SVL is led by @drfeifei @silviocinguetta @jcniebles @jiajunwu_cs and works on machine learning, computer vision, robotics and language
Stanford, CA
- Tweets 372
- Following 152
- Followers 17,726
- Likes 335
Street View Image, Pose & 3D Cities Dataset. 25 million images, 3D models of 8 cities, camera pose & correspondences github.com/amir32002/3D_Stre…
3
138
289
#TransferLearning is crucial for general #AI, and understanding what transfers to what is crucial for #TransferLearning. Taskonomy (#CVPR18 oral) is one step towards understanding transferability among #perception tasks. Live demo and more: taskonomy.stanford.edu/
2
127
258
Introducing the RoboTurk Real Robot Dataset - one of the largest, richest, and most diverse robot manipulation datasets ever collected using human creativity and dexterity!
111 hours
54 non-expert demonstrators
2144 demonstrations
Download: roboturk.stanford.edu/realro…
[1/2]
1
47
151
Stanford Vision and Learning Lab is presenting 7 papers at #CORL2023, including 3 oral presentations, and 3 award nominations, see below:
1
8
69
46,551
[1/2] Our lab has 3 papers accepted to NeurIPS 2019:
1. HYPE: Human Eye Perceptual Evaluation of Generative Models. Zhou and Gordon et al. (Oral)
2. SOCIAL-BIGAT: Multimodal Trajectory Forecasting using Bicycle-GAN and Graph Attention Networks. Kosaraju et al.
2
7
59
Stanford Vision and Learning Lab: Performing Research at the Forefront of Computer Vision, Machine Learning, and Robotics - HostingAdvice.com
@drfeifei @silviocinguetta @jcniebles hostingadvice.com/blog/a-loo…
12
55
Stanford Vision and Learning Group website is online now! @drfeifei @silviocinguetta @jcniebles svl.stanford.edu
19
41
We are hosting one of the 3 challenges of embodied-ai.org at CVPR20. Train your navigating agent in our simulator Gibson (svl.stanford.edu/gibson2) and we will test it in the real world! The best solutions will showcase live during CVPR. More info: svl.stanford.edu/gibson2/cha…
1
7
37
Learning from hints (not demonstrations): A new paper on an important direction of RL for control where expert intuition can be used to guide learning without the need to provide optimal or even complete solutions.
Our new work at #Corl2019 will present RL with Ensemble of Suboptimal Teachers -aka- specify as much as you can easily, let learning handle the rest.
Blog: buff.ly/2O99xWr
Paper: buff.ly/2NY9KeU
w\ @andrey_kurenkov, A. Mandlekar, @RobobertoMM, @silviocinguetta
4
31
Work from our group in Robot Learning for Manipulation is finalist for best paper award at @icra2019 and is being presented tomorrow in Montreal. @drfeifei @silviocinguetta @animesh_garg @yukez @michellearning @leto__jean
Excited to be at #ICRA2019
Best Paper Award talk
Making Sense of Vision and Touch: Self-Supervised Learning of Multimodal Representations for Contact-Rich Tasks
Paper: arxiv.org/abs/1810.10191
Video: piped.video/TjwDJ_R2204
6
28
Wiki: github.com/StanfordVL/robotu…
Paper and more details: arxiv.org/abs/1911.04052
[2/2]
6
25
We are happy to announce our ICCV19 Workshop on Visual Perception for Navigation in Human Environments: The JackRabbot Social Robotics Dataset and Benchmark. Submission deadline August 20. jrdb.stanford.edu/workshops/… For more info, contact @SHamidRezatofig and Roberto Martin-Martin
5
22
Are you a passionate and experienced researcher in robotics with knowledge in computer vision? Do you want to build impactful robotic systems? Stanford Vision and Learning lab (SVL) is searching for a Postdoctoral Fellow with your skills.
1
3
22
4,336
Our focus on robot learning from single example of a task through a video has resulted in a line of work that combines symbolic systems with neural networks
This continues our efforts in neuro-symbolic planning for one-shot imitation in multi-step reasoning domains.
