Zackory Erickson (@ZackoryErickson) profile picture

Assistant Professor, Robotics Institute at CMU. Physical HRI, assistive robotics, multimodal perception, simulation, mobile manipulation. he/him

Pittsburgh, PA
Joined August 2013
Zackory Erickson retweeted
How far can we push dexterous robot manipulation with human video-only supervision and minimal assumptions? 🚫 No teleop. 🚫 No wearables. 🚫 No external sensors. 🚫 No robot demos. Introducing VIDEOMANIP: 🎥 Just monocular RGB, 🌍 in-the-wild human video → dexterous robot manipulation 🤚[1/6]
6
34
9
167
32,169
Zackory Erickson retweeted
Introducing FMVP: a method that adapts to natural arm motions during robot-assisted dressing. Pre-trained on vision in sim, fine-tuned with limited real-world vision+force data, and tested in a 12-user, 264-trial study, FMVP is robust across garments and motions. #CoRL2025
1
15
3
38
8,056
Zackory Erickson retweeted
How do we discover a robot's failure modes before deploying it in the real world? Standard benchmarks often don't capture the full picture, leaving policies vulnerable to plausible variations in object shape. Thrilled that our work, "Geometric Red-Teaming for Robotic Manipulation," has been accepted as an oral presentation at #CoRL2025! We introduce a framework to automatically find these geometric blindspots. georedteam.github.io 🧵
3
14
4
46
9,914
Zackory Erickson retweeted
🚨🚨New paper: if you want robots to do bimanual long-horizon tasks well, try RaC: a human in-the-loop data collection protocol that naturally amplifies Recovery & Correction behaviors + trains on them. 📈 data efficiency 10x vs prior results (+many nice properties). 🧵⬇️
5
52
3
267
25,935
Zackory Erickson retweeted
Introducing ArticuBot🤖at #RSS2025, in which we learn a single policy for manipulating diverse articulated objects across 3 robot embodiments in different labs, kitchens & lounges, achieved via large-scale simulation and hierarchical imitation learning. articubot.github.io 🧵
3
33
3
89
7,763
Zackory Erickson retweeted
[1/7] Teaching dexterous robot hands to perform functional grasps usually needs hours of teleoperation, manual labeling, or pre-scanning object meshes. Not anymore. 🔥We are excited to introduce Web2Grasp that learns functional multi-finger grasps straight from web images of human hand-object interactions (HOI). No human demos. No object scans. Just web images. 👉web2grasp.github.io/ @CMU_Robotics @CarnegieMellon
1
26
4
76
9,270
Zackory Erickson retweeted
Happy to announce our paper on LAMS: LLM-Driven Automatic Mode Switching at HRI 2025! Our research introduces the use of off-the-shelf LLMs for automatic mode switching in robot teleoperation. 📄Paper: arxiv.org/abs/2501.08558 (1/4)
1
1
1
218
VoicePilot, the first work to involve human subjects interacting directly with an LLM-integrated physically assistive robot, was accepted to UIST 2024! sites.google.com/andrew.cmu.…
4
13
5
85
14,896
CMU experts in human-robot interaction propose VoicePilot: a framework + design guidelines for incorporating LLMs as speech interfaces for physically assistive robots 🤖🦾 The paper was accepted at @ACMUIST - held in Pittsburgh this October! Read: ri.cmu.edu/voicepilot-framew…
2
8
1
44
3,938
Zackory Erickson retweeted
How to build a safe, generalizable robot-assisted dressing system? Our #RAL2024 paper introduces a novel multi-modal learning method, combining a simulation-trained vision policy and a real-world-learned force dynamics model, with strong results in a 10-participant human study.🧵
1
8
21
2,308
Our research won a Best Paper Award at HRI 2024! Henry Evans, an individual with quadriplegia, used our wearable HAT (Head-Worn Assistive Teleoperation) interface with @hellorobot Stretch in a 7-day in-home study to complete 32 self-care/household tasks. sites.google.com/view/hat2-t…
1
7
23
1,252
Zackory Erickson retweeted
Can we eliminate human effort for reward engineering in RL? Introducing RL-VLM-F, which uses preference feedback from VLMs to automatically generate rewards to learn new tasks, using only a text description of the task and the agent's visual observations! arxiv.org/abs/2402.03681
2
38
3
129
37,978
Is your robot 🤖 safe around humans🚸? Or will it fail catastrophically🤕? Excited to introduce "Natural Adversarial Frontier", a framework for probing the robustness of Human-Robot Interactions. To appear at #CoRL2023 🧵(1/8) Joint work with @ancadianadragan @daniel_s_brown @ZackoryErickson arxiv.org/abs/2310.10610 ood-human.github.io
1
10
1
51
11,269
New research from @CMU_Robotics Institute aims to increase autonomy for individuals with motor impairments by introducing a head-worn device that will help them control a mobile manipulator. https://xitter.gq/t.co/L23SyiEYCu
9
2
24
3,220