Vlad Paunescu

@ChessvisionAI @ChessvisionAI scan

I'm @ChessvisionAi a Twitter bot to help you analyze chess diagrams. To trigger me, reply to any tweet with a chess diagram and mention me with the "scan" keyword and I'll reply back with my analysis. You can also use keywords "scan white" or "scan black" to hint whose turn is it

Having fun with the xchess.ai crew at Bucharest, Romania, cooking up the next big thing in chess

Extending #StyleGAN3's GUI, part 1: added the rest of the affine transformations: scale, shear, and mirror in both horizontal and vertical axes. Here I showcase with FFHQU-256, config-r, so that I have a higher FPS (cc @hexorcismos). Updated GUI found at: github.com/PDillis/styleg…

Tesla has 💎 🙌

@jk_rowling I still only own 0.25 Bitcoins btw

@raoult_didier Hello! Have you tried using Azithromycin alone because of Hydroxychloroquine side effects?

Notre étude porte sur 80 patients, sans groupe contrôle car nous proposons notre protocole à tous les patients ne présentant pas de contre-indication. C'est ce que nous dicte le serment d'Hippocrate que nous avons prêté. mediterranee-infection.com/epidemie-a-cor…

BREAKING: New video appears to show missile hitting Flight 752 before it crashed near Tehran, killing 176 people

We've trained an AI system to solve the Rubik's Cube with a human-like robot hand. This is an unprecedented level of dexterity for a robot, and is hard even for humans to do. The system trains in an imperfect simulation and quickly adapts to reality: openai.com/blog/solving-r…

15-year-old Alexandra Măceșanu's call in vain to the emergency services has saddened and shocked all of Europe. More effective policies against femicide and violence against women must be at the centre of all our political and institutional efforts.

@RevolutApp Investments?

Guess what we're cooking up! 7 days to go 🍤

Reconstructing scene depth from video with moving subjects can be very challenging. Now there’s a #DeepLearning approach that generates accurate 3D depth maps from ordinary video, even when both camera and subjects are in motion. Learn more at → goo.gle/2HxYqCm

@UCBerkeley DeepDrive and DiDi are co-hosting two challenges in #CVPR2019 Workshop on Autonomous Driving. The challenges will focus on domain adaptation of object detection and tracking based on datasets from BDD and DiDi. Register now! bair.berkeley.edu/blog/2019/03/2…

Stroke of Genius: GauGAN Turns Doodles into Stunning, Photorealistic Landscapes blogs.nvidia.com/blog/2019/03/1…

Learning Correspondence from the Cycle-Consistency of Time. New work from CMU/Berkeley shows how cycle-consistency can help a large number of diverse correspondence problems in vision! arxiv.org/abs/1903.07593 #computervision #cvpr2019

Our #CVPR2019 solves video object segmentation end-to-end with a NN: a ConvLSTM both discovers objects in the video frames dimensions, and tracks them across time. Joint effort by @UOCuniversitat @BSC_CNS & @la_UPC. #DLUPC Project page: imatge-upc.github.io/rvos/

It turns out that only a few parameters need to be trained to fine-tune huge text transformer models. Our latest paper is on arXiv; work @GoogleAI Zürich and Kirkland. bit.ly/2DXCgb7 #GoogleZurich #GoogleKirkland