#AudioML - Twitter Hashtag

7 days ago

Pretrained audio encoders encode spatial info unevenly. SARL benchmark reveals systematic biases in azimuth, elevation, distance, and room factors. Open-source evaluation framework released. #AudioML #SpatialAudio https://t.co/smNsn5nXkO

0

9

Orchestral Pit @OrchestralPit

7 days ago

Universal audio encoder outperforms domain specialists via hybrid SSL and supervised distillation. Scales to 1B params. Drop-in for audio LLMs. #AudioML #AudioRepresentation https://t.co/ipLNQsvTG0

0

10

Anthony Rodrigues @anthonyrod443

about 1 month ago

This isn't deepfake detection. It's codec artifact detection. Reads the encoder fingerprint, not the synthesis. Müller et al. (2022): lab detectors under 1% EER hit 30%+ on real-world audio. A perfect score isn't a model. It's a flag for what's leaking. #AudioML

0

2

0

50

Mark Pors 🦖

@pors

3 months ago

Imagine a noisy group call where 3 people talk at once. This paper builds a model that can focus on a single speaker (using a short voice sample) and extract that voice. This cleanup results in better, faster audio transcription. Summary and full paper 👇 #AudioML #SpeechToText

pors's tweet photo. Imagine a noisy group call where 3 people talk at once.

This paper builds a model that can focus on a single speaker (using a short voice sample) and extract that voice.

This cleanup results in better, faster audio transcription.

Summary and full paper 👇

#AudioML #SpeechToText

1

2

0

87

Camille Noufi @CamilleNoufi

about 2 years ago

🔊Join us at the upcoming BISH Bash @ MathWorks! 📅 May 16 in Sunnyvale, CA/Virtual. 🤝Talks, networking, & bites! Topics: Spatial Audio Modeling 🎧 Simulating a Clarinet 🎷 Neural Network Tuning 🧠 Music for Hearing Aids 🔊 RSVP: https://t.co/j9MtHkB2CU #AudioTech #AudioML

1

3

0

733

Burooj Ghani @calvin_the_cat

over 2 years ago

Has anyone experimented with #mixup in embedding space for audio-ML tasks? Curious about its effectiveness. Does it enhance model generalisation, improve performance? How would it compare to standard mixup? Would love to hear any experiences/insights! #AudioML #MachineLearning

1

2

0

167

Stefan & the Agents @stefan_loops

about 3 years ago · San Francisco

who’s doing cool stuff in audio & music tech in music games at GDC? let’s chat! #audioml #audioai #GDC2023

0

2

0

124

Jess Holland @JessKHolland

over 3 years ago

Delivery day! Building my own microphone-amplifier circuits for experiments with #conditionmonitoring using audio. #audioML #prototyping #machinelearning @Water_UofExeter

JessKHolland's tweet photo. Delivery day! Building my own microphone-amplifier circuits for experiments with #conditionmonitoring using audio.
#audioML #prototyping #machinelearning @Water_UofExeter https://t.co/tNLOrXCY9I

1

2

0

Sarthak Yadav @yadav_sar

about 4 years ago

Latest work on learning audio features in a self-supervised setting with @neilzegh. Mel-scale initialized frontends perform worse than random initialization for SSL, and other differences! #audioml #deeplearning https://t.co/P1Pwak053p

0

3

1

0

Suraj Subramanian @subramen

about 4 years ago

This was so much fun to build, but even more fun to use! I love plugging in old #Motown songs just to listen to the bassists - damn they knew how to funk. psst check out the source code for an easter egg https://t.co/RI4SAVKBNj #audioML #karaoke #demuxr

Meta Open Source

@MetaOpenSource

about 4 years ago

With web app Demuxr, musicians can input an audio clip and isolate each instrument in order to create custom remixes they can practice with. 🎵 @subramen and @jtannady break down how #opensource models PyTorch and Demucs make this happen: https://t.co/10aJhhdJs5

MetaOpenSource's tweet photo. With web app Demuxr, musicians can input an audio clip and isolate each instrument in order to create custom remixes they can practice with. 🎵

@subramen and @jtannady break down how #opensource models PyTorch and Demucs make this happen: https://t.co/10aJhhdJs5 https://t.co/eDwFX5fLHi

2

7

1

4

0

2

1

0

Peroptyx

@peroptyx

about 4 years ago

Congratulations to Anthony, Emma and team, doing pioneering work in the field of biodiversity and audio ML. We are very proud of you Anthony! #audioml #machinelearning #artificialintelligence #thefutureofdataishuman

peroptyx's tweet photo. Congratulations to Anthony, Emma and team, doing pioneering work in the field of biodiversity and audio ML.

We are very proud of you Anthony!

#audioml #machinelearning #artificialintelligence #thefutureofdataishuman https://t.co/IlnS2bCFkW

0

6

8

0

OHI Studio

@ohigames

almost 5 years ago

Scan Band is a new #MachineLearning #AR tool that transforms objects around you into virtual instruments you can play with your body. An amazing collab with @Artiphon 🙏🎶 pushing @SnapAR #LensStudio to the maxxx 💪 #ML #ScanBand #audioML Try it here → https://t.co/iblHCZuYGJ