1. All left ideologies going back to classical liberalism itself are revolutionary and utopian at heart. The whole of the enlightenment IS the culture war writ large. Whether it is capitalism, Marxism, socialism, or fascism, they are part of the enlightenment drive for utopia.
This superb article denounces a false left promoting wokery and dissolving all basic pillars of human civilisation, while not falling into the dangers of simplistic right-wing reaction. Increasingly the real left, including Marxists, are breaking with such liberal extremism. https://t.co/5RLvHMqixM
— john milbank (@johnmilbank3) December 5, 2022
More from All
How can we use language supervision to learn better visual representations for robotics?
Introducing Voltron: Language-Driven Representation Learning for Robotics!
Paper: https://t.co/gIsRPtSjKz
Models: https://t.co/NOB3cpATYG
Evaluation: https://t.co/aOzQu95J8z
🧵👇(1 / 12)
Videos of humans performing everyday tasks (Something-Something-v2, Ego4D) offer a rich and diverse resource for learning representations for robotic manipulation.
Yet, an underused part of these datasets are the rich, natural language annotations accompanying each video. (2/12)
The Voltron framework offers a simple way to use language supervision to shape representation learning, building off of prior work in representations for robotics like MVP (https://t.co/Pb0mk9hb4i) and R3M (https://t.co/o2Fkc3fP0e).
The secret is *balance* (3/12)
Starting with a masked autoencoder over frames from these video clips, make a choice:
1) Condition on language and improve our ability to reconstruct the scene.
2) Generate language given the visual representation and improve our ability to describe what's happening. (4/12)
By trading off *conditioning* and *generation* we show that we can learn 1) better representations than prior methods, and 2) explicitly shape the balance of low and high-level features captured.
Why is the ability to shape this balance important? (5/12)
Introducing Voltron: Language-Driven Representation Learning for Robotics!
Paper: https://t.co/gIsRPtSjKz
Models: https://t.co/NOB3cpATYG
Evaluation: https://t.co/aOzQu95J8z
🧵👇(1 / 12)
Videos of humans performing everyday tasks (Something-Something-v2, Ego4D) offer a rich and diverse resource for learning representations for robotic manipulation.
Yet, an underused part of these datasets are the rich, natural language annotations accompanying each video. (2/12)
The Voltron framework offers a simple way to use language supervision to shape representation learning, building off of prior work in representations for robotics like MVP (https://t.co/Pb0mk9hb4i) and R3M (https://t.co/o2Fkc3fP0e).
The secret is *balance* (3/12)
Starting with a masked autoencoder over frames from these video clips, make a choice:
1) Condition on language and improve our ability to reconstruct the scene.
2) Generate language given the visual representation and improve our ability to describe what's happening. (4/12)
By trading off *conditioning* and *generation* we show that we can learn 1) better representations than prior methods, and 2) explicitly shape the balance of low and high-level features captured.
Why is the ability to shape this balance important? (5/12)
You May Also Like
Stan Lee’s fictional superheroes lived in the real New York. Here’s where they lived, and why. https://t.co/oV1IGGN8R6
Stan Lee, who died Monday at 95, was born in Manhattan and graduated from DeWitt Clinton High School in the Bronx. His pulp-fiction heroes have come to define much of popular culture in the early 21st century.
Tying Marvel’s stable of pulp-fiction heroes to a real place — New York — served a counterbalance to the sometimes gravity-challenged action and the improbability of the stories. That was just what Stan Lee wanted. https://t.co/rDosqzpP8i
The New York universe hooked readers. And the artists drew what they were familiar with, which made the Marvel universe authentic-looking, down to the water towers atop many of the buildings. https://t.co/rDosqzpP8i
The Avengers Mansion was a Beaux-Arts palace. Fans know it as 890 Fifth Avenue. The Frick Collection, which now occupies the place, uses the address of the front door: 1 East 70th Street.
Stan Lee, who died Monday at 95, was born in Manhattan and graduated from DeWitt Clinton High School in the Bronx. His pulp-fiction heroes have come to define much of popular culture in the early 21st century.
Tying Marvel’s stable of pulp-fiction heroes to a real place — New York — served a counterbalance to the sometimes gravity-challenged action and the improbability of the stories. That was just what Stan Lee wanted. https://t.co/rDosqzpP8i
The New York universe hooked readers. And the artists drew what they were familiar with, which made the Marvel universe authentic-looking, down to the water towers atop many of the buildings. https://t.co/rDosqzpP8i
The Avengers Mansion was a Beaux-Arts palace. Fans know it as 890 Fifth Avenue. The Frick Collection, which now occupies the place, uses the address of the front door: 1 East 70th Street.
