Ana Brassard

she/her

Sendai, Miyagi, Japan

Technical Staff I at the Riken AIP Natural Language Understanding Team

PhD student at the Tohoku NLP Group, Tohoku University


Email me: brassard.ana@gmail.com

Download my CV

See my code: github.com/a-brassard

I work in Natural Language Processing with a particular interest in commonsense reasoning. My PhD topic is thinking of ways to diagnose current systems' reasoning capabilities beyond what a single performance score can tell us.

(2022) I made a crowdsourced semi-structured explanation dataset for a commonsense reasoning benchmark which can be used to train explanation-generating models or to compare with existing knowledge bases. Available here!

(2024) I also made a dataset of explanations with fine-grained quality ratings to aid in the development of automatic explanation evaluation metrics, as well as analyzing the behavior of large language models as evaluators. Available here!

In the upcoming future, I hope to work on developing more robust, transparent, and reliable reason-capable AI systems, with increasingly sophisticated evaluation methods to match.

Education

PhD, Information Sciences (in progress)
April 2021 - *March 2025 (planned)
Graduate School of Information Sciences, Tohoku University
Master of Science in Computing
October 2016 - September 2018
Faculty of Electrical Engineering and Computing, University of Zagreb
Bachelor of Science in Computing
October 2013 - September 2016
Faculty of Electrical Engineering and Computing, University of Zagreb

