Arkadiy Saakyan Last updated Oct 2023 --- EDUCATION Columbia University in the City of New York PhD in Computer Science (Natural Language Processing) Advisor: Prof. Smaranda Muresan May 2026 (expected) Columbia University in the City of New York M.S. in Computer Science GPA 4.30/4.33 May 2023 Columbia University in the City of New York, Columbia College B.A. in Computer Science, Concentration in Mathematics GPA 3.73/4.33 May 2021 --- PUBLICATIONS Google Scholar: https://tinyurl.com/scholar-saakyan Sociocultural Norm Similarities and Differences via Situational Alignment and Explainable Textual Entailment Sky CH-Wang*, Arkadiy Saakyan*, Oliver Li, Zhou Yu, and Smaranda Muresan Proceedings of EMNLP 2023 NormDial: A Comparable Bilingual Synthetic Dialog Dataset for Modeling Social Norm Adherence and Violation Oliver Li, Mallika Subramanian, Arkadiy Saakyan, Sky CH-Wang, and Smaranda Muresan Proceedings of EMNLP 2023 Learning to Follow Object-Centric Image Editing Instructions Faithfully Tuhin Chakrabarty, Kanishk Singh, Arkadiy Saakyan, Smaranda Muresan Findings of EMNLP 2023 Arkadiy Saakyan*, Tuhin Chakrabarty*, Olivia Winn*, Artemis Panagopoulou, Yue Yang, Marianna Apidianaki, and Smaranda Muresan. I Spy a Metaphor: Large Language Models and Diffusion Models Co-Create Visual Metaphors (Findings of ACL 2023) - Evaluate language model and text-to-image diffusion model collaboration for visual metaphor creation. Tuhin Chakrabarty, Arkadiy Saakyan, Debanjan Ghosh, and Smaranda Muresan. FLUTE: Figurative Language Understanding and Textual Explanations. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing (EMNLP 2022). (Long paper, Poster) - Generate a dataset for figurative language understanding with textual explanations via model-in-the-loop framework with GPT-3. Distilled resulting generations to smaller T5 model to significantly improve the quality of rationales generated to explain the reasoning behind figurative natural language inference problems. Tuhin Chakrabarty, Arkadiy Saakyan, and Smaranda Muresan (2021). Don’t Go Far Off: An Empirical Study on Neural Poetry Translation. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP 2021). (Long paper, Oral) - Present an empirical investigation on poetry translation along several dimensions: 1) size and style of training data (poetic vs. non-poetic), including a zero-shot setup; 2) bilingual vs. multilingual learning; and 3) language-family-specific models vs. mixed-multilingual models Arkadiy Saakyan, Tuhin Chakrabarty, and Smaranda Muresan (2021). COVID-Fact: Fact Extraction and Verification of Real-World Claims about COVID-19 Pandemic. In Proceedings of the Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL 2021). (Long paper, Oral) - Introduces a FEVER-like dataset COVIDFact of 4,086 claims concerning the COVID-19 pandemic, generated automatically with language models rather than employing human annotators. --- WORK IN PROGRESS ICLEF: In-Context Learning with Expert Feedback for Explainable Style Transfer Arkadiy Saakyan and Smaranda Muresan - Introduce a dataset for explainable style transfer and a method to learn from scarce expert feedback using in-context learning. --- RESEARCH EXPERIENCE Columbia Computer Science Department May 2020 - Present Graduate Research Assistant | Advisor: Prof. Smaranda Muresan Columbia Computer Science Department & Columbia Law School Sept 2019 - May 2020 Research Assistant | Advisors: Prof. Daniel Bauer, Dr. Frank Giaoui Columbia Business School Sept 2018 - May 2019 Research Assistant | Advisor: Prof. Tania Babina --- INDUSTRY EXPERIENCE Amazon Web Services, New York, NY May 2023 - Aug 2023 Applied Scientist Intern Data Analytics Lab, UBS Global Banking, New York, NY July 2021 - Aug 2022 Data Scientist Data Analytics Lab, UBS Global Banking, New York, NY Summer 2019, Summer 2020 Data Science Intern --- HONORS & AWARDS Columbia Data Science Institute Best Student Project Awarded for EMNLP 2021 paper Don’t Go Far Off: An Empirical Study on Neural Poetry Translation ACL 2021 Diversity & Inclusion Award Summer Research Grant Grant to perform research during the summer ($3,500) Columbia Work Exemption Program Grant to perform research during academic year ($3,600) Columbia Undergraduate Scholars Program John Jay National Scholarship "This award program provides financial aid and special programming to enhance the academic and extracurricular experiences of outstanding students." --- SERVICE Reviewer at - EACL 2023 - ACL 2023 - EMNLP 2023 --- ORGANIZATIONAL MEMBERSHIPS ACL (Association for Computational Linguistics) Summer 2021 - Present Queer in AI Summer 2021 - Present -- SKILLS Programming Python (including huggningface, fairseq, pytorch, sklearn, matplotlib, pandas, numpy, networkx, Flask), SQL, Java, C/C++, Prolog Tools Git, Amazon Mechanical Turk, Microsoft Excel, LaTeX Languages Armenian (heritage speaker), Russian (native), Italian (intermediate), French (basic), Brazilian Portuguese (basic) --- RELEVANT COURSEWORK NLP Multilingual Language Technologies and Language Diversity, Semantic Representations in NLP, Natural Language Processing, Semantic and Declarative Technologies ML Machine Learning, Geometric Data Analysis, Artificial Intelligence Math Statistical Inference, Probability Theory, Graph Theory, Structure and Dynamics of Complex Networks, Linear Algebra, Accelerated Multivariable Calculus Logic Mathematical Logic II, Mathematical Logic I, Modal Logic, Theory of Computing CS Advanced Algorithms, Introduction to Databases, Advanced Programming, Fundamentals of Computer Systems, Data Structures, Agile Project Management --- IMMIGRATION Approved EB-2 National Interest Waiver I-140 petition