
PhD student, working on speech AI with David Harwath @utsaltlab (@UTAustin) and David R. Mortensen @dmort27 (@LTIatCMU).
Joined August 2023
- Tweets125
- Following172
- Followers223
- Likes280
Kwanghee Choi retweeted
I've always believed that intelligence should come from simple, general building blocks that learn their own abstractions directly from raw data; in other words, the bitter lesson. I think we're finally seeing a real step towards AI that can learn from all data.
Language modeling on raw bytes is a microcosm of this ideal, and to me it's a central problem in AI: if we can't build LLMs that learn end-to-end from bytes, how will we handle the rest of the world's data that we understand even less?
Last year we introduced hierarchical networks (H-Nets) as a first step. Our new version finds up to 3-5x efficiency gains in controlled scaling law experiments, and at larger scale matches an open-source model trained on 11T tokens with ~10x less data and compute. The posts below walk through what we found and why we think it works.
Kwanghee Choi retweeted
๐ข New paper: RUBATO: Decoupled Clocks Allow Full-Duplex Models to Think as Fast as They Can
Current full-duplex models tie thinking to the conversation's pace. RUBATO decouples them, giving each stream its own clock, so thinking runs as fast as the hardware allows.
Kwanghee Choi retweeted
(1/7)๐ข New paper! Speech AI is getting better at hearing and speaking many languages. But what if the input mixes two languages, or is in one the model never learned? It guesses one language it knows and imposes that language's sounds on everything.
We find why, and fix it ๐งต
Kwanghee Choi retweeted
๐ค How well do agents learn from experience?
We study agents that amortize experience into reusable artifacts and introduce Agent Plasticity to measure how efficiently they improve on held-out environments.
Top performer โ most efficient learner.
Kwanghee Choi retweeted
Excited to share our new paper to be presented at SLT 2026 in Palermo! ๐ฎ๐น ๐
We show that phone segmentation and recognition can be solved jointly by manipulating SSL representations using less than a minute of labeled speech. (1/n)
3 papers submitted & accepted at #SLT2026 ๐ Great to see that phonological vector arithmetic is indeed effective across several new applications. Bunch of computational linguistics this time, see you in Palermo!
Kwanghee Choi retweeted
Foundation models are built in various stages, from pre-training to post-training and beyond. We still lack a principled understanding of how these stages interact.
Excited to co-organize โTransitioning from Pre-Training to Post-Trainingโ at NeurIPS 2026 in Sydney.
Submit your work by Aug 29!
announcing the NeurIPS 2026 workshop on *Transitioning from Pre-training to Post-training*!๐
come bring your theory, empirics, & methods to enable better foundations, mechanics, & evaluation for LLM training!
submit your work by 29th August!๐๏ธ
pretrain2posttrain.github.ioโฆ
๐งต
See you at San Diego on Sunday 14:00-15:30, Poster Session B! I'll be presenting the [b] = [d] - [t] + [p] paper ๐ค
4 papers submitted & accepted at ACL 2026! ๐ So grateful to work alongside & learn from amazing minds, pushing the boundaries of speech technologies, machine learning, and computational linguistics. See you in San Diego!
4
271
3 papers accepted at Interspeech 2026! ๐ Proud to see my role slowly transitioning from a mentee to a mentor. Huge thanks to my collaborators for letting me be a part of their journey!
Kwanghee Choi retweeted
What if you had nano-banana for audio?
AudioChat is a multi-modal LM that performs fine-grained understanding, generation, and editing of multi-source scenes
By diffusing continuous latents, it generates 48khz stereo edits with great input adherence:
wanchichen.github.io/audiochโฆ
Kwanghee Choi retweeted
โจNew paperโจ
We find script (e.g. Cyrillic, Latin) to be a linear direction in the activation space of Whisper, enabling transliteration at test-time by adding such script directions to the activations โ producing e.g. Cyrillic Japanese transcriptions.
4 papers submitted & accepted at ACL 2026! ๐ So grateful to work alongside & learn from amazing minds, pushing the boundaries of speech technologies, machine learning, and computational linguistics. See you in San Diego!
Kwanghee Choi retweeted
Shikhar Bharadwaj, Chin-Jou Li, Kwanghee Choi, Eunjung Yeo, William Chen, Shinji Watanabe, David R. Mortensen, "An Empirical Recipe for Universal Phone Recognition," arxiv.org/abs/2603.29042
Thanks a lot for the interest in our work! Here's the recording for people who missed the seminar: youtu.be/DtFYKvNo9IQ
๐๐๐ฅ๐-๐ฌ๐ฎ๐ฉ๐๐ซ๐ฏ๐ข๐ฌ๐๐ ๐๐ฉ๐๐๐๐ก ๐๐จ๐๐๐ฅ๐ฌ ๐๐ซ๐ ๐๐ก๐จ๐ง๐จ๐ฅ๐จ๐ ๐ข๐๐๐ฅ ๐๐๐๐ญ๐จ๐ซ ๐๐๐๐ก๐ข๐ง๐๐ฌ!
๐ฃ๏ธ Excited to be giving an invited talk this Thursday (March 19th, 3pm Amsterdam time)!
Huge thanks to @mariannedhk at Univ. of Amsterdam for the invite ๐
7
375
๐๐๐ฅ๐-๐ฌ๐ฎ๐ฉ๐๐ซ๐ฏ๐ข๐ฌ๐๐ ๐๐ฉ๐๐๐๐ก ๐๐จ๐๐๐ฅ๐ฌ ๐๐ซ๐ ๐๐ก๐จ๐ง๐จ๐ฅ๐จ๐ ๐ข๐๐๐ฅ ๐๐๐๐ญ๐จ๐ซ ๐๐๐๐ก๐ข๐ง๐๐ฌ!
๐ฃ๏ธ Excited to be giving an invited talk this Thursday (March 19th, 3pm Amsterdam time)!
Huge thanks to @mariannedhk at Univ. of Amsterdam for the invite ๐
๐งต Together, both papers take a step beyond the usual "what info do S3Ms encode" probing paradigm. We aim to answer how is that info actually encoded geometrically? Come see for yourself Thursday! ๐
Slides: docs.google.com/presentationโฆ
1
1
119















