about¶

I am a Research Scientist at Grammarly working on Machine Learning and Natural Language Processing. My work focuses on representation learning, affective computing, and personalization. Most of my research draws from Computational Psychology, Natural Language Processing, and crowdsourcing to build scalable unsupervised solutions that perform structured prediction, personalization, semantic searches, sentiment analysis, and text classification.

Before that, I was a PhD student with a focus on Computational Psycholinguistics and Semantics at the Department of Theoretical and Applied Linguistics at the University of Cambridge, supervised by Dr John Williams.

My research employed methods from statistics and machine learning, in order to gain a better understanding of how we learn and process language. My primary interest was to identify, through computational modelling and behavioural experiments, the strategies that allow people to unconsciously pick up linguistic regularities from their environment.

My Erdős number is 5.

publications¶

peer-reviewed publications:

• Raheja V., Alikaniotis, D. (2020). Adversarial Grammatical Error Correction, EMNLP 2020
• Alikaniotis, D., Raheja V. (2019). The Unreasonable Effectiveness of Transformer Language Models in Grammatical Error Correction, In Proceedings of the 14th Workshop on Innovative Use of NLP for Building Educational Applications (BEA-2019) Florence, Italy.
• Bentz, C., Alikaniotis, D., Cysouw, M., Ferrer-Cancho, R. (2017). The Entropy of Words — Learnability and Expressivity across More than 1000 Languages, Entropy 19(6)
• Bentz, C. & Alikaniotis, D., (2016). The Word Entropy of Natural Languages
• Alikaniotis, D., Yannakoudakis, H. & Rei, M. (2016). Automatic Text Scoring using Neural Networks in Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (ACL-2016)
• Bentz, C., Alikaniotis, D., Samardžić, T. & Buttery, P. (2017). Variation in word frequency distributions: Definitions, measures and implications for a corpus-based language typology, Journal of Quantitative Linguistics.
• Zhang W., Caines A., Alikaniotis D. & Buttery P. (2016). Predicting author age from Weibo microblog posts. Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC-2016)
• Caines, A. P., Bentz, C., Alikaniotis, D., Katushemererwe F. & Buttery, P. (2016). The Glottolog Data Explorer: Mapping the world's languages. Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC-2016)
• Alikaniotis, D. & Williams, J. N. (2015). A Distributional Semantics Approach to Implicit Language Learning, in Vito Pirrelli, Claudia Marzi, Marcello Ferro (eds.): Word Structure and Word Usage. (NetWordS-2015).

other:

• Alikaniotis, D. (2017). Understanding Semantic Implicit Learning through distributional linguistic patterns. A computational perspective, University of Cambridge, Cambridge, UK.
• Williams, J. N., Sheehan M., Paciorek, A. & Alikaniotis, D.. (2016). Linguistic naturalness and semi-artificial language learning. Fifth Implicit Learning Seminar, Lancaster, UK.
• Alikaniotis, D. & Williams, J. N. (2015). Semantic Constraints in Implicit Language Learning. Cambridge Language Sciences Annual Symposium, Cambridge, UK.
• Alikaniotis, D. (2015). Modelling implicit language learning with distributional semantics, NLIP Seminar Series, University of Cambridge.
• Alikaniotis, D. & Williams, J. N. (2015). Predicting semantic constraints in implicit language learning with distributional semantics, Interdisciplinary Advances in Statistical Learning, San Sebastian, Spain.
• Alikaniotis, D. (2014). Computational Methods for Calculating Semantic Similarity. Workshop in Computational Linguistics. University of Cambridge, February 2014.
• Alikaniotis, D. (2012). Implicit Rule Learning in a Computer Language Game. MPhil Research Project, University of Cambridge, April 2012.
• Alikaniotis, D. (2011). Pragmatic Motivations for Erroneous Reasoning. MPhil Research Project, University of Cambridge, December 2011.
• Alikaniotis, D. (2011). Computational Linguistics and Electronic Lexicography. Bachelor's Thesis, National and Kapodistrian University of Athens.

software¶

• References management:
  These two packages were designed to speed up productivity in Emacs/LaTeX. Pyopl is a standalone extendable command-line tool which queries several engines and returns the reference in bib format. Empos is an emacs-lisp wrapper which lets you invoke pyopl from within emacs.
• Python Online Paper Locator.
• EMacs Paper Online Search.
• Reproducible work:
  I believe that proper scientific work should be as transparent as possible. In my view, this means that methods/stimuli/data/settings or software used should be publicly available (at least wherever possible). Below you can find snippets from my own work or -hopefully- useful code for setting up experiments.
• Statistical learning experiment (as in Saffran et al. 1996)
• Mask Maker Generates masks from words (e.g. _li__t_or for alligator) including useful information.
• Attractor Network (as in Rabovsky & McRae, 2014)

teaching¶

Courses I taught while at Cambridge:
• Li3: Language, brain and society
• Li16: Psychology of Language Processing and Learning
• Li18: Computational Linguistics
• MPhil Research Seminars
• Foundations in Applied Statistics
• Basic Quantitative Analysis

I also supervised the following undergraduate dissertations
• Stride, A. Methods for modelling relevance of paragraphs within legal corpora.
• Searle, M. Conceptual metaphor in the embodiment of abstract concepts.

contact¶