I’m excited to share one of two papers accepted to #Interspeech2025! @interspeech.bsky.social

“Spectrotemporal Modulation: Efficient & Interpretable Feature Representation for Classifying Speech, Music & Environmental Sounds”
📄 Paper: arxiv.org/abs/2505.23509
#NeuroInspiredML #AudioAI

🚀 **7B model tops MMAU!** Xiaomi used DeepSeek-R1's GRPO to boost Alibaba's Qwen2-Audio-7B accuracy to 64.5%, beating GPT-4o by 10%. 🎧🤖

#AI #AudioAI #ReinforcementLearning #MMAU #Xiaomi

aidisruption.ai/p/xiaomis-7b...

Equation: Waveforms×SSL → Audio IQ 🎧🧠
Detect events; triage noise; pretrain unlabeled audio→fine-tune; align w/ vision+text; verify labels/drift - GLCND.IO
Explore → https://glcnd.io/training-ai-to-understand-sound/
#AI #AudioAI

FCPE reaches 96.79% raw pitch accuracy on the MIR‑1K dataset and runs with a 0.0062 real‑time factor on an RTX 4090, enabling faster‑than‑real‑time processing. Read more: https://getnews.me/fcpe-model-offers-fast-accurate-pitch-estimation-for-audio/ #fcpe #audioai

Bí Ẩn Miền Tây: “Nổ” Giải Đặc Biệt Xổ Số – Âm Thanh AI hé lộ điều gì? #XổSố/hashtag/X%E1%BB%95S%E1%BB%91" class="hover:underline text-blue-600 dark:text-sky-400 no-card-link">#XổSố #MiềnTâyashtag/Mi%E1%BB%81nT%C3%A2y" class="hover:underline text-blue-600 dark:text-sky-400 no-card-link">#MiềnTây #AudioAI"/hashtag/AudioAI" class="hover:underline text-blue-600 dark:text-sky-400 no-card-link">#AudioAI #BíẨn"/hashtag/B%C3%AD%E1%BA%A8n" class="hover:underline text-blue-600 dark:text-sky-400 no-card-link">#BíẨn #GiảiĐắcBiệti%E1%BA%A3i%C4%90%E1%BA%AFcBi%E1%BB%87t" class="hover:underline text-blue-600 dark:text-sky-400 no-card-link">#GiảiĐắcBiệt #TinTức/hashtag/TinT%E1%BB%A9c" class="hover:underline text-blue-600 dark:text-sky-400 no-card-link">#TinTức

Bí Ẩn Miền Tây: "Nổ" Giải Đặc Biệt Xổ Số - Âm Thanh AI hé lộ điều gì? #XổSố #MiềnTây #AudioAI #BíẨn #GiảiĐắcBiệt #TinTức Phát hiện gây chấn động: Sự kiện xổ số miền Tây liên…

Futuri Launches Futuri AudioAI™, The World’s First 100% AI-Driven Local Content System https://podnews.net/press-release/futuri-audioai

idk #audioclassification #musicinformationretrieval #ml #audioai #audioanalysis #musictheory

Audio-Reasoner, a language model trained on the CoTA dataset of 1.2 million samples, improves benchmarks by +25.42% on MMAU‑mini and +14.57% on AIR‑Bench chat. Read more: https://getnews.me/audio-reasoner-boosts-reasoning-skills-in-large-audio-language-models/ #audioreasoner #audioai #multimodal

AudioAI in a blobby AI nutshell 💙
I see this one is going to improve in the next few years and maybe even hit the charts one day?

Listen to Joy to the World performed by Blob Opera. Then play four opera voices to create your own composition 🎼 artsandculture.google.com/experiment/b...

നിങ്ങൾക്കിഷ്ട്ടപ്പെട്ട വാർത്തകൾ ഓഡിയോ രൂപത്തിൽ കേൾക്കാം; എഐ ഫീച്ചര്‍ അവതരിപ്പിച്ച് ഗൂഗിൾ
keralatimeslive.news/2025/01/14/g...

#googleaifeature #audioai

Voxtral: Open Source AI Audio Model—Capabilities, Features, and How to Access.

See here - techchilli.com/artificial-i...

#Voxtral #AI2025 #AudioAI #MistralAI #OpenSource

🔊 Hiring: Audio AI Research Engineer! 📍San Francisco. Unleash your potential with cutting-edge AI technologies. Join our innovative team! #TechJobs #AudioAI #SanFranciscoJobs educativ.net/jobs/job/312...

Ex‑Google NotebookLM engineers launched Huxe, an audio‑first app that turns emails, calendars and web topics into spoken briefings. It's free on iOS and Android and raised $4.6 million. Read more: https://getnews.me/huxe-launches-audio-ai-app-for-news-briefs-and-deep-dive-podcasts/ #huxe #audioai

And just like that, it’s over. We dived in at #CES2025 and never stopped. Every meeting surprised us with great technologies and demonstrations. Thank you for the water, the lip balm, and the swag. We will be back next year and couldn’t recommend it more.
#CES2026 #AudioAI #audioinnovations

Audiobook Generator GUI that can clone your voice like 11Labs but hosted locally. Fun little project so I could listen to the books I've written and make sure everything sounded right.

github.com/Jeremy-Harpe...

#chatterbox
#audiobook
#author
#localllama
#audible
#audioAI

Bộ trưởng Giáo dục lên tiếng về Thông tư 29/2024: ‘Oan cho một số địa phương nếu nói dạy thêm, học thêm không hiệu quả’ – #AudioAI #QuốcHội #ChấtVấn #BộTrưởngGDĐT

1 giờ trước1 liên quanGốcBộ trưởng Bộ Giáo dục và Đào tạo Nguyễn Kim Sơn cho rằng nếu nói Thông tư 29/2024 về dạy thêm, học thêm không…

Current audio models struggle with human-like nuances: pitch, emotion, and accents. This is due to learning difficulty vs. text, reliance on synthetic data, and even intentional safeguards to prevent misuse. #AudioAI 2/5

Vertex AI: Google lancia Lyria, potenzia Veo e Chirp #AI #AudioAI #Chirp3 #CloudAI #CreazioneContenuti #GenAI #Google #GoogleAI #Imagen3 #ImmaginiAI #IntelligenzaArtificiale #Lyria #MediaGenerativi #Notizie #Novità� #TechNews #Tecnologia #Veo2 #VertexAI #VideoAI www.ceotech.it/vertex-ai-go...

SSEU‑Bench tests speech, scene and event understanding with independent, joint and energy‑aware settings; chain‑of‑thought prompts boost joint task performance. Submitted 16 Sep 2025. Read more: https://getnews.me/benchmark-targets-speech-scene-event-understanding-in-audio-ai/ #audioai #sseubench

Latent Bridge Models enable audio super‑resolution up to 192 kHz, setting state‑of‑the‑art scores for any‑to‑48 kHz across speech, music and environmental sounds. https://getnews.me/latent-bridge-models-boost-audio-super-resolution-quality/ #latentsuperresolution #audioai

ChatGPT Deep Research is useful but lacks audio summaries. Google's NotebookLM nails this with Audio Overviews, turning reports into podcasts. With its mobile apps arriving soon, NotebookLM could become the go-to for on-the-go research. #AI #AudioAI #NotebookLM