Applied Scientist | PhD | Working on multimodal healthcare AI. Works on "potentially" cool projects over weekends.
Remote
Joined July 2017
- Tweets212
- Following409
- Followers120
- Likes1.1K
Pinned Tweet
Made something cool using #LLM and #diffusion models. Guess the #YouTube video given hints created by #generativeAI. The hints are based on the context of the title of the "target video". More details 👇
#mkbhd #Mrwhosetheboss
Sanket Thakur retweeted
New Anthropic research: A global workspace in language models.
Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.
We found a strikingly similar divide inside Claude.
Did you know that "subscribe" button on youtube glows, when the word is mentioned in the video for that channel.
#YouTube #AI #subscribe
Things are going crazy with people making songs of recipes :
suno.com/song/4a77dea7-19f3-…
#GenerativeAI #spaghetti #suno
Sanket Thakur retweeted
Check out our Short-Term Object #Interaction #Anticipation method, 🌟NAOGAT🌟, presented at the @wacv_official #WACV2024 #egocentricvision @ego4_d
Joint work with @sanketsans97 @pmorerio @vmurino @alessiodelbue @PAVIS_IIT @IITalk
Project page: sanketsans.github.io/leverag…
Sanket Thakur retweeted
You know how image generation went from blurry 32x32 texture patches to high-resolution images that are difficult to distinguish from real in roughly a snap of a finger? The same is now happening along the time axis (extending to video) and the repercussions boggle the mind just a bit. Every human becomes a director of multi-modal dreams, like the architect in Inception.
Coming back to Earth for a second, image/video generation is a perfect match for data-hungry neural nets because data is plentiful, and the pixels of each image or video are a huge source of bits (soft constraints) on the parameters of the network. When you're training giant neural nets in supervision-rich settings, your train loss = validation loss, and life is so good.
My favorite place to keep an eye on the AI video space unfold atm is probably reddit.com/r/aivideo/ , or the individual Discords.
Introducing Pika 1.0, the idea-to-video platform that brings your creativity to life.
Create and edit your videos with AI.
Rolling out to new users on web and discord, starting today. Sign up at pika.art
Thanks to everyone who came by to visit our work at #ICIP2023. We propose a guided-attention based approach to estimate next-active-object for egocentric videos.
Our work on Guided attention also won the #CVPR23 #ego4d challenge earlier.
sanketsans.github.io/guided-…
Work done in collaboration with @BeyanCigdem @pmorerio @vmurino @alessiodelbue at @PAVIS_IIT @IITalk
Absolutely delighted to have won the #CVPR23 @ego4_d STA challenge 🏆. A big thanks to all my advisors!
@BeyanCigdem @pmorerio @vmurino @alessiodelbue
Can you tell me what's the active object in the image ?
#egocentric #computervision
🎉Exciting news! @PAVIS_IIT research line won the @ego4_d Short-Term-Anticipation (STA) challenge! 🎖️ The team used guided-attention to enhance the motion and contextual information in egocentric video.
Congrats @sanketsans97 @BeyanCigdem @pmorerio @vmurino @alessiodelbue #CVPR23
Sanket Thakur retweeted
yay the ability to share ChatGPT conversations is now rolling out. I can share a few favorites.
E.g. GPT-4 is great at generating Anki flash cards, helping you to memorize any document. Example:
chat.openai.com/share/eef34f…
Easy to then import in Anki: apps.ankiweb.net