1. Neural Task Programs: arxiv.org/abs/1710.01813
2. Neural Task Graphs: arxiv.org/abs/1807.03480
3. Continuous Relaxation of Symbolic Planner: arxiv.org/abs/1908.06769
4
22
The group wins the best paper award at #ICRA2019 for work on multimodal state representations in robot policy learning
1
20
New work on pose tracking from our group
We have just released our new work on 6D pose estimation from RGB-D data -- real-time inference with end-to-end deep models for real-world robot grasping and manipulation! Paper: arxiv.org/abs/1901.04780 Code: github.com/j96w/DenseFusion w/ @danfei_xu @drfeifei @silviocinguetta
1
3
19
Stanford Vision and Learning Lab members and alumni: we are organizing a SVL+Alumni meet up at CVPR’24! Please check your emails for more details.
1
1
27
33,798
Congrats @silviocinguetta for the appointment as the inaugural Mindtree Faculty Scholar at Stanford
I am delighted to share the news that I have been appointed the inaugural Mindtree Faculty Scholar at Stanford. More info: markets.businessinsider.com/…
3
20
Check out New work from the group on task oriented grasping at #RSS18 by @KuanFangAI @yukez @animesh_garg @andrey_kurenkov @thebigmehtaphor @drfeifei @silviocinguetta
Check out our paper on Task-Oriented Grasping at #RSS2018 at 930am on Wed 06/27 buff.ly/2K9WEKE (find full paper here: buff.ly/2KkPiQK)
5
17
A new work on structuring diverse semantics in 3D space that yielded the 3D Scene Graph! It’s showcased on the Gibson database by annotating the models with diverse semantics using a semi-automated method.
What space should diverse semantics be grounded & what should be the structure? 3D Scene Graph is a 4-layer structure for unified semantics, 3D space &camera. We demonstrate it on Gibson models with an automated labeling method. Data available to download! 3dscenegraph.stanford.edu/
7
18
Composable Part-based Manipulation uses object-part decomposition & part-part correspondences to improve generalization of robotic manipulation across object categories. @Weiyu_Liu_
Poster: Thu 9 2:45-3:30 pm
Website: cpmcorl2023.github.io
w/ Tucker Hermans & @animesh_garg
2
19
9,898
1
3
19
This is continuing an important line of work in policy learning with large Datasets.
More importantly this is the question to answer if we need to create a analog of "Imagenet" for Robotics.
We need to both collect large Datasets and have Algorithms to leverage this data!
New work on exploring if a policy can be learned only from offline, off-policy dataset?
IRIS: Implicit Reinforcement without Interaction at Scale
arxiv.org/abs/1911.05321
Video: piped.video/H9KZgrI2I7I
Seattle Robotics @NvidiaAI
@AjayMandlekar @drfeifei B. Boots F. Ramos D. Fox
1
18
Joint 2D-3D-Semantic data for scene understanding is out! 70K+ images and 6 mutually registered modalities in 2D&3D: 3Dsemantics.stanford.edu
9
14
Congratulations to Ashesh, Amir, Silvio and Ashutosh for the best student paper award at CVPR 2016!!Structural-RNN!!
1
1
11
Continued efforts in larger scale crowdsourcing robotics for dataset creation in setups where engineered solutions are hard, simulation is tricky and pure compute has low success. The diversity of human cognitive reasoning and dexterity provides so many ways do do the same task!
