AI bot which hallucinates colors into B&W image. I was built using deep learning and trained on millions of images by @vivek_kumar. #DeepLearning#AI#AIBot#DL
It's so awesome to see the impact of the computational audio capabilities we developed featured in
@madebygoogle π π π
Congrats to John Hershey, @ScottTWisdom, @PGetreuer & everyone who contributed for pioneering new computational audio capabilities in Pixel8 #MadeByGoogle
Check out the 4 new Google Photos features coming first to Pixel 8 and 8 Pro β
Whether itβs noise from wind, traffic, or barking dogs, Audio Magic Eraser in Google Photos reduces distracting sounds in your video in just a few taps! πͺ
It's so awesome to see the impact of the computational audio capabilities we developed featured in
@madebygoogle π π π
Congrats to John Hershey, @ScottTWisdom, @PGetreuer & everyone who contributed for pioneering new computational audio capabilities in Pixel8 #MadeByGoogle
Our work to bring the power of LLMs into non-linguistic modalities.
We tokenize images using frozen LLM translating them into lexical words which carry both rich semantics and fine details. This tokenization enables a diverse set of image understanding and generation tasks.
Fascinating research by Google reveals the power of Language Models (LLMs) like PaLM or GPT in tackling visual tasks using in-context learning. This novel method enables LLMs to perform image generation tasks without requiring any parameter updates. #palm#GPT4#LLMs
*How to inspire world-class researchers?*
Creating the AI research team (dolby.io/audio-research) @Dolby was a crash course in leading and inspiring experts more accomplished than me.
Sharing some learnings
π§΅
The Slow Death of Expertise: Murray talks about how advances in machine learning have made cutting-edge research less reliant on domain experts in
@azeem podcast "Grading AI: The Hits and Misses" ~
exponentialview.co/grading-aβ¦
Leading researchers - make sure they have agency unshackled from immediate business needs. Enable them to pursue their intrinsic motivations: autonomy, mastery, and purpose. Innovation requires deep thought and reflection.
h/t (Drive by @DanielPink)
Excited about pushing the state of art in audio & speech. We are hiring in our San Francisco & Barcelona #AI teams.
Join the AI research team in a company creating spectacular audio & video experiences.
linkedin.com/jobs/view/26806β¦@Dolby
πthis. All startups are hard & need long term commitment. Working on "problem worth solving" strengthens the will to persevere. It also attracts passionate talent - increasing the probability of success
@sama wrote a great blogpost about the paradox blog.samaltman.com/hard-starβ¦
Matthew Claudel @MIT showed that spatial relations mater more than departmental & institutions.
Researchers located in the same workspace are > 3x likely to collaborate when compared to those who are 400m apart. Collaboration further halves at 800m.
pubmed.ncbi.nlm.nih.gov/2864β¦
As the world struggles together, we have an opportunity for introspection and to pull together.
Vivek has a piece today about the divide and bonding that Covid presents.
#StrangeNewNormalstrangenewnormal.com/2020/05β¦
Whoohoo it's amazing to share what the team has been working on.
A Platform to make your content sound great with a single API: dolby.iox.com/i/status/1260223271703β¦
Yay! At long last, excited to be able to announce what I've been working on for the past year. Our new developer platform and APIs are live. Use them to build apps with interactive communications or enhanced sound in your media:
news.dolby.com/en-WW/188636-β¦
Thanks for sharing this and being frank about the hardships startups go through.
"Thatβs the hard thing about hard thingsβthere is no formula for dealing with them.β - @bhorowitzx.com/alexjc/status/12574075β¦
3/ However, when you're in the trenches, everything can feel like a critical moment β with the full weight of potential failure behind them.
ππππππ¦ π πππ ππ‘ π€ππ πππ π¦, ππ’π‘ ππ πππ ππ£ππ π πππ ππ‘ π€ππ’ππ ππ π‘βππ βπππ.
Proud of our AI teams contributing to 3 papers at #ICASSP2020
* Source Coding of Audio Signals with a Generative Model
* Audio Codec Enhancement with Generative Adversarial Networks
* An Empirical Study of Conv-Tasnet
ieeexplore.ieee.org/xpl/conhβ¦#IEEEICASSP2020@Dolby
1/4
Blown away by the site for the #ICLR2020 virtual conference. Impressive interactive visualization to find similar papers. Thanks for putting it all together at a really aggressive timeline @iclr_conficlr.cc/virtual/paper_vis.htβ¦
Breaks my heart. In January 2020 we learned about the new coronavirus and didn't pay attention to the consequences.
We had time to prepare and we squandered it.