Published in International Journal on Natural Language Computing, 2024
Applies RAG-Fusion to Infineon product Q&A and manually evaluates answer accuracy, relevance, and comprehensiveness.
Recommended citation: Rackauckas, Z. (2024). "RAG-Fusion: A New Take on Retrieval-Augmented Generation." International Journal on Natural Language Computing, 13(1), 37–47.
Download Paper
Published in LLM4Eval @ SIGIR 2024, 2024
Proposes an automated Elo-based evaluation methodology for comparing retrieval-augmented generation systems.
Recommended citation: Rackauckas, Z., Câmara, A., & Zavrel, J. (2024). "Evaluating RAG-Fusion with RAGElo: An Automated Elo-Based Framework." LLM4Eval @ SIGIR 2024.
Download Paper
Published in arXiv, 2025
Preprint evaluating learning engagement and interaction outcomes in stylized LLM-driven dialogue agents.
Recommended citation: Rackauckas, Z., & Hirschberg, J. (2025). "Animating Language Practice: Engagement with Stylized Conversational Agents in Japanese Learning." arXiv.
Download Paper
Published in MAGMaR @ ACL 2025, 2025
Introduces a multimodal retrieval-augmented generation framework for spoken question answering without full transcription reliance.
Recommended citation: Rackauckas, Z., & Hirschberg, J. (2025). "VoxRAG: A Step Toward Transcription-Free RAG Systems in Spoken Question Answering." MAGMaR @ ACL 2025.
Download Paper
Published in JALTCALL Trends, 2025
Empirical analysis of learner engagement patterns in an LLM-driven character-based language learning application.
Recommended citation: Rackauckas, Z. (2025). "Learner interactions with Jouzu: A mobile application with conversational characters." JALTCALL Trends.
Download Paper
Published in IEEE UEMCON 2025, 2025
Evaluates expressive Japanese TTS systems in character-based conversational settings.
Recommended citation: Rackauckas, Z., & Hirschberg, J. (2025). "Comparative Evaluation of Expressive Japanese Character Text-to-Speech with VITS and Style-BERT-VITS2." IEEE UEMCON 2025.
Download Paper
Published in HCI+NLP @ EMNLP 2025, 2025
A dialogue system that integrates personal memory videos into emotionally grounded LLM-based question generation.
Recommended citation: Rackauckas, Z., Minematsu, N., & Hirschberg, J. (2025). "Re:Member: Emotional Question Generation from Personal Memories." HCI+NLP @ EMNLP 2025.
Download Paper
Published in International Conference on Human-Agent Interaction (HAI 2025), 2025
Extended abstract describing multimodal AI agents integrated into engineering lab instruction.
Recommended citation: Shigi, M., Rackauckas, Z., Akiyama, Y., & Minematsu, N. (2025). "LangInLab: Augmenting Engineering Lab Instruction with Vision- and Voice-Enabled AI Agents for Language Learning." HAI 2025.
Download Paper
Published in BEA @ ACL 2026, 2026
Surveys automated oral-presentation coaching systems across pronunciation, prosody, pacing, and content feedback.
Recommended citation: Wen Liang, Li Siyan, Zackary Rackauckas, and Julia Hirschberg. 2026. A Survey of Automated Presentation Coaching: Systems, Methods, and Open Challenges. In Proceedings of the 21st Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2026), pages 34–47, San Diego, California, USA. Association for Computational Linguistics.
Download Paper
Published in Interspeech 2026 Show & Tell, 2026
Demonstrates a speech-to-speech English practice system that estimates spoken fluency and adapts response vocabulary and speech rate to the learner’s level.
Recommended citation: Rackauckas, Z. (2026). AdaptLingo: A Speech-to-Speech English Practice System with Fluency-Adaptive Responses. Proc. Interspeech 2026, 2382-2383.
Download Paper
Published in Interspeech 2026 Show & Tell, 2026
Presents Jouzu, a speech-first Japanese practice system combining character-conditioned dialogue, expressive voice synthesis, persona comparison, and in-place vocabulary support.
Recommended citation: Rackauckas, Z. (2026). A Speech-First Character Interface for Stylized Japanese Dialogue Practice. Proc. Interspeech 2026, 2388-2389.
Download Paper
Published in arXiv, 2026
Benchmarks 13 speech models on phonetic error detection across children’s, acted, and synthetic speech.
Recommended citation: Shi, Z.*, Rackauckas, Z.*, Yip, J., Halpern, S., Campos, S., Brace, L., Yang, E., Feng, T., Shi, X., Narayanan, S., Satterfield, T., & Xu, Y. (2026). "PEDBench: Can Speech Models Detect Phonetic Errors?" arXiv. (*Equal contribution)
Download Paper