Publications

generated by bibbase.org
  2025 (2)
Quantifying the Influence of Evaluation Aspects on Long-Form Response Assessment. Kamoda, G.; Asai, A.; Brassard, A.; and Sakaguchi, K. In Rambow, O.; Wanner, L.; Apidianaki, M.; Al-Khalifa, H.; Eugenio, B. D.; and Schockaert, S., editor(s), Proceedings of the 31st International Conference on Computational Linguistics, COLING 2025, Abu Dhabi, UAE, January 19-24, 2025, pages 8787–8808, 2025. Association for Computational Linguistics
Quantifying the Influence of Evaluation Aspects on Long-Form Response Assessment [link]Paper   link   bibtex  
Evaluating Model Alignment with Human Perception: A Study on Shitsukan in LLMs and LVLMs. Shiono, D.; Brassard, A.; Ishizuki, Y.; and Suzuki, J. In Rambow, O.; Wanner, L.; Apidianaki, M.; Al-Khalifa, H.; Eugenio, B. D.; and Schockaert, S., editor(s), Proceedings of the 31st International Conference on Computational Linguistics, COLING 2025, Abu Dhabi, UAE, January 19-24, 2025, pages 11428–11444, 2025. Association for Computational Linguistics
Evaluating Model Alignment with Human Perception: A Study on Shitsukan in LLMs and LVLMs [link]Paper   link   bibtex  
  2024 (2)
ACORN: Aspect-wise Commonsense Reasoning Explanation Evaluation. Brassard, A.; Heinzerling, B.; Kudo, K.; Sakaguchi, K.; and Inui, K. CoRR, abs/2405.04818. 2024.
ACORN: Aspect-wise Commonsense Reasoning Explanation Evaluation [link]Paper   doi   link   bibtex  
Think-to-Talk or Talk-to-Think? When LLMs Come Up with an Answer in Multi-Step Reasoning. Kudo, K.; Aoki, Y.; Kuribayashi, T.; Sone, S.; Taniguchi, M.; Brassard, A.; Sakaguchi, K.; and Inui, K. CoRR, abs/2412.01113. 2024.
Think-to-Talk or Talk-to-Think? When LLMs Come Up with an Answer in Multi-Step Reasoning [link]Paper   doi   link   bibtex  
  2023 (6)
Empirical Investigation of Neural Symbolic Reasoning Strategies. Aoki, Y.; Kudo, K.; Kuribayashi, T.; Brassard, A.; Yoshikawa, M.; Sakaguchi, K.; and Inui, K. In Vlachos, A.; and Augenstein, I., editor(s), Findings of the Association for Computational Linguistics: EACL 2023, Dubrovnik, Croatia, May 2-6, 2023, pages 1124–1132, 2023. Association for Computational Linguistics
Empirical Investigation of Neural Symbolic Reasoning Strategies [link]Paper   doi   link   bibtex  
Do Deep Neural Networks Capture Compositionality in Arithmetic Reasoning?. Kudo, K.; Aoki, Y.; Kuribayashi, T.; Brassard, A.; Yoshikawa, M.; Sakaguchi, K.; and Inui, K. In Vlachos, A.; and Augenstein, I., editor(s), Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2023, Dubrovnik, Croatia, May 2-6, 2023, pages 1343–1354, 2023. Association for Computational Linguistics
Do Deep Neural Networks Capture Compositionality in Arithmetic Reasoning? [link]Paper   doi   link   bibtex  
Prompting for explanations improves Adversarial NLI. Is this true? Yes it is true because it weakens superficial cues. Kavumba, P.; Brassard, A.; Heinzerling, B.; and Inui, K. In Vlachos, A.; and Augenstein, I., editor(s), Findings of the Association for Computational Linguistics: EACL 2023, Dubrovnik, Croatia, May 2-6, 2023, pages 2120–2135, 2023. Association for Computational Linguistics
Prompting for explanations improves Adversarial NLI. Is this true? Yes it is true because it weakens superficial cues [link]Paper   doi   link   bibtex  
Do Deep Neural Networks Capture Compositionality in Arithmetic Reasoning?. Kudo, K.; Aoki, Y.; Kuribayashi, T.; Brassard, A.; Yoshikawa, M.; Sakaguchi, K.; and Inui, K. CoRR, abs/2302.07866. 2023.
Do Deep Neural Networks Capture Compositionality in Arithmetic Reasoning? [link]Paper   doi   link   bibtex  
Empirical Investigation of Neural Symbolic Reasoning Strategies. Aoki, Y.; Kudo, K.; Kuribayashi, T.; Brassard, A.; Yoshikawa, M.; Sakaguchi, K.; and Inui, K. CoRR, abs/2302.08148. 2023.
Empirical Investigation of Neural Symbolic Reasoning Strategies [link]Paper   doi   link   bibtex  
Chat Translation Error Detection for Assisting Cross-lingual Communications. Li, Y.; Suzuki, J.; Morishita, M.; Abe, K.; Tokuhisa, R.; Brassard, A.; and Inui, K. CoRR, abs/2308.01044. 2023.
Chat Translation Error Detection for Assisting Cross-lingual Communications [link]Paper   doi   link   bibtex  
  2022 (5)