Scaling crowdsourcing robotics with Human Reasoning & Dexterity for Large-Scale Dataset Creation
Blog: ai.stanford.edu/blog/robotur…
Paper: arxiv.org/abs/1911.04052
Webpage: roboturk.stanford.edu/realro…
@AjayMandlekar @yukez @silviocinguetta @drfeifei
J. Booher, M. Spero, A. Tung, A. Gupta
1
11
A thorough evaluation of possible action spaces and their efficiency with RL for manipulation tasks. The answer is best of both worlds - Op. Space control in end effector space with learnable gains fares better than end to end image-to-torque
Which control spaces work well for RL in contact-rich robot manipulation -- 'Variable Impedance Control in End-Effector Space' #IROS2019
Paper: arxiv.org/abs/1906.08880
Video: piped.video/AozIUIW3Ghs
w\ Roberto M-M., @michellearning, R. Gardener, @silviocinguetta, @leto__jean
ALT Policy Learning in Image State Space with VICES action space transfer from Sim2Real and across Robots
2
9
We created 4D ConvNets for 3D video perception :D
Please checkout our paper: 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks, @cvpr2019 arxiv.org/abs/1904.08755
1
9
PhD student @lynetcha gives a great overview of ongoing work in point cloud segmentation
Be sure to check out #TWiMLTalk #123! Joined by @lynetcha, we discuss her work on SEGCloud, an end-to-end framework that performs 3D point-level segmentation. Head over to buff.ly/2uuDQ1z to listen!
6
10
This is a very exciting new direction of research in our group: perception by combining vision with touch for manipulation.
Very excited to share our new paper on Making Sense of Vision and Touch: Self-Supervised Learning of Multimodal Representations. Featuring RL on real-robots from scratch in a matter of hours without any simulation!
Video: piped.video/TjwDJ_R2204
arXiv: arxiv.org/abs/1810.10191
3
10
Thanks @ToyotaResearch and @SAILToyota for the exciting new experimental platform!
New baby rolls in SVL lab (@StanfordCVGL) - the @ToyotaResearch HSR. looking forward to servicing some humans!
@drfeifei @silviocinguetta
3
9
Semantic Segmentation of 3D Point Clouds #3DV @lynetcha, @ChrisChoy208
Iro Armeni, @JunYoungGwak, @silviocinguetta segcloud.stanford.edu/
8
If you are at #CVPR18, there is a workshop organized by Berkeley, Stanford, and MIT on Beyond Supervised Learning, with a great lineup of speakers. At Ballroom G. Drop by! beyond-supervised.ai/ #perception #unsupervisedlearning #selfsupervisedlearning #transferlearning
1
8
8
Exciting news! CVGL got 6 papers accepted at ECCV, with an amazing unprecedented acceptance rate of ~85%!
3
7
Hot new paper on 3D reconstruction by @NoSleepMusings @ChrisChoy208 @animesh_garg and team
Our new work on Free-Form Deformation Network for 3D Shape Reconstruction from a Single Image aka DeformNet j.mp/2vK8UqX
1
8
Congrats to all the authors
Congratulations to Best Paper Award by @zamir_ar @silviocinguetta @StanfordCVGL & collaborators at #CVPR2018, a computational map of perceptual task transfer learning!! taskonomy.stanford.edu/
8
SegCloud was a recent Spotlight at 3DV and is the current leader on the Semantic3D.net reduced-8 benchmark. segcloud.stanford.edu/
Semantic Segmentation of 3D Point Clouds #3DV @lynetcha, @ChrisChoy208
Iro Armeni, @JunYoungGwak, @silviocinguetta segcloud.stanford.edu/
1
7
Humans have the remarkable ability to make and use tools to help them solve tasks. We introduce a framework for robots also to jointly learn to design and use tools via reinforcement learning.
Poster: Thu 9th 2:45-3:30 pm
Website: robotic-tool-design.github.i…
A robot may be unable to complete a task when limited by its morphology. Remarkably, people and some animals can get around this by not only using but also *designing* tools. We explore whether robots can also do this in our latest work!
