Xiaohan Zhang @XiaohanZhang220 Twitter profile

Last Seen Profiles

@excell305

@couples_mot7arr

@Bokepindoo2023

@cd_aktif

@bokeplokalmalam

@PlutoRockDog

@ibubohay2

@mohmmaed2S

@stwmaniax

@bokeplokalmalam

@stw46

@Acquire_Fi

@Joelle_K23

@stwmaniax

@Rubitvcuerna

@jzuckman

@stanfordpedcard

@katarinaellen

@bokeplokalmalam

@Bokepindoo2023

@q753159

@stw_pdg

@ozC3x0ghrGlnARU

@CrescentJun_PH

@snmk_info

@smail22315549

@Viniciuskuster7

@bokeplokalmalam

@Noor_Naser2000

@stw_pdg

@Erni66809403

@udo_mbakara1

@muahid_hafiz

@stwmaniax

@KristaMckinstry

@althabty91798

Xiaohan Zhang

@XiaohanZhang220

3 months

We have openings for Fall interns in our Foundation Model team at Boston Dynamics AI Institute. Please feel free to DM if you are interested.

5

23

249

Xiaohan Zhang

@XiaohanZhang220

2 years

How do you combine Large Language Models (LLMs) with Task and Motion Planning (TAMP)? 📢 Introducing LLM-GROP ✅ Use prompting to extract commonsense knowledge for semantically valid arrangements ✅ Instantiation with TAMP in order to generalize to varying scene geometries 🧵👇

2

25

127

Xiaohan Zhang

@XiaohanZhang220

1 year

S3O: Symbolic State Space Optimization tl;dr: Solving Task and Motion Planning problems without predefining task-level state space in mobile manipulation domains. w/ @yifengzhu_ut @yding25 @yuqian_jiang @yukez @PeterStone_TX @ShiqiZhang7 Check it out at @IROS2023 next week!

1

9

53

Xiaohan Zhang

@XiaohanZhang220

1 year

LLM-GROP will also be presented next week at #IROS2023 ! Come to chat with us on LLMs and classical planning🧐

Xiaohan Zhang

@XiaohanZhang220

2 years

How do you combine Large Language Models (LLMs) with Task and Motion Planning (TAMP)? 📢 Introducing LLM-GROP ✅ Use prompting to extract commonsense knowledge for semantically valid arrangements ✅ Instantiation with TAMP in order to generalize to varying scene geometries 🧵👇

2

25

127

0

9

23

Xiaohan Zhang

@XiaohanZhang220

3 years

Excited to share our new work on visually grounded task and motion planning for mobile manipulation. #ICRA2022 Paper: Project page: Amazing collaborators: @yifengzhu_ut @yding25 @yukez @PeterStone_TX @ShiqiZhang7

0

1

18

Xiaohan Zhang

@XiaohanZhang220

1 year

Language-conditioned mobile manipulation skills learning from only a few demonstrations👇

Priyam Parashar

@priyam8parashar

1 year

Robot learning of language and manipulation tasks needs to be sample efficient. SLAP combines language and point-cloud embeddings as spatial-language tokens within a Transformer, to do just that – learn free-form language-conditioned robot policies. 🧵

4

20

141

0

1

13

Xiaohan Zhang

@XiaohanZhang220

2 years

Everyone deserves a wiping robot 🤖 Very interesting work led by @thomas__lew during my last internship with the robotics team @GoogleAI

Thomas Lew

@thomas__lew

2 years

📢Excited to share our #ICRA2023 work on robotic table wiping via RL + optimal control! 📖 🎥 💡RL (for high-level planning) + trajectory optimization (for precise control) can solve complex tasks without on-robot data collection ⬇️

3

7

47

0

1

8

Xiaohan Zhang

@XiaohanZhang220

7 months

It's always exciting to me how foundation models redefine the future of robotics and embodied AI, then we really need reliable benchmarks, especially for long-horizon vision&language understanding. We build real-world datasets and provide clean and simple baselines in OpenEQA.

AI at Meta

@AIatMeta

7 months

Today we’re releasing OpenEQA — the Open-Vocabulary Embodied Question Answering Benchmark. It measures an AI agent’s understanding of physical environments by probing it with open vocabulary questions like “Where did I leave my badge?” More details ➡️

38

258

1K

2

8

Xiaohan Zhang

@XiaohanZhang220

2 years

Paper: Webpage: Co-lead with @yding25 , collaboration with @chris_j_paxton and @ShiqiZhang7 🦾

LLM-GROP

Task and Motion Planning with Large Language Models for Object Rearrangement Yan Ding* 1, Xiaohan Zhang* 1, Chris Paxton 2, Shiqi Zhang 1 (* equal contribution) 1 SUNY Binghamton; 2 Meta AI Accepted...

sites.google.com

2

0

6

Xiaohan Zhang

@XiaohanZhang220

2 years

Table wiping is indeed a non-trivial task for robot perception and MoMa whole-body motions. Really nice blog post summarizing the project led by @thomas__lew

Google AI

@GoogleAI

2 years

Read how we enabled a robot to reliably wipe up crumbs and spills with an approach for robotics applications in complex environments that uses an #RL policy (trained with a stochastic differential equation simulator) followed by a trajectory optimizer. →

23

82

319

1

0

3

Xiaohan Zhang

@XiaohanZhang220

3 months

DM is open now

0

3

Xiaohan Zhang

@XiaohanZhang220

1 year

@theo_gervet @MistralAI @arthurmensch @GuillaumeLample @tlacroix6 Congrats!

0

2

Xiaohan Zhang

@XiaohanZhang220

9 months

@jdvakil @ieee_ras_icra congrats!

0

Xiaohan Zhang

@XiaohanZhang220

11 months

@thomas__lew Congrats congrats!

0

1

Xiaohan Zhang

@XiaohanZhang220

2 years

@ShiqiZhang7 Congrats, Shiqi!

0

2

Xiaohan Zhang

@XiaohanZhang220

8 months

@BingCompSci @ShiqiZhang7 @yding25 congrats!

1

0

1

Xiaohan Zhang

@XiaohanZhang220

2 years

We generate symbolic spatial relationships between objects using LLMs. Furthermore, by using an adaptive sampler, those **symbolic** descriptions are grounded to a set of valid **geometric** configurations.

2

0

1

Xiaohan Zhang

@XiaohanZhang220

7 months

OpenEQA is accepted to CVPR this year. Paper: Project website:

0

1

Xiaohan Zhang

@XiaohanZhang220

1 month

@shahdhruv_ @Princeton Congrats!

0

3

Xiaohan Zhang

@XiaohanZhang220

2 years

In the proposed system, valid geometric configurations are goal candidates for TAMP. Plans are optimized towards maximizing long-term utility (seeking the best trade-off between motion feasibility and task completion efficiency).