JuanJo Nieto
@juanjo_3ns
I like doing handstands and tracking things - Applied Scientist @ Lace Lithography
ID: 404413711
https://github.com/juanjo3ns 03-11-2011 22:01:56
441 Tweet
170 Followers
1,1K Following
Today at 1pm CET we are hosting the last of our guest lectures UPC Telecos 📡 dseupc. Carlos Florensa (Carlos Florensa) from Covariant (formerly in Berkeley AI Research) will present "What supervision scales? Practical Learning through interaction." youtu.be/47vzPmOLdps
Last days for signing up to our 6th postgraduate course on Artificial Intelligence with Deep Learning UPC School Barcelona. 120 teaching hours in English covering core DL, vision, NLP, speech/audio & RL (50% hands on). From 17/2 to 14/7. Fee: 3.900 €. talent.upc.edu/ing/estudis/fo…
Seasonal Contrast: Unsupervised Pre-Training from Uncurated Remote Sensing Data (arxiv.org/abs/2103.16607) w/ Pau Rodríguez, Alexandre Lacoste, David Vazquez & Xavi Giró at @element_ai @servicenow A method for pretraining remote sensing representations from large-scale unlabeled data! 🧵👇
Sign up to the info sessions on June 1 & 2 of our program on Artificial Intelligence with Deep Learning UPC School Barcelona. 120 teaching hours in English covering core DL, vision, NLP, speech/audio & RL. Fee: 3.900 €. talent.upc.edu/ing/estudis/fo…
How to self-supervise learning in RL ? Juan José Nieto (JuanJo Nieto) maximizes mutual information between states and latents by applying Explore, Discover & Learn (EDL) in MineRL Project. MSc thesis UPC Telecos 📡 Barcelona, co-advised with Víctor Campos . youtu.be/lqHXQmAmrxs
Team lab lunch, Summer 2021 (COVID-19 limited edition). After 4 PhD defenses in 1.5 years, the new batch is ready. Looking forward to new adventures in sign language, RL, 3D reconstruction and even genomics, with Amanda Duarte, Laia, JuanJo Nieto, Pol et al.
- In #ICML2020, EDL showed how RL agents can learn skills without supervision by maximizing MI. - Only in a toy maze. Too easy. - Today at #CVPR2021 Embodied AI we scale it up to 3D pixels from MineRL Project and AI Habitat. imatge-upc.github.io/PixelEDL/ imatge-upc.github.io/PiCoEDL/
It was a pleasure working with you Roger Creus Castanyer Xavi Giró Víctor Campos Oscar Mañas 😁 I couldn't be happier with the work done and people seem to have liked it too 🎉 Looking forward to present this work at the Unsupervised RL workshop ICML Conference.
The program for the Deep Learning Barcelona Symposium 2021 is out. 3 keynotes (@joanbruna Elisenda Bou-Balust Joan Serrà), 29 talks, 47 posters & 2 mentorship sessions on December 22-23 at Campus Nord UPC. sites.google.com/view/deep-lear… Book your agendas, but registration is not open yet.
Today data scientists Dani Fojo (Dani Fojo) & Juan José Nieto (JuanJo Nieto) from Glovo taught a practical on deep reinforcement learning at UPC Telecos 📡 Barcelona. Our data science students (dseupc) learned how RL can be applies to solve logistic problems at scale.
Our teaching team of the Postgradute on Artificial Intelligence with Deep Learning UPC School kicked off the students projects for this Winter term.
In Deep Learning Barcelona Symposium, Antonio Bonafonte points to a problem of our PhD students: if they have a public grant, they cannot stop it for an internship in industry. I’ll add that public PhD grants are not even enough for a living. I invite our minister Gemma Geis i Carreras 🎗 to address this non-sense.
Dr. Adrià Salvador from Glovo Data Science (GlovoDSCORE) presents the second of our Careers Sessions at the Deep Learning Barcelona Symposium. Thanks for being the gold sponsor of DLBCN 2021.
Instructors dinner of our Postgraduate on Artificial Intelligence at Deep Learning at UPC School Barcelona. We are already preparing the 11th edition, online, starting on November 2022: talent.upc.edu/ing/estudis/fo…
