11 papers
BaLEEN: Biasing with Latent Encoded Entities for Context-Aware ASR
Chihiro Taguchi, Yotaro Kubo, Rujikorn Charakorn
Transcribing domain-specific entities and rare proper nouns remains a major challenge in automatic speech recognition (ASR). In this paper, we propose BaLEEN (Biasing with Latent E…
Pretrained self-supervised speech models can recognize unseen consonants
Chihiro Taguchi, Éric Le Ferrand, Hirosi Nakagawa +4
Modern pretrained self-supervised automatic speech recognition models are trained on large-scale audio data to encode speech into contextualized representations. However, their tra…
Automatic Speech Recognition for Documenting Endangered Languages: Case Study of Ikema Miyakoan
Chihiro Taguchi, Yukinori Takubo, David Chiang
Language endangerment poses a major challenge to linguistic diversity worldwide, and technological advances have opened new avenues for documentation and revitalization. Among thes…
Creating ConLangs to Probe the Metalinguistic Grammatical Knowledge of LLMs
Chihiro Taguchi, Richard Sproat
We present a system that uses LLMs as a tool in the development of Constructed Languages -- ConLangs, which we call IASC (Interactive Agentic System for ConLangs). The system is mo…
Building Tailored Speech Recognizers for Japanese Speaking Assessment
Yotaro Kubo, Richard Sproat, Chihiro Taguchi +1
This paper presents methods for building speech recognizers tailored for Japanese speaking assessment tasks. Specifically, we build a speech recognizer that outputs phonemic labels…
Languages Still Left Behind: Toward a Better Multilingual Machine Translation Benchmark
Chihiro Taguchi, Seng Mai, Keita Kurabe +4
Multilingual machine translation (MT) benchmarks play a central role in evaluating the capabilities of modern MT systems. Among them, the FLORES+ benchmark is widely used, offering…
Efficient Context Selection for Long-Context QA: No Tuning, No Iteration, Just Adaptive-
Chihiro Taguchi, Seiji Maekawa, Nikita Bhutani
Retrieval-augmented generation (RAG) and long-context language models (LCLMs) both address context limitations of LLMs in open-domain question answering (QA). However, optimal exte…
SoftMatcha: A Soft and Fast Pattern Matcher for Billion-Scale Corpus Searches
Hiroyuki Deguchi, Go Kamoda, Yusuke Matsushita +4
Researchers and practitioners in natural language processing and computational linguistics frequently observe and analyze the real language usage in large-scale corpora. For that p…
Language Complexity and Speech Recognition Accuracy: Orthographic Complexity Hurts, Phonological Complexity Doesn't
Chihiro Taguchi, David Chiang
We investigate what linguistic factors affect the performance of Automatic Speech Recognition (ASR) models. We hypothesize that orthographic and phonological complexities both degr…
Killkan: The Automatic Speech Recognition Dataset for Kichwa with Morphosyntactic Information
Chihiro Taguchi, Jefferson Saransig, Dayana Velásquez +1
This paper presents Killkan, the first dataset for automatic speech recognition (ASR) in the Kichwa language, an indigenous language of Ecuador. Kichwa is an extremely low-resource…
Universal Automatic Phonetic Transcription into the International Phonetic Alphabet
Chihiro Taguchi, Yusuke Sakai, Parisa Haghani +1
This paper presents a state-of-the-art model for transcribing speech in any language into the International Phonetic Alphabet (IPA). Transcription of spoken languages into IPA is a…