Thanks, trying to find the most sensible route to update my workflow while maintaining style compatibility with old lecture videos I do not want to rerecord.
29.01.2026 20:41
👍 0
🔁 0
💬 0
📌 0
Thanks, trying to find the most sensible route to update my workflow while maintaining style compatibility with old lecture videos I do not want to rerecord.
Are you going via beamer or direct to slides? Notebook lm did an amazing job on one of mine, but not editable.
Our ARC DP on Skillphabets for robot tool use was funded!
We will have two PhD positions open on this grant (dextrous manipulation with reusable skills, human-robot information asymmetry) and a postdoc (learning skill decomposition from demonstrations). Drop me a message if you are interested.
Here is the recording of the Robotics Worldwide Workshop, held at MIT on April 4, 2025: www.youtube.com/watch?v=0OdO... - includes "All Robotics in 90 minutes" lightning talks and a panel discussion on the future of robotics. not for the faint of heart. #mitSparkLab @mitofficial.bsky.social
Ending the #RSS AC meeting on a high note with @mgb45.bsky.social, Rudolf Lioutikov, Jana Tumova, and Abhinav Valada. Can’t wait to see all of you again soon!
This was a lot of fun! Livestream seems to now be a recording of you want to watch.
I'm very excited to host the Robotics Worldwide Workshop at MIT with Amanda Prorok next week! We have an incredible lineup of roboticists from around the world who will join us for lightning talks, a panel discussion, and a poster session. Join us if you are in the area!
This is great!
Here's how I used up all my Sora credits.
First, my number one most impossible video model benchmark, "person blowing out a candle"
3 days of keynotes, 10-min paper talks, and 1-min spotlights on the latest research, all free and open to anyone interested! Conference on Robot Learning (CoRL) talks are now online:
youtube.com/watch?v=0joZwi…
#Robotics #DeepLearning #CoRL2024
I often think of this in the same way as the bitter lesson. If we ignore morphology as a solution parameter, are we just self indulgently toying with puppets while waiting for hardware engineers to put the work in.
My favourite control textbook is Seborgs process control and dynamics - I feel like those chem eng examples just make everything so much easier to follow.
Adding my love letter to
arxiv.org/pdf/2304.01315
Empirical Design in Reinforcement Learning
by
Andrew Patterson, Samuel Neumann, Martha White, Adam White
JMLR 25 (2024) 1-63
#ReinforcementLearning
These aren’t the heroes we deserve, but they are the heroes we need.
You can use this feed to get only mentions bsky.app/profile/did:...
The case of the missing teaspoons doi.org/10.1136/bmj....
This also means inequity in reviewing, papers that look good are reviewed by more stringent experienced reviewers, while run of the mill papers go the way of grad students, who tend to give noncommittal reviews for fear of making wrong decisions -> ave conf score of 5, endless revise& resubmit.
The problem is that it is almost impossible to get a tenured academic to agree to review anything right now. The last paper I handled for RA-L needed 35 requests before securing 3 experienced reviewers. You can maybe get a new assistant prof. or postdoc, but grad students are still the most willing.
Challenge accepted! Our incredible undergrads have just released a new open quadruped for you to make at home.
youtu.be/vo-NXgQ52jU?...
github.com/MonashRoboti...
The Onion should buy Elsevier next
This is great!
Thanks for getting this started!
