. In: Proc. of ACL 2017. [ anthology ] [ arXiv ] [ code ] [ data ] [ video ]

• Vlad Niculae and Cristian Danescu-Niculescu-Mizil. Conversational markers of constructive discussions. In: Proc. of NAACL-HLT 2016. [ website ] [ anthology ]

• Chenhao Tan, Vlad Niculae, Cristian Danescu-Niculescu-Mizil, Lillian Lee. Winning Arguments: Interaction Dynamics and Persuasion Strategies in Good-faith Online Discussions. In: Proc. of WWW 2016. [ website ]

• Vlad Niculae, Srijan Kumar, Jordan Boyd-Graber, Cristian Danescu-Niculescu-Mizil. Linguistic harbingers of betrayal: A case study on an online strategy game. In: Proc. of ACL 2015. [ website ] [ anthology ]

• Vlad Niculae*, Caroline Suen*, Justine Zhang*, Cristian Danescu-Niculescu-Mizil, Jure Leskovec. QUOTUS: The structure of political media coverage as revealed by quoting patterns. In: Proc. of WWW 2015. [ website ] [ slides ]

• Marcos Zampieri, Alina Maria Ciobanu, Vlad Niculae, Liviu P. Dinu. AMBRA: A ranking approach to temporal text classification. In: Proc. of Semeval 2015. [ anthology ] [ code ]

• Vlad Niculae and Cristian Danescu-Niculescu-Mizil. Brighter than Gold: Figurative language in user generated comparisons. In: Proc. of EMNLP 2014. [ website ] [ anthology ]

• Vlad Niculae, Marcos Zampieri, Liviu P. Dinu, Alina Maria Ciobanu. Temporal text ranking and automatic dating of texts. In: Proc. of EACL 2014. [ anthology ] [  slides  ]

• Vlad Niculae. Comparison pattern matching and creative simile recognition. In: Proc. of JSSP 2013. [ anthology ] [ poster ] [ code ]

• Vlad Niculae and Octavian Popescu. Determining is-a relationships for textual entailment. In: Proc. of JSSP 2013. [ anthology ] [ poster ]

• Lars Buitinck, Gilles Louppe, Mathieu Blondel, Fabian Pedregosa, Andreas Müller, Olivier Grisel, Vlad Niculae, Peter Prettenhofer, Alexandre Gramfort, Jaques Grobler, Robert Layton, Jake Vanderplas, Arnaud Joly, Brian Holt and Gaël Varoquaux. API design for machine learning software: experiences from the scikit-learn project. In: ECML/PKDD 2013 Workshop: Languages for Data Mining and Machine Learning. [ PDF ]

• Vlad Niculae, Victoria Yaneva, Computational considerations of comparisons and similes. In: Proceedings of ACL Student Research Workshop, 2013. [ anthology ] [ poster ]