Joost Bastings

Welcome

Hi there! I’m a PhD candidate at the Institute for Logic, Language and Computation (ILLC) at the University of Amsterdam.

I work on Neural Machine Translation and Machine Learning for NLP under supervision of Wilker Aziz, Ivan Titov, and Khalil Sima’an. My research is part of the project Machine Translators: Teaching Computers to Translate Using Their Own Words.

News & Activities

• May 2019. Our paper “Interpretable Neural Predictions with Differentiable Binary Variables” was accepted to ACL 2019!
• Sep 2018. I released a jupyter notebook & blogpost on NMT in PyTorch: The Annotated Encoder-Decoder with Attention.
• Sep 2018. I will be at BlackboxNLP (EMNLP) presenting the paper “Jump to better conclusions: SCAN both left and right”
• July 2018. I’m presenting a poster at ACL NMT workshop 2018 on Modeling Sentence Structure in Neural Machine Translation.
• Jan-April 2018. Research intern at Facebook AI Research. New York, USA.
• 25 Sep 2017. Google NLP Summit. Zurich, Switzerland.
• 9 Sep 2017. EMNLP 2017.
• 7 Sep 2017. Invited talk at DLID workshop, BMVC 2017, London.
• Jul/Aug 2017. ACL 2017 and ACL NMT workshop.
• 12 July 2017. Invited talk at Textkernel.
• Jul 2017. Lab teacher at Lisbon Machine Learning Summer School (LxMLS 2017). Lisbon, Portugal.
• Jul 2016. Lisbon Machine Learning Summer School (LxMLS 2016). Lisbon, Portugal.
• Oct 2015. Deep Learning 4 Machine Translation (DL4MT) Winter School. Dublin, Ireland.
• Sep 2015. Google NLP Summit. Zurich, Switzerland.

Publications

• Joost Bastings, Wilker Aziz, Ivan Titov (2019). Interpretable Neural Predictions with Differentiable Binary Variables. ACL. Florence, Italy. [bib] [code]

• Joost Bastings, Marco Baroni, Jason Weston, Kyunghyun Cho, Douwe Kiela (2018). Jump to better conclusions: SCAN both left and right. Workshop on Analyzing and Interpreting Neural Networks for NLP, EMNLP. Brussels, Belgium. [bib] [poster]

• Joost Bastings. 2018. The Annotated Encoder-Decoder with Attention. https://bastings.github.io/annotated_encoder_decoder/ [bib]

• Diego Marcheggiani, Joost Bastings, Ivan Titov (2018). Exploiting Semantics in Neural Machine Translation with Graph Convolutional Networks. NAACL-HLT’18. New Orleans, USA. [bib]

• Joost Bastings, Wilker Aziz, Ivan Titov, Khalil Sima’an (2018). Modeling Latent Sentence Structure in Neural Machine Translation. ACL’18 NMT workshop (extended abstract). Melbourne, Australia. [poster]

• Joost Bastings, Ivan Titov, Wilker Aziz, Diego Marcheggiani, Khalil Sima’an (2017). Graph Convolutional Encoders for Syntax-aware Neural Machine Translation. EMNLP’17. Copenhagen, Denmark. [bib] [code] [data]

• Jan-Thorsten Peter, Hermann Ney, Ondrej Bojar, Ngoc-Quan Pham, Jan Niehues, Alex Waibel, Franck Burlot, François Yvon, Mārcis Pinnis, Valters Šics, Joost Bastings, Miguel Rios, Wilker Aziz, Philip Williams, Frédéric Blain, and Lucia Specia (2017). The QT21 Combined Machine Translation System for English to Latvian. WMT’17. Copenhagen, Denmark. [bib]

• Joost Bastings and Khalil Sima’an (2014). All Fragments Count in Parser Evaluation. Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14). Reykjavik, Iceland. [bib] [code]

Reviewing

• CoNLL 2019
• EMNLP 2019
• ACL 2019
• ICLR 2019 Debug ML workshop
• WMT 2018
• CONLL 2018
• EMNLP 2018
• ACL NMT 2018

Supervision

• Laura Ruis (June 2018). MSc AI individual project.
• Joost Baptist (2017). MSc Thesis AI. Graduated.

Teaching

• MSc AI, Natural Language Processing 1. Fall 2018. TA & guest lecturer on SMT.
• MSc AI, Natural Language Processing 1. Fall 2017. TA & guest lecturer on SMT.
• MSc AI, Natural Language Processing 2. Spring 2017. TA & guest lecturer on NMT.
• MSc AI, Natural Language Processing 1. Fall 2016. TA & guest lecturer on SMT.
• MSc AI, Logic Programming / Prolog. Fall 2015. TA.

Talks

• 7 Sep 2017. Invited talk at Deep Learning on Irregular Domains (DLID) workshop, BMVC 2017, London. Graph Convolutional Networks for Natural Language Processing.
• 12 July 2017. Invited talk at Textkernel. Using Graph Convolutional Encoders for improved Neural Machine Translation.

Short CV

• 2018 Research Intern at Facebook AI Research. New York, USA.
• 2015–Now. PhD Candidate at the Institute for Logic, Language and Computation, University of Amsterdam.
• 2012–2014. Trainee at Volvo Cars. Gothenburg, Sweden.
• 2010 Extreme Blue & Explovate intern at IBM. Amsterdam, The Netherlands.

Education

MSc Artificial Intelligence

University of Amsterdam. 2009 – 2012. Cum laude.

Thesis: The Graph Symbol Model: Context-Free Parsing with Context

BSc Artificial Intelligence

Utrecht University. 2006 – 2009. Cum laude.

Thesis: Finding Words in Continuous Speech

Contact

Check out my Github profile @bastings or find my contact details at my ILLC page.

Follow @joostbastings