@HaiNguy69482974 Profile picture

Hai Nguyen

@HaiNguy69482974

Ph.D. Student on Robot Learning under Partial Observability @Northeastern / Applied Scientist Intern @Amazon / Ex-Intern @omron_sinicx / @CheveningAA Alumni

Joined October 2018
Similar User
Dian Wang photo

@Dian_Wang_

Robert Platt photo

@RobotPlatt

Helping Hands Lab @ Northeastern photo

@HelpingHandsLab

Peter Stone photo

@PeterStone_TX

Chris Amato photo

@cjdamato

Ondrej Biza photo

@BizaOndrej

Haojie Huang photo

@HaojieHuang13

Yuchen Xiao photo

@YuchenXiao5

Guanang (Shirley) Su photo

@Guanang_Su

Xupeng Zhu photo

@XupengZ

Mingxi Jia photo

@SaulBadman12

Robin Walters photo

@RobinSFWalters

Linfeng Zhao photo

@LinfengZhaoZLF

Eric Rosen photo

@_ericrosen

Jung Yeon Park photo

@park_jungy

Hai Nguyen Reposted

#CoRL2024 OrbitGrasp: SE(3)-Equivariant Grasp Learning Led by @boce_hu Orbitgrasp maps each point in the cloud to a continuous grasp quality function using spherical harmonics. Our method outperforms all baselines across all settings and tasks. Presenting at Poster Session 1.

Tweet Image 1

Please check out my final publication during my PhD. Happy to chat more at #CoRL2024, poster session 2.

#CoRL2024 Leveraging Mutual Information for Asymmetric Learning under Partial Observability led by @HaiNguy69482974 Addressing asymmetric learning under partial observability (state availability at training) by rewarding actions leads to histories that gain info about the state.

Tweet Image 1


Hai Nguyen Reposted

#CoRL2024 ThinkGrasp: A Vision-Language System for Strategic Part Grasping in Clutter by @RubyFreax A plug-and-play vision-language grasping system that uses GPT-4o’s advanced contextual reasoning for heavy clutter environment grasping strategies. Presenting at Poster Session 3.

Tweet Image 1

Hai Nguyen Reposted

#CoRL2024 IMAGINATION POLICY: Using Generative Point Cloud Models for Learning Manipulation Policies Led by @HaojieHuang13 A key-frame multi-task policy can generate key poses (imagine) and do manipulation precisely with sample efficiency. Presenting at Poster Session 4.

Tweet Image 1

Hai Nguyen Reposted

#CoRL2024 Equivariant Diffusion Policy Led by @Dian_Wang_ A sample efficient BC algorithm based on equi diffusion. It leverages symmetry to boost learning with 5x less training data and mastering complex tasks with <60 demos. Presenting at Oral Session 1 and Poster Session 2.


Hai Nguyen Reposted

🤖 Excited for #RSS2024? Don’t miss our workshop on "Geometric and Algebraic Structure in Robot Learning"! Submit workshop papers (by 6/10 AOE) and dive into discussions on leveraging these structures for enhanced #robotics. 🚀🔍 Join us on 07/19 in Delft, Netherlands!

Tweet Image 1

Hai Nguyen Reposted

I wrote a survey covering decentralized training and execution methods for cooperative MARL. Hopefully, people find it useful. Any feedback is more than welcome! arxiv.org/abs/2405.06161


Happy to share a @ieee_ras_icra paper from my last summer internship in beautiful Tokyo at @omron_sinicx We learned the Peg-In-Hole from tactile feedback directly on a soft wrist in 3 hours by leveraging domain symmetry and demonstrations.

