🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

This is a random trash can in Bolzano, Italy, on which “thank you” is written in a number of languages… but they represent the beautiful linguistic landscape of Bolzano—German, Italian, Ladin, English, Arabic, Chinese, Albanian, Romanian, French, Ukrainian, Urdu (Punjabi?).

My student Chihiro Taguchi carolin T will also be around for the workshops! Talk to him about endangered languages, Kichwa (Ecuadorian Quechua), universal speech recognition, etc.

I’ll be at #NAACL2024 workshops next week (from June 18)! Please contact me if you’re interested in meeting up to talk about NLP for endangered languages, Universal Dependencies, NLP’s interface with linguistic theories, or anything else! (credit: template from David Chiang)

Tomorrow I’ll present my (previously published) work to create a new ASR dataset and model for Kichwa, an indigenous language of Ecuador, at AmericasNLP NAACL HLT 2025 from 2:30pm to 3:30pm! If you are still around at the workshops, come and check out my poster 😊

Today at ACL 2024 I’ll present my poster “Language Complexity and Speech Recognition Accuracy: Orthographic Complexity Hurts, Phonological Complexity Doesn't” at Poster Session D from 10:30am! Please stop by :) ArXiv: arxiv.org/abs/2406.09202

Missed my poster? Don’t worry! I’ll be giving an oral presentation in the session “Speech recognition, text-to-speech and spoken language understanding” in World Ballroom C from 10:30am! (my talk will start at 11am).

Congratulations to Chihiro Taguchi ctaguchi for winning an SAC/Outstanding Paper Award at #ACL2024 for his paper, "Language complexity and speech recognition accuracy: orthographic complexity hurts, phonological complexity doesn't"!

The paper studies the impact of various linguistic factors on automatic speech recognition accuracy. Orthographic complexity makes ASR more difficult, while phonological complexity doesn't. arxiv.org/abs/2406.09202

I was awarded an Outstanding Paper Award and an Area Chair’s Award at #ACL2024! I’m thrilled to win them for the research question I proposed myself and worked on with my supervisor David Chiang.

🚀次回のNLPコロキウムのお知らせ 講演者：田口智大 ctaguchi (@notredame) 日時：9/11 (水) 12:00- JST 人間👶にとって学びづらい言語は音声認識モデル🤖にとっても学びづらいのでしょうか。文字体系と音韻に着目した調査結果のご紹介です。[ACL'24 Outstanding Paper] nlp-colloquium-jp.github.io/schedule/2024-…

Congratulations to ctaguchi for winning a Lucy Family Institute for Data & Society Societal Impact Award for his work on creating language technologies for the Kichwa language community in Ecuador! Notre Dame CSE youtube.com/watch?v=HjsNOE…

Our paper has been accepted at #ICLR2025🎉 SoftMatcha finds query patterns that match not only exactly, e.g., grep, but also semantically similar ones levearaging word embeddings. Moreover, our soft yet efficient matcher can search billion-word corpora in 0.1 sec on CPUs🚀

バカでかコーパスを超高速に検索できる柔らかい grep/KWIC を作る話が #ICLR2025 に採択されました。 NLP + 理論計算機科学 + 言語学 のプレイヤーの合作です。 異なる分野のプロが集まる魅力をひしひし感じましたし、信じられないスピード感での議論・アルゴリズム改善・論文生成が超愉しかったです。

Web ツールも公開しています： softmatcha.github.io ノートルダムの田口さんに参加いただいたことにより、英語・日本語に次いで、ラテン語（？！）でも検索できるようになりました。西洋古典にご興味のあるみなさん、ぜひ遊んでみてください。 📑はこちら： openreview.net/forum?id=Q6PAn…

Megagon Labs has welcomed ctaguchi, PhD student at the University of Notre Dame, to our office for his winter internship. 🎉 He is conducting research on Retrieval-augmented generation (#RAG) and long-context language models (#LCLMs), mentored by Seiji Maekawa and @Nikita_Bhutani #AI #LLMs #NLProc

Midwest Speech and Language Days will be held Apr 15-16 at University of Notre Dame! Abstract submissions are due Mar 20, and registration deadline is Mar 27. Financial assistance for students (lodging, poster printing) is available. nlp.nd.edu/msld25

I completed my 3-month internship at Megagon Labs last week. Massive thanks to Seiji Maekawa and Nikita Bhutani I learned so much, and it was a very productive learning experience. I hope to see the team members again soon! (also hopefully the paper is coming out soon 🚀)