
Applied AI/ML, PhD @TampereUni, Previously @facebook (@meta), @AdobeResearch, @FreshworksInc, @Krutrim
Bengaluru
Joined July 2009
- Tweets1.2K
- Following721
- Followers560
- Likes1.3K
Sharath Adavanne retweeted
Good news for Indian tech companies who use Google Maps APIs.
Google announces "India specific pricing" for Maps APIs, with up to 70% lower prices on most APIs.
In my 5 years at Swiggy, I never saw Google Maps drop their prices. Maps API billing moved in only one direction - up.
It's great to see competition drive prices lower for consumers. Kudos to @MapmyIndia
and @Krutrim for shaking up this monopoly.
When companies compete, consumers benefit.
inc42.com/buzz/amid-competit…
Sharath Adavanne retweeted
🤯 Human-like TTS is now ~11x cheaper(no pun)
Text to Speech was an entirely new one in OpenAI DevDay, adding directly to their multi-modal stack
Sharing some samples - 6 voices, multiple languages, cheaper, streamable and very human like
🧵...
Sharath Adavanne retweeted
Bravo! The audio-visual dataset used @DCASE_Challenge task3, Sony-TAu Realistic Spatial Soundscapes 2023 (STARSS23), will be presented @NeurIPSConf
Paper: arxiv.org/abs/2306.09126
Dataset: zenodo.org/record/7880637
Great effort by Kazuki, TAU and Sony colleagues
#NeurIPS2023
Our work done at @RakutenRIT on making text-to-speech models expressive for automatic production of audiobooks just got accepted at #ICASSP2023
Thanks to our collaborators Xin-san and @forthshinji-san from @UTokyo_News_en
and my colleagues @fm_ang @kulashish @Obs25oct
``Improving Speech Prosody of Audiobook Text-to-Speech Synthesis with Acoustic and Textual Contexts. (arXiv:2211.02336v1 [cs.SD]),'' Detai Xin, Sharath Adavanne, Federico Ang, Ashish Kulkarni, Shinnosuke Takamichi, Hiroshi Saruwatari, ift.tt/nkmD71c
With just ~5-10 minutes of your speech recording this tech will make YOU speak in any language! 🤯🤯🤯
Content creators and celebrities can now cater to audience beyond their native language!
#voicecloning #multilingual
We're at @dubverse_ai proud to announce the development of NeoDub, a production-ready Cross-Lingual Voice Cloning (CLVC) model that allows a single AI model to generate speech in 13 different languages.
🧵
#ArtificialIntelligence #research #startup #LANGUAGE #MachineLearning
1
10
911
Checkout @narendramodi greeting Indians in all the supported languages on the occasion of Republic day.
dbvr.app/rdayli
Cc: @BJP4India @PMOIndia @GoI_MeitY @AshwiniVaishnaw
2
177
Sharath Adavanne retweeted
We have an open Research Engineer internship position at @dubverse_ai
The internship is focused on NMT and requires you to have a piece of knowledge of NLP, Transformers, and general AI modeling.
Details + how to apply: https://teapawt.org/t.co/BLPuQRtFnJ
#research #internship #NLP
We are hiring multiple audio and speech researchers for internships, junior and mid-level careers at @RakutenRIT. We are the research wing of @RakutenGroup and support their 70+ businesses. So you get to work on a variety of problems. Check links below for the job descriptions.
Sharath Adavanne retweeted
Hey everyone!
@Iamshennoying is teaching veena online for beginners.
Please get in touch, and RT :)
📢 New paper: A multi-source localization and tracking approach for dynamic sound scenes with varying numbers of sources.
'Differentiable Tracking-Based Training of Deep Learning Sound Source Localizers' with Archontis Politis and @TuomasVirt at WASPAA
arxiv.org/abs/2111.00030
We further formulate the popular tracking-based metrics of CLEAR-MOT as a differential objective function and optimise the multi source localizer directly on it.
1
The proposed training procedure results in large improvements in localization error, detection metrics, and tracking capabilities. Code available here: github.com/sharathadavanne/d…
Sharath Adavanne retweeted
This is how we do safe physical distancing in paradise, Innahura Maldives-style 🌴
Book your last-minute getaway with us and enjoy savings up to 25%: innahura.com☀️
📸 adavanne
#Innahura #InnahuraMaldives #HereLifeIsEasy #Maldives #MaldivesGetaway #VisitMaldives










