SISL
@SISLaboratory
Followers
1K
Following
128
Media
114
Statuses
375
Stanford Intelligent Systems Laboratory, Aeronautics & Astronautics Department. Advancing Research on Autonomous Systems and Decision Making Under Uncertainty.
Stanford, CA
Joined September 2015
Great collaboration between the @Stanford SISL and @NASAAmes with Grace Ra Kim, Hailey Warner, Duncan Eddy, Evan Astle, Zachary Booth, Edward Balaban, and @aiprof_mykel. Paper Link: https://t.co/xNGIXsofk0 Check them out at ISPARO https://t.co/skWCqI0FH2!
0
0
1
Using the Enceladus Orbilander as a case study, the novel approach cuts sample identification errors by nearly 40%, even in off-nominal scenarios. The key: offline-validated policies that still adapt in real-time.
1
0
0
[New SISL paper 📣 ] Autonomous science operations for deep space to be presented at ISPARO in a few days! SISLers Grace Kim et al. develop a decision-making framework that lets spacecraft independently run science campaigns under extreme communication delays.
1
1
2
[New SISL webinar 📚 ] Safety validation methods for AI in high-stakes environments by long-time SISLer Sydney Katz! Whether you're building autonomous systems, deploying AI in regulated industries, or evaluating these technologies, this covers the essentials.
1
2
3
4/4 Website: https://t.co/BS2J3UxCE6 Paper: https://t.co/GRzZmPqOT7 Youtube: https://t.co/sHLR4e3OFX Bernard Lange, Anil Yildiz, Mansur Maturidi Arief, Shehryar Khattak, @aiprof_mykel, Georgios Georgakis
0
0
0
3/4 ... answering, goal-based navigation, text path-following, to open-ended exploration. Beyond the technical results, this work offers a practical blueprint for building modular, reasoning-driven autonomy that integrates cleanly with existing robotic stacks.
1
0
0
2/4 The core innovation is embedding an LVLM agent that dynamically orchestrates perception, reasoning, and navigation modules, breaking away from rigid, task-specific pipelines. ARNA demonstrates