我々は、ICRA2024で柔軟なロボットの部品挿入の研究を発表します。手先の姿勢を観測できない部分観測な環境で、力覚を使用した学習を考えました。 部品の幾何的な対称性を考慮し効率的な強化学習も提案しました。インターン @HaiNguy69482974さん、OSX @Tdash_Kozさん、@cambel07さんとの研究です。



Hai Nguyen Reposted

How can we combine the benefits of Fourier Transformation to further facilitate the efficiency of policy learning? In our Fourier Transporter (FouTran), we introduce a novel 3D Pick-Place Model with Bi-equivariance property. Paper: arxiv.org/pdf/2401.12046

#ICLR24 We proposed FourTran, a very sample-efficient 3D manipulation pick-place model. 1. It can learn a nontrivial 3D policy with less than 10 demos. 2. It represents 3D action distribution in Fourier Space. Check it in the Poster Session 4 at 4:30 PM Vienna time (10:30 EDT)

Tweet Image 1


Happy to chat more on today's poster session from 4.15 to 5.00 pm at CoRL.

Happy to share our #CoRL23 paper about leveraging symmetry for POMDPs. We formulate group-invariant POMDPs that are symmetric in the history space. As expected, equivariant recurrent policies shine in solving them. For more, please see sites.google.com/view/equi-rl-p….



New #IROS2023 paper alert: A Bayesian-Adaptive RL with mixed observability assumption, training directly on hardware, greatly more sample-efficient >> RL methods. Joint w/ SammieKatt @YuchenXiao5 @cjdamato @LLPR_NEU Paper: arxiv.org/abs/2307.11954 Video: youtube.com/watch?v=H9xp60…


Hai Nguyen Reposted

Delighted to announce the schedule of the Workshop on Symmetries in Robot Learning @RoboticsSciSys! Join the discussion with our incredible speakers @tesssmidt @haosu_twitr @GhaffariMaani @andyzeng_ @erikjbekkers @HughWang19 Can't wait to see you there! sites.google.com/view/rss23-sym.

Tweet Image 1

Hai Nguyen Reposted

Excited to announce the #RSS2023 Workshop on Symmetries in Robot Learning! Submit your work and join us to explore how symmetries can help in robotics with our incredible speakers @andyzengtweets @HughWang19 @haosu_twitr @tesssmidt @erikjbekkers Website: sites.google.com/view/rss23-sym.

Tweet Image 1

Hai Nguyen Reposted

We(@cjdamato) recently published MacroMARL(github.com/yuchen-x/Macro…), including both value-based and policy-gradient-based algorithms for multi-agent asynchronous learning and execution, as well as three macro-action-based multi-agent domains. To find more, please check it out!


Interested in partial observability? Come to my online poster session at @CoRL2022 (online session 2, Sunday, Dec 18, 7:30-8.00 AM, Auckland time, i.e., Saturday, Dec 17, 1:30 - 2:00 PM, EST time) w/ @RobotPlatt @cjdamato @DianWang1007 @AndreaBaisero @LLPR_NEU @HelpingHandsLab

Glad to share our #CoRL22 paper arxiv.org/abs/2211.01991 about using MDP solutions to efficiently learn POMDPs during offline training. We proposed a SAC-like agent that balances between acting like an MDP expert and for environment rewards.



Hai Nguyen Reposted

Tmr(Thurs), 4-6pm, Hall J 613 #NeurIPS2022, my advisor Chris @cjdamato is going to present how we enable agents to asynchronously learn and execute. Please stop by to check the details and know more about the lab @LLPR_NEU Also, feel free to contact me for any questions!

Tweet Image 1

I am very excited to share that our paper "Asynchronous Actor-Critic for Multi-Agent Reinforcement Learning" (arxiv.org/abs/2209.10113) has been accepted to #NeurIPS2022. A step towards asynchronous and hierarchical behaviors in real-world multi-agent/robot systems.



Hai Nguyen Reposted

LLPR member Daniel Melcer shows decentralized shielding can be an effective solution for multi-agent RL and doesn’t require communication between agents #NeurIPS2022 Link to paper: openreview.net/pdf?id=JO9o3Dg…

Tweet Image 1

Hai Nguyen Reposted

Looking for an easy-to-read review paper on Grasp Learning? I just wrote one for Annual Reviews that I'm excited about. Check it out: arxiv.org/abs/2211.04895


Loading...

Something went wrong.


Something went wrong.