🌐robotic-tool-design.github.i…
🧵👇
1
6
2,108
Lab PI @drfeifei wins the PAMI Longuest-Higgins Prize for the effort on ImageNet with @lijiali_vision @jiadeng @RichardSocher and team. Congrats everyone
6
Discussing remarkable vision capacity of humans with roboticists at @IROS2017 #inspiration
Plenary by @drfeifei at @IROS2017 The Cambrian explosion in robots: Robots need eyes - Perception meets Robotics nitter.app/animesh_garg/status/91…
5
7
Come check out the poster on recent work in imitation learning at #NeurIPS
"AC-Teach: A Bayesian Actor-Critic Method for Policy Learning with an Ensemble of Suboptimal Teachers"
Deep RL Workshop - Sat Dec 14 - sites.google.com/view/deep-r…
Paper: arxiv.org/abs/1909.04121
@andrey_kurenkov @AjayMandlekar @RobobertoMM @silviocinguetta @StanfordSVL
1
7
Three new papers from our lab to appear at #CoRL2017 @silviocinguetta @ozansener @NoSleepMusings @danfei_xu
5
6
Students from the lab @RanjayKrishna and Apoorva Doradula describe their new work on engagement learning!
Ranjay Krishna and Apoorva Doradula use conversations as a strategy for training AI systems. They call it engagement learning – an AI “learns what kinds of concepts people like to discuss and how to ask questions to get an informative response.” stanford.io/2I6U57g
ALT Magic Grant Winners 2018
2
5
NOIR is a brain-robot interface that enables humans to command robots to perform 20 challenging everyday activities using their brain signals, such as cooking, cleaning, & playing games.
Poster: Wed 8th 5:15pm
Website: noir-corl.github.io
Introducing our new work @corl_conf 2023, a novel brain-robot interface system: NOIR (Neural Signal Operated Intelligent Robots).
Website: noir-corl.github.io/
Paper: arxiv.org/abs/2311.01454
🧠🤖
1
5
765
Check out the workshop on a building causal models to go beyond reinforcement learning in robotics organized by lab members @animesh_garg and @yukez
Excited about the lineup of speakers to explore causality in robotics.
RSS Workshop on Causal-Imitation. CfP is out and deadline for posters is Jun 3. Consider submitting! @yukez @ermonste @jiajunwu_cs and Michael Laskey
buff.ly/2Lig8K0
1
5
[2/2]
3. Regression Planning Networks. Xu et al.
Congratulations to all the authors for their amazing work!
1
6
Exciting work on video analysis and another step on the path to having autonomous agents attempting to make us pasta
A great talk by Shyamal Buch at #CVPR18 on our paper on Weakly supervised reference aware visual grounding. @jcniebles @drfeifei @deanh619 Lucio Dery from @StanfordCVGL
2
6
Fresh work from the lab!
Our new work on weakly supervised GANs for 3D shape from images
Paper: arxiv.org/abs/1705.10904
Podcast: piped.video/D2dLHCrvDvs
1
5
Stanford scientists are teaching a robot how not to be awkward in public qz.com/893068 via @qz
1
3
VoxPoser uses LLM+VLM to create 3D value maps in robot workspace, which guides motion planner to synthesize behaviors for everyday manipulation tasks w/o requiring robot data.
Oral: Wed 8th 11:50 am
Poster: Wed 8th 5:15-6:00 pm
How to harness foundation models for *generalization in the wild* in robot manipulation?
Introducing VoxPoser: use LLM+VLM to label affordances and constraints directly in 3D perceptual space for zero-shot robot manipulation in the real world!
🌐 voxposer.github.io
🧵👇
1
2
5
2,038
MimicPlay is an imitation learning algorithm that uses cheap human play data to unlock real-time planning for long-horizon manipulation.
Oral: Thu 9th 8:30 am
Poster: Thu 9th 2:45-3:30 pm
Best paper/Best student paper finalist
Best system paper finalist
How to teach robots to perform long-horizon tasks efficiently and robustly🦾?
Introducing MimicPlay - an imitation learning algorithm that uses "cheap human play data". Our approach unlocks both real-time planning through raw perception and strong robustness to disturbances!🧵👇
1
4
1,415
Exciting direction of research by joint group, SVL, of @silviocinguetta and @drfeifei
Our new paper on Neural Task Programming: Learning to Generalize Across Hierarchical Tasks j.mp/2y8qYOS
5
Glad to share some of our efforts in healthcare
Check out our first blog post about SAIL research, about how computer vision can be used to enable smart hospitals
By Albert Haque & Michelle Guo (@mshlguo), led by professors Terry Platchek (@TerryPlatchek), Arnold Milstein, & Fei-Fei Li (@drfeifei)
ai.stanford.edu/blog/measuri…
4
Sequential dexterity is a system that learns to chain multiple dexterous manipulation policies for tackling long-horizon manipulation tasks in both simulation and real-world.
