“Task-Based” or “Semantic” robot grasping — where the grasp criteria are based on how the object will be used —has a long history. Humans grasp objects intuitively based on affordances. We believe this is the first paper that applies LLM + CLIP to infer affordances from words.
@ken_goldberg
-
The Real AI Revolution in Astronomy Hasn’t Happened Yet
By
–
Starting now: The Real AI Revolution in Astronomy Hasn’t Happened Yet. @profjsb Josh Bloom, Astronomy,
@UCBerkeley Online:
https://
berkeley.zoom.us/j/98700193604 -
Linear Autoregressive Predictors Learn All Turing-Computable Functions
By
–
“Any Turing-computable function can be learned with a linear autoregressive next token predictor.”
-
Ryan and Gaurav Present AI Research Paper at CoRL2023
By
–
Looking fwd to Ryan and Gaurav presenting this exciting paper at #CoRL2023 in Atlanta! https://t.co/M35BjuY66F
— Ken Goldberg (@Ken_Goldberg) 12 septembre 2023Looking fwd to Ryan and Gaurav presenting this exciting paper at #CoRL2023 in Atlanta!
-
250 Years of Robotics Research: Berkeley Panel Honoring Ruzena Bajcsy
By
–
250 Years of Robotics Research : @UCBerkeley panel honoring Ruzena Bajcsy from @GRASPlab @StanfordAILab @Berkeley_AI https://
youtube.com/live/HHiVAeRDd
Vs?si=FaUvKPfY-QFU1WP6
… -
250 Years of Robotics Research Berkeley AI Livestream Today
By
–
250 Years of Robotics Research: Today at noon. @berkeley_ai
. @UCBerkeley @rodneyabrooks Livestream @ https://
youtube.com/@BerkeleyEECS/
live
… -
Self-driving cars progress and future challenges discussion
By
–
Interesting discussion of past and future of self-driving cars with
@CadeMetz
. Engineers have made remarkable progress, but much remains to be done… -
Shane Wighton Tests His Latest Robot Creation
By
–
Hats off to Shane Wighton of @stuffmadehere for self-testing his latest robot!
— Ken Goldberg (@Ken_Goldberg) 20 août 2023
Hairstylists: don't worry; your jobs are safe… https://t.co/aG3BpLeyuHHats off to Shane Wighton of @stuffmadehere for self-testing his latest robot! Hairstylists: don't worry; your jobs are safe…
-
Cruise robotaxi crashes not company’s fault, green light cases
By
–
I don't think we should blame @Cruise for this crash or the one with an SF emergency vehicle, both on Thurs night. In both cases, the robotaxi had the green light and was hit by the other vehicle. The @sfchronicle headline is misleading: "2 Driverless Cars Crash on Same Night".
-
AI Pushes Comedians Toward Authenticity and Uniquely Human Expression
By
–
“AI gives comedians incentive to abandon the more generic forms of their work…and inspires them to be more eccentric and personal. AI will force artists to ..,think harder about what makes them, and their work, distinctly human.”