Jorge Mendez-Mendez
@JorgeAMendez_Embodied lifelong learning (compositionality, RL, TAMP, robotics). Assistant Professor at @SBU_ECE. Postdoc at @MIT_LISLab. PhD from @GRASPlab.
Similar User
@drmapavone
@fcdl94
@AljundiRahaf
@ContinualAI
@CoLLAs_Conf
@mundt_martin
@JasonMa2020
@MassCaccia
@GMvandeVen
@TylerLHayes
@NikolaiMatni
@TLesort
@duke_zzwang
@DPalenicek
@YuchenXiao5
I am excited to share that I will join the ECE department at Stony Brook University @SBU_ECE @stonybrooku as an Assistant Professor starting this fall. I will be recruiting PhD students during this application cycle to work on robot learning!
Very excited to head back to Philly next year for CoLLAs! Submit your best work on lifelong learning!
📢 Exciting News! The Fourth Conference on Lifelong Learning Agents (CoLLAs 2025) will be held at the University of Pennsylvania (@Penn) in Philadelphia, USA 🇺🇸 🗓️ Important Dates: Abstract Deadline: Feb 21, 2025 Submission Deadline: Feb 26, 2025 Conference Dates: Aug 11 - Aug…
I am hiring multiple PhD students this year to launch the Princeton Robot Planning and Learning lab: tomsilver.github.io/hiring Thanks for helping me spread the word!
We're excited to have 4 different lab members moving on to start exciting new labs as faculty members! 🧵👇
As an author, I appreciate the opportunity to explain nuanced points to clear up confusions. As a reviewer, I often receive various paraphrases of the same response in hopes that one will stick and change my mind. In balance, I agree that open discussion phases are not working.
Opinion: we should either do single page pdf rebuttals, single pdf revision, or direct decisions This OpenReview craze of continuing to post more & more detailed boxes (that won't ever satisfy page limits) with the expectation for reviewers to agree eventually is simply absurd
Underappreciated finding in this paper: We compare single-task and multi-task MLPs to a compositional network. With few tasks, all achieve ~40% success. With 224 tasks, still 40% for STL and MTL but comp. solves them all! Task variety is key to enable the comp. inductive bias.
We don't need more #RL #benchmarks tasks. We need suites that test properties or capture real problems. That's why I loved CompoSuite. Sadly, it hasn't caught on as much as I'd like; likely because I didn't advertise/use it enough. On it now, stay tuned! arxiv.org/abs/2207.04136
We don't need more #RL #benchmarks tasks. We need suites that test properties or capture real problems. That's why I loved CompoSuite. Sadly, it hasn't caught on as much as I'd like; likely because I didn't advertise/use it enough. On it now, stay tuned! arxiv.org/abs/2207.04136
How can we make classifiers more robust when we can't modify the weights or assume its architecture — effectively making it a black box? In our preprint, we demonstrate that we can improve robustness by augmenting the inputs to the model using an LLM. 1/8 arxiv.org/abs/2402.08225
Looking forward to meeting with folks and chatting about continual learning and causality at #AAAI24!
Undecided whether to attend the #AAAI24 Continual Causality Bridge on Feb 20+21? continualcausality.org -> 4 more reasons to reg early (before Dec 20) for this unique #continuallearning + #causality event! Invited talks by @white_martha @JorgeAMendez_ @egavves & Elina Robeva!
Excellent piece from @nishanthkumar23 on the most common debate at #CoRL23: is scale all we need to solve robotics? Have a read for a balanced and well crafted perspective (or summary of perspectives)!
There was a lot of good and interesting debate on "is scaling all we need to solve robotics?" at #CoRL23. I spent some time writing up a blog post about all the points I heard on both sides: nishanthjkumar.com/Will-Scaling-S…
Exciting!
Thrilled to announce the first annual Reinforcement Learning Conference @RL_Conference, which will be held at UMass Amherst August 9-12! RLC is the first strongly peer-reviewed RL venue with proceedings, and our call for papers is now available: rl-conference.cc.
