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
A brief analysis and comparison of the CSS for Twitter's PWA vs Twitter's legacy desktop website. The difference is dramatic and I'll touch on some reasons why.
Legacy site *downloads* ~630 KB CSS per theme and writing direction.
6,769 rules
9,252 selectors
16.7k declarations
3,370 unique declarations
44 media queries
36 unique colors
50 unique background colors
46 unique font sizes
39 unique z-indices
https://t.co/qyl4Bt1i5x
PWA *incrementally generates* ~30 KB CSS that handles all themes and writing directions.
735 rules
740 selectors
757 declarations
730 unique declarations
0 media queries
11 unique colors
32 unique background colors
15 unique font sizes
7 unique z-indices
https://t.co/w7oNG5KUkJ
The legacy site's CSS is what happens when hundreds of people directly write CSS over many years. Specificity wars, redundancy, a house of cards that can't be fixed. The result is extremely inefficient and error-prone styling that punishes users and developers.
The PWA's CSS is generated on-demand by a JS framework that manages styles and outputs "atomic CSS". The framework can enforce strict constraints and perform optimisations, which is why the CSS is so much smaller and safer. Style conflicts and unbounded CSS growth are avoided.
Legacy site *downloads* ~630 KB CSS per theme and writing direction.
6,769 rules
9,252 selectors
16.7k declarations
3,370 unique declarations
44 media queries
36 unique colors
50 unique background colors
46 unique font sizes
39 unique z-indices
https://t.co/qyl4Bt1i5x
PWA *incrementally generates* ~30 KB CSS that handles all themes and writing directions.
735 rules
740 selectors
757 declarations
730 unique declarations
0 media queries
11 unique colors
32 unique background colors
15 unique font sizes
7 unique z-indices
https://t.co/w7oNG5KUkJ
The legacy site's CSS is what happens when hundreds of people directly write CSS over many years. Specificity wars, redundancy, a house of cards that can't be fixed. The result is extremely inefficient and error-prone styling that punishes users and developers.
The PWA's CSS is generated on-demand by a JS framework that manages styles and outputs "atomic CSS". The framework can enforce strict constraints and perform optimisations, which is why the CSS is so much smaller and safer. Style conflicts and unbounded CSS growth are avoided.
I'm going to do two history threads on Ethiopia, one on its ancient history, one on its modern story (1800 to today). 🇪🇹
I'll begin with the ancient history ... and it goes way back. Because modern humans - and before that, the ancestors of humans - almost certainly originated in Ethiopia. 🇪🇹 (sub-thread):
The first likely historical reference to Ethiopia is ancient Egyptian records of trade expeditions to the "Land of Punt" in search of gold, ebony, ivory, incense, and wild animals, starting in c 2500 BC 🇪🇹
Ethiopians themselves believe that the Queen of Sheba, who visited Israel's King Solomon in the Bible (c 950 BC), came from Ethiopia (not Yemen, as others believe). Here she is meeting Solomon in a stain-glassed window in Addis Ababa's Holy Trinity Church. 🇪🇹
References to the Queen of Sheba are everywhere in Ethiopia. The national airline's frequent flier miles are even called "ShebaMiles". 🇪🇹
I'll begin with the ancient history ... and it goes way back. Because modern humans - and before that, the ancestors of humans - almost certainly originated in Ethiopia. 🇪🇹 (sub-thread):
The famous \u201cLucy\u201d, an early ancestor of modern humans (Australopithecus) that lived 3.2 million years ago, and was discovered in 1974 in Ethiopia, displayed in the national museum in Addis Ababa \U0001f1ea\U0001f1f9 pic.twitter.com/N3oWqk1SW2
— Patrick Chovanec (@prchovanec) November 9, 2018
The first likely historical reference to Ethiopia is ancient Egyptian records of trade expeditions to the "Land of Punt" in search of gold, ebony, ivory, incense, and wild animals, starting in c 2500 BC 🇪🇹
Ethiopians themselves believe that the Queen of Sheba, who visited Israel's King Solomon in the Bible (c 950 BC), came from Ethiopia (not Yemen, as others believe). Here she is meeting Solomon in a stain-glassed window in Addis Ababa's Holy Trinity Church. 🇪🇹
References to the Queen of Sheba are everywhere in Ethiopia. The national airline's frequent flier miles are even called "ShebaMiles". 🇪🇹
