Out some of their latest work, see these papers: -Diffusion for multi-agent systems: arxiv.org/abs/2307.06244 -Diffusion for robotics arxiv.org/abs/2409.15528 -VLMs for CAD Code Design: arxiv.org/pdf/2410.05340 -LLMs for Alignment in Planning arxiv.org/abs/2208.08374

Delighted to cap this year off holding @Seancye and @Pradyumna_T. Their legacy for the @core_robotics is immense, pioneering work on generative AI (diffusion models and LLMs). Your leadership will be missed. @zoox and Toyota Research Institute are lucky to have you! To check…

How can everyday users teach robots to perform complex, long-horizon tasks (e.g., meal prep) without relying on robotics experts for help? Our latest study published in #IJRR tackles this challenge, paving the way for more accessible human-robot collaboration. A 🧵...

Congrats to Sam Yi Ting, Dr. Erin Botti, and Manisha Natarajan from the @core_robotics who are tremendous researchers and who helped me pursue this @NASA Early Career Funded research. 6/6 @GTrobotics @gtcomputing

In future work, the @core_robotics will be developing adaptive robots that adjust to human stress levels and modulate workload to max performance in safety-critical settings. Here's an AAMAS'24 paper showing one way to modulate latent human variables: ifaamas.org/Proceedings/aama… 5/n

The *right* kind of stress can boost performance (to a point) while workload degrades it. Stress actually has 2 components: distress and eustress, which have differential effects. Overall, our results show that AI systems need to dynamically adapt based on all these factors. 4/n

24 participants, multiple stress (s) and workload (w) conditions. The result? 70% performed better under moderate stress -- predicted in rat studies by the Yerkes-Dodson Law (YDL). However, performance declined as workload increased. 3/n

Robot teleoperation is vital in safety-critical applications, such as space exploration and surgery. Understanding how stress and workload impact performance can drive safer, more efficient systems. 2/n c.f. researchgate.net/publication…

Can stress improve human performance in robot teleoperation? We found that moderate stress enhances effectiveness in critical tasks, but high workload? It takes a toll. Let’s dive into the findings...🧵1/6 Link: ieeexplore.ieee.org/document…

“Trust and Dependence on Robotic Decision Support” has been published in @IeeeTro! Available here: ieeexplore.ieee.org/document… Bravo to Manisha Natarajan on her strong experimental work to inform how we design for appropriate trust in robotic systems. @core_robotics

Interested in how else diffusion models can be leveraged for robot model learning? Check out more papers by @Seancye et al. from @core_robotics. Check them out: arxiv.org/abs/2409.15528 arxiv.org/pdf/2403.10794 ieeexplore.ieee.org/stampPDF…

Diffusion models can forecast trajectories, which can help with predicting aircraft flight paths to helping robots generate motion plans in complex environments. Check out our latest work on this topic that Dr. Sean Ye presented #IROS2024! Teaser Talk: tiny.cc/nwjrzz

We are hiring tenure-track faculty in Robotics, HRI, AI/ML, and Graphics at @ICatGT! I am grateful for the collaborative, supportive ecosystem of amazing faculty and students in @ICatGT and strongly encourage you to apply to join us! ic.gatech.edu/faculty-hiring