Context Limitations Make Neural Language Models More Human-Like. Kuribayashi, T.; Oseki, Y.; Brassard, A.; and Inui, K. In Goldberg, Y.; Kozareva, Z.; and Zhang, Y., editor(s), Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, EMNLP 2022, Abu Dhabi, United Arab Emirates, December 7-11, 2022, pages 10421–10436, 2022. Association for Computational Linguistics
Context Limitations Make Neural Language Models More Human-Like [link]Paper   doi   link   bibtex  
Chat Translation Error Detection for Assisting Cross-lingual Communications. Li, Y.; Suzuki, J.; Morishita, M.; Abe, K.; Tokuhisa, R.; Brassard, A.; and Inui, K. In Deutsch, D.; Udomcharoenchaikit, C.; Opitz, J.; Gao, Y.; Fomicheva, M.; and Eger, S., editor(s), Proceedings of the 3rd Workshop on Evaluation and Comparison of NLP Systems, Eval4NLP 2022, Online, November 20, 2022, pages 88–95, 2022. Association for Computational Linguistics
Chat Translation Error Detection for Assisting Cross-lingual Communications [link]Paper   doi   link   bibtex  
COPA-SSE: Semi-structured Explanations for Commonsense Reasoning. Brassard, A.; Heinzerling, B.; Kavumba, P.; and Inui, K. In Calzolari, N.; Béchet, F.; Blache, P.; Choukri, K.; Cieri, C.; Declerck, T.; Goggi, S.; Isahara, H.; Maegaard, B.; Mariani, J.; Mazo, H.; Odijk, J.; and Piperidis, S., editor(s), Proceedings of the Thirteenth Language Resources and Evaluation Conference, LREC 2022, Marseille, France, 20-25 June 2022, pages 3994–4000, 2022. European Language Resources Association
COPA-SSE: Semi-structured Explanations for Commonsense Reasoning [link]Paper   link   bibtex   1 download  
COPA-SSE: Semi-structured Explanations for Commonsense Reasoning. Brassard, A.; Heinzerling, B.; Kavumba, P.; and Inui, K. CoRR, abs/2201.06777. 2022.
COPA-SSE: Semi-structured Explanations for Commonsense Reasoning [link]Paper   link   bibtex   1 download  
Context Limitations Make Neural Language Models More Human-Like. Kuribayashi, T.; Oseki, Y.; Brassard, A.; and Inui, K. CoRR, abs/2205.11463. 2022.
Context Limitations Make Neural Language Models More Human-Like [link]Paper   doi   link   bibtex  
  2021 (2)
Learning to Learn to be Right for the Right Reasons. Kavumba, P.; Heinzerling, B.; Brassard, A.; and Inui, K. In Toutanova, K.; Rumshisky, A.; Zettlemoyer, L.; Hakkani-Tür, D.; Beltagy, I.; Bethard, S.; Cotterell, R.; Chakraborty, T.; and Zhou, Y., editor(s), Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT 2021, Online, June 6-11, 2021, pages 3890–3898, 2021. Association for Computational Linguistics
Learning to Learn to be Right for the Right Reasons [link]Paper   doi   link   bibtex   1 download  
Learning to Learn to be Right for the Right Reasons. Kavumba, P.; Heinzerling, B.; Brassard, A.; and Inui, K. CoRR, abs/2104.11514. 2021.
Learning to Learn to be Right for the Right Reasons [link]Paper   link   bibtex   1 download  
  2019 (2)
Diamonds in the Rough: Generating Fluent Sentences from Early-Stage Drafts for Academic Writing Assistance. Ito, T.; Kuribayashi, T.; Kobayashi, H.; Brassard, A.; Hagiwara, M.; Suzuki, J.; and Inui, K. In van Deemter, K.; Lin, C.; and Takamura, H., editor(s), Proceedings of the 12th International Conference on Natural Language Generation, INLG 2019, Tokyo, Japan, October 29 - November 1, 2019, pages 40–53, 2019. Association for Computational Linguistics
Diamonds in the Rough: Generating Fluent Sentences from Early-Stage Drafts for Academic Writing Assistance [link]Paper   doi   link   bibtex   1 download  
Diamonds in the Rough: Generating Fluent Sentences from Early-Stage Drafts for Academic Writing Assistance. Ito, T.; Kuribayashi, T.; Kobayashi, H.; Brassard, A.; Hagiwara, M.; Suzuki, J.; and Inui, K. CoRR, abs/1910.09180. 2019.
Diamonds in the Rough: Generating Fluent Sentences from Early-Stage Drafts for Academic Writing Assistance [link]Paper   link   bibtex   1 download  
  2018 (1)
TakeLab at SemEval-2018 Task12: Argument Reasoning Comprehension with Skip-Thought Vectors. Brassard, A.; Kuculo, T.; Boltuzic, F.; and Snajder, J. In Apidianaki, M.; Mohammad, S. M.; May, J.; Shutova, E.; Bethard, S.; and Carpuat, M., editor(s), Proceedings of The 12th International Workshop on Semantic Evaluation, SemEval@NAACL-HLT 2018, New Orleans, Louisiana, USA, June 5-6, 2018, pages 1133–1136, 2018. Association for Computational Linguistics
TakeLab at SemEval-2018 Task12: Argument Reasoning Comprehension with Skip-Thought Vectors [link]Paper   doi   link   bibtex