明らかに最近はまともな研究開発をするために要求されるエンジニアリングの水準が高くなっていると思う
いやまあレールの上に乗っかってるだけならそんなでもないんやけど
外れたりするとヤバくなる

Yuto Nishida, Makoto Morishita, Hiroyuki Deguchi, Hidetaka Kamigaito, Taro Watanabe
Long-Tail Crisis in Nearest Neighbor Language Models
https://arxiv.org/abs/2503.22426

🔔 The Emerging Bioacousticians’ Days come back !

🗓️ 24 to 26 June 2025

📍Vernon, France

👋🏼 See you soon !

#scientific #conference #bioacoustics #science

View of the Granlibakken Tahoe resort and Lake Tahoe

It’s official! 🎉
WASPAA 2025 will be held October 12-15 at the Granlibakken Tahoe resort, in Tahoe City, CA🏞️
Important dates:
Abstract deadline: April 23, 2025 (23:59 AOE)
Paper deadline: April 30, 2025 (23:59 AOE)
Acceptance: July 2, 2025
Camera-ready: July 16, 2025
More info: waspaa.com
Please RT🤗

RANF uses retrieved HRTFs from a dataset to augment neural field upsampling; it improves HRTF upsampling from sparse measurements and was part of a winning solution in a challenge.

An ADMM-based method for mel-spectrogram inversion jointly estimates magnitude and phase, improving upon cascaded methods through efficient variable updates, showing effectiveness on speech and foley sounds.

Diffusion Models for Audio Restoration: A review [Special Issue On Model-Based and Data-Driven Audio Signal Processing] With the development of audio playback devices and fast data transmission, the demand for high sound quality is rising for both entertainment and communications. In this quest for better sound quality...

Our article, "Diffusion Models for Audio Restoration: A Review," is now published in the IEEE Signal Processing Magazine!

A huge thank you to all co-authors Jean-Marie Lemercier, Julius Richter, Simon Welker, Eloi Moliner, and Vesa Välimäki for a great collaboration.

doi.org/10.1109/MSP....

We're sad to bid farewell to Kevin Wilkinghoff at the end of his 6-month stay as a visiting research scientist at MERL🍻
We learned a lot and had a great time🤗

In #SLT2024, my co-authors will present a Mamba-based decoder-only approach for ASR (MADEON) at P1-24-ASR and ESPnet-Codec at P3-23-SS07. Please enjoy😆

MADEON: arxiv.org/abs/2411.06968
ESPnet-Codec: arxiv.org/abs/2409.15897

Task-Aware Unified Source Separation Several attempts have been made to handle multiple source separation tasks such as speech enhancement, speech separation, sound event separation, music source separation (MSS), or cinematic audio sour...

New paper out with Kohei Saijo, J. Ebbers, F. Germain, G. Wichern: "Task-Aware Unified Source Separation" in which we strive to pave the way for a truly universal / unified / to-infinity-and-beyond source separation framework🦸
arxiv.org/abs/2410.23987