Excited to present our work on features fields for robotic manipulation at #CoRL2023 🤖! Honored to be selected as a finalist for the Best Paper/Student Paper award. Check out our oral talk on Thursday morning (Oral 6) and poster in the afternoon (Poster 5, 2:45-3:30pm)
New paper at #CoRL2023! "Distilled Feature Fields Enable Few-Shot Language-Guided Manipulation" How should robots represent the world around them? This paper's answer: as a field of foundation model features localized in 3D space. paper+code: f3rm.github.io 1/n
If you’re looking for research interns for your team on RL, @marcel_hussing is a fantastic researcher! He has worked on applied and theoretical RL and has very deep insight into both.
Looking for summer research internship opportunities with focus on publication. Seems like most stuff is LLMs now. If you know (or you are) somebody with RL positions, feel free to hit me up! Would love to chat, even if it's just to get to know new people. RTs appreciated.
Looking forward to visiting @RutgersCS tomorrow! I will be talking about modularity, lifelong learning, and robots.
Our next seminar will be presented by @JorgeAMendez_ In this talk, Jorge will explain how we can leverage various forms of modularity that arise in robot systems to develop powerful lifelong learning mechanisms. cs.rutgers.edu/news-events/cs…
Our #CoRL2023 paper shows that by composing the energies of diffusion models trained to sample for individual constraint types such as collision-free, spatial relations, and physical stability, it can solve novel combinations of known constraints. diffusion-ccsp.github.io (🧵 1/N)
Interesting read on how embodiment might help lifelong learning. We often talk about compositionality, physical grounding, active learning and test-time learning. It is crucial to find a comprehensive framework to bring them together. #EmbodiedAI #Robotics #LifelongLearning
"Embodied intelligence is the ultimate lifelong learning problem." I discuss how to tackle this problem via compositionality in this piece. I will be on the job market for faculty positions this cycle—reach out if I'd be a good fit for your department! lis.csail.mit.edu/embodied-lifel…
Excited that our paper on embodied lifelong learning for TAMP got accepted @corl_conf Looking forward to presenting it and meeting folks in Atlanta!
For the past year I have been thinking a lot more seriously about what #LifelongLearning entails in the context of #Robotics. This *new preprint* is one first step towards answering that question! 🧵👇 arxiv.org/abs/2307.06870
"Embodied intelligence is the ultimate lifelong learning problem." I discuss how to tackle this problem via compositionality in this piece. I will be on the job market for faculty positions this cycle—reach out if I'd be a good fit for your department! lis.csail.mit.edu/embodied-lifel…
🚨Another pre-print alert! 🚨 In the spirit of our CompoSuite benchmark from last year, we are releasing four datasets to enable the study of offline compositional RL. 🔗arxiv.org/abs/2307.07091 👨🎓👩🎓 @JorgeAMendez_, @Singrodiaanisha, Cassandra Kent, @EricREaton 🧵👇 (1/5)
United States Trends
- 1. #TysonPaul 137 B posts
- 2. Serrano 202 B posts
- 3. #NetflixFight 57,2 B posts
- 4. #netflixcrash 13,4 B posts
- 5. Shaq 12,8 B posts
- 6. #buffering 9.535 posts
- 7. Rosie Perez 11,4 B posts
- 8. My Netflix 67,4 B posts
- 9. Tori Kelly 4.240 posts
- 10. ROBBED 87,6 B posts
- 11. Cedric 18 B posts
- 12. Ramos 67,6 B posts
- 13. Barrios 48,4 B posts
- 14. #boxing 39,3 B posts
- 15. Roy Jones 4.851 posts
- 16. Father Time 8.132 posts
- 17. Phil Collins 3.330 posts
- 18. Jerry Jones 8.312 posts
- 19. Christmas Day 14,5 B posts
- 20. He's 58 6.757 posts
Who to follow
-
Marco Pavone
@drmapavone -
Fabio Cermelli
@fcdl94 -
Rahaf Aljundi
@AljundiRahaf -
ContinualAI
@ContinualAI -
CoLLAs 2025
@CoLLAs_Conf -
Martin Mundt
@mundt_martin -
Jason Ma
@JasonMa2020 -
Massimo Caccia
@MassCaccia -
Gido van de Ven
@GMvandeVen -
Tyler Hayes
@TylerLHayes -
Nikolai Matni
@NikolaiMatni -
Timothée Lesort
@TLesort -
Zizhao Wang
@duke_zzwang -
Daniel Palenicek
@DPalenicek -
Yuchen Xiao
@YuchenXiao5
Something went wrong.
Something went wrong.