Poster: Tue 7th 2:45-3:30 pm
How to chain multiple dexterous skills to tackle complex long-horizon manipulation tasks?
Imagine retrieving a LEGO block from a pile, rotating it in-hand, and inserting it at the desired location to build a structure.
Introducing our new work - Sequential Dexterity 🧵👇
1
3
750
Lab PI @silviocinguetta meets up with dignitaries from the IT ministry of India to discuss AI challenges for India
It was an extra-ordinary visit to Stanford University. Had a wonderful exchange with its brilliant faculty on use of technology for human development and the application of AI and the challenge it poses. The faculty was deeply impressed with India's story of digital inclusion.
4
Four papers have been accepted in CVPR 2015 including one oral presentation! Congratulations to all the authors!
4
Exciting work on one shot robot learning at #ICRA2018 @animesh_garg @yukez @danfei_xu @silviocinguetta @drfeifei
This is being presented at #ICRA2018 by @StanfordCVGL student @danfei_xu on Wed 05/23 in Brisbane. Come say hi if you attending -buff.ly/2Ismg4S
1
4
A new effort from SVL and the JackRabbot team! New dataset and benchmark for robot perception in human environments. The winners of the first challenge on pedestrian detection and tracking will be presented at our workshop at #ICCV19! jrdb.stanford.edu/workshops/…
Want to see what JackRabbot sees? Finally a comprehensive dataset from the point of view of a mobile manipulating robot!
3
New study that analyzes different action spaces for RL in robot manipulation in the quest to find the best one. Guys what? The best method is a combination of operational space control (1986) with learned adaptive gain tuning. To appear at #IROS2019
What action space should we use for contact-rich manipulation? We show that Variable Impedance Control in End-Effector Space outperforms most other choices.
Paper: buff.ly/2KACdq7
w\ R. Martin-Martin, @michellearning, R. Gardner, @silviocinguetta, @leto__jean
@StanfordSVL
1
3
RL in multitask domains with shared underlying latent information can be made more effective through learning action space manifolds.
Check out this effort on Learning latent Action Spaces at #ICRA2021
When you try to open a new door, do you try to yank it up? Likely, no.
Why should your robot continue to do so!
Check out our #ICRA2021 paper on learning action spaces for efficient contact-rich manipulation.
paper: arxiv.org/abs/2103.15793
piped.video/7iFRdvV_pTU
2
4
Stanford’s ‘Jackrabbot’ robot will attempt to learn the arcane and unspoken rules tcrn.ch/25D5dCf via @techcrunch
3
Our paper "Robust Single-View Instance Recognition", by D. Held, S. Thrun , S. Savarese, has been accepted to ICRA arxiv.org/abs/1507.08286
1
2
Exciting new line of work from the group on 3D vision at #CVPR2019 by @lynetcha @SHamidRezatofig @silviocinguetta
Announcing the launch of "Completion3D: Stanford 3D Point Cloud Completion Benchmark” completion3d.stanford.edu/ . Come check out our associated paper "TopNet: Structural Point Cloud Decoder" at #CVPR2019! @silviocinguetta @SHamidRezatofig @StanfordSVL
2
Come check out labs new work at #ICRA2019
14:40-15:55, Grasping II 1.2.12 MoB1-12, Room 220
@silviocinguetta @ken_goldberg @animesh_garg @andrey_kurenkov
Presenting on Monday at #ICRA2019
Mechanical Search: Multi-Step Retrieval of a Target Object Occluded by Clutter. w\ @ken_goldberg @andrey_kurenkov
Paper: arxiv.org/abs/1903.01588
Video: piped.video/lCwdGSDkbG4
3
Do you know how to make a dumpling🥟? Our robot🤖does! RoboCook is a robot system designed for long-horizon manipulation of elasto-plastic objects with a variety of tools.
