Stefan Lee

I am an Associate Professor at Oregon State working on problems in embodied AI that combine perception, control, and natural language.

[Hunter L Brown @HunterLBrown]

New preprint 🤖

RL policies that switch between position and force control per-axis for precision assembly. 1,600+ real Franka FR3 trials, zero-shot sim-to-real.

35% higher success and 30% less force than VICES.
1/5

[Alan Fern @AlanPaulFern1]

1/ Introducing Humanoid Hanoi — a distribution over “Tower of Hanoi” box rearrangement problems for humanoids.

https://t.co/jpSe2B6SBi

[Stefan Lee @stefmlee]

RT @fuxinli2: Are you passionate about connecting 3D perception / world modeling with learning-based control for humanoid mobile manipulati…

[AIGSA @ Oregon State @osu_aigsa]

We are excited to announce the 2026 Artificial Intelligence Application Support Program (AIASP) for prospective students in #ArtificialIntelligence applying to AI graduate degree programs at @OregonState! More info in the thread. 🧵1/5

[Akhil @aperinch]

Excited to share our work to appear at #CVPR2025! This work in collaboration with @jacob__krantz and @stefmlee investigates if visual representations (imaginations) of natural language instructions can improve performance of vision-and-language navigation (VLN) agents.
TL;DR-yes

[AIGSA @ Oregon State @osu_aigsa]

We are excited to announce the 2024 Artificial Intelligence Application Support Program (AIASP) for prospective students in #ArtificialIntelligence applying to AI graduate degree programs at @OregonState! More info in the thread. 🧵1/5

[Skand @skand_2021]

I'm happy to present our work titled "Point Cloud Models Improve Visual Robustness in Robotic Learners" tomorrow, May 16 (Thu).

Paper: https://t.co/OOah9WEz07
Project Page: https://t.co/UpGPZvxGzo

Poster session: 10:30 - 12:00 pm (JST)
Oral talk: 16:30 - 18:00 (JST) @ AX-F206

[Mohit Bansal @mohitban47]

🚨 We have postdoc openings at UNC 🙂

Exciting+diverse NLP/CV/ML topics**, freedom to create research agenda, competitive funding, very strong students, many collabs w/ other faculty & universities+companies, superb quality of life/weather. Please apply + help spread the word 🙏

Ping me for any questions!

**multimodal generative models, grounded+embodied semantics, LLM reasoning+planning agents/multi-agent interaction+collaboration, faithfulness+hallucination, interpretability+model editing, efficiency+model merging, generalizability/OOD, etc. (more details of our publications/talks + students/postdocs + labpage, etc. in links below👇)

[Eric Slyman (he/they) @EricSlyman]

I'm happy to present "VLSlice: Interactive Vision-and-Language Slice Discovery" at @ICCVConference this Thursday! Find me 2:30pm - 4:30pm in room "Foyer Sud" - 121. See https://t.co/VoOaKJ0Wes for the paper, code, and presentations. TL;DR in the thread🧵1/8

[AIGSA @ Oregon State @osu_aigsa]

We are excited to announce the 2023 Artificial Intelligence Application Support Program (AIASP) for prospective students in #ArtificialIntelligence applying to AI graduate degree programs at @OregonState! More info in the thread. 🧵1/5
https://t.co/x48HvRbIGc

[Jialu Li @JialuLi96]

🚨Excited to share: “𝗣𝗮𝗻𝗼𝗚𝗲𝗻: Text-Conditioned Panoramic Environment Generation for Vision-and-Language Navigation”!🚨

PanoGen creates diverse 360-degree panorama via recursive outpainting, achieving SotA on R2R, R4R, CVDN.

https://t.co/rhjMKsPVXz

@mohitban47 @uncnlp
🧵

[Franziska Meier @_kainoa_]

The Embodied AI team @MetaAI is looking for postdocs!

We're looking for researchers that are interested in advancing foundation models for Embodied AI and robotics.

If interested, apply here: https://t.co/zTtA1QVC4j

[Thomas G. Dietterich @tdietterich]

Oregon State EECS is recruiting (including in AI/ML). We will soon have the most powerful academic computing facility in the US thanks to a gift from Jen-Hsun and Lori Huang. https://t.co/8SR7WHPF1G

[Jacob Krantz @jacob__krantz]

Language-guided navigators shouldn’t be carted off to a new environment with each new instruction. Instead, they should persist & improve over time.

We present Iterative Vision-and-Language Navigation to evaluate exactly this!
https://t.co/N8fOkGyCqa
https://t.co/ocAAcT1cYm

[International Conference on 3D Vision (3DV) @3DVconf]

Rebuttal period is coming in a few days!
Check these resources to make the most out of your rebuttal:

@Awfidius at CVPR20:
- slides: https://t.co/SUOvqQ75kN
- video: https://t.co/iFhXRWKFLC

@deviparikh, @DhruvBatraDB and @stefmlee blog post: https://t.co/wEqkHP8jmN

[Aishwarya Agrawal @aagrawalAA]

To the V&L community: should we rethink current evaluation practices in VQA?

We conducted a systematic study of recent pretrained V&L models for OOD generalization in VQA.

Check out the findings!

[Stefan Lee @stefmlee]

@YicongHong @panderson_me Permission to use this? It is awesome.

[Jacob Krantz @jacob__krantz]

Thrilled to announce the 2nd RxR-Habitat Challenge at the #CVPR2022 Embodied AI Workshop!

Agents must navigate indoor environments following human instructions. The environments are realistic, the instructions multilingual.

Is your agent up for the task?
https://t.co/zaOyTxuUoQ

[Karthika Mohan @Carthica]

I welcome PhD applications from students interested in working with me on topics at the intersection of Artificial Intelligence and Causal Inference. Application deadline: 15th Dec 2021. Link: https://t.co/10Kipo21BK