Oral: Tue 7th 8:50 am
Poster: Tue 7th 2:45-3:30 pm
Best system paper finalist
Do you know how to make a dumpling🥟? Our robot🤖does!
Introducing RoboCook: Long-Horizon Elasto-Plastic Object Manipulation with Diverse Tools.
Project website: hshi74.github.io/robocook/
Here we show how RoboCook makes a dumpling under external human perturbation. Thread🧵👇
1
2
1,030
Congrats @silviocinguetta. Looking forward to exciting research Stanford AI Center nitter.app/animesh_garg/status/82… nitter.app/animesh_garg/status/82…
2
3
Our work on automated process for acquiring detailed building information featured in Stanford News stanford.io/295G6Rj via @Stanford
1
3
Something all conferences should do
2
3DV program and tutorials are now online. Register now and mark your calenders for. Oct 25-28
The program for #3DV2016 is now available! Very exciting list of speakers, talks, demos, tutorials, and exhibitors! 3dv.stanford.edu/program.htm…
3
2
A very timely workshop on causality and robotics. Consider participating!
Also see a recent piece on Judea Pearl's perspective on causality models in AI quantamagazine.org/to-build-…
Excited about the lineup of speakers to explore causality in robotics.
RSS Workshop on Causal-Imitation. CfP is out and deadline for posters is Jun 3. Consider submitting! @yukez @ermonste @jiajunwu_cs and Michael Laskey
buff.ly/2Lig8K0
1
2
Our work on deep Layout estimation have been picked up and featured on CVPR daily: rsipvision.com/CVPR2016-Wedn…
2
2
JackRabbot has been featured in the Stanford news: news.stanford.edu/2016/06/01…
1
1
2
Elias Boreinboim from Purdue explaining the causal heirarchy from Judea Pearl school of thought at workshop on causality and imitation #RSS2018
2
Stanford’s ‘Jackrabbot’ paves way for social robotics stanforddaily.com/2016/07/31… via @Stanford_Daily
1
2
Attending @cvpr2016? Vote for our 2019 CVPR Bid!
More info at cv-foundation.org/CVPR2019/C…
2
The #3dv2016 website is live! 3dv.stanford.edu/
2
2
Our paper goo.gl/vIYVMJ won the Google runner-up best paper award at the Vision from Satellite to Street 2015 Workshop!
1
1
1
The deadline of early bird registration of 3DV 2016 is approaching (9/16). Do not forget to register here 3dv.stanford.edu/registratio… #3dv2016
2
1
5 papers accepted to #cvpr2016: 2 oral, 2 spotlight, 1 poster!
1
1
#3DV Program is online now. Exciting keynotes and talks starting next Tues 10/25 3dv.stanford.edu/program.htm…
1
Stanford team teaching robots manners abc7news.com/1376025/ via @abc7newsbayarea
1
New work from our group on zero shot transfer in RL for Robotics
1
1
@silviocinguetta adding to the great conversation on ar/vr in vision at @ToyotaResearch workshop
1
1
We got the Stanford CIFE Seed Research Award to conduct research on parsing of buildings. See our latest work on it: buildingparser.stanford.edu
1
3DV 2016 is sold out. We look forward to seeing you all next Tuesday! #3dv2016
1
1
This Monday (10/17) we'll discuss building parsing, its potential and the AEC/FM industry at "The Modern Architect" radio show @KZSU 10-11AM
1
Stanford SVL is a vibrant community of faculty, postdocs and students. Our alumni have landed in prestigious positions in academia and industry. We strive for an integrative and diverse group. We especially encourage applications from traditionally underrepresented groups in AI.
1
1
1,055
BBC News - Jackrabbot: Why this robot is watching how you move bbc.com/news/technology-3683…
1
1
Exciting new datasets for progress in videos nitter.app/animesh_garg/status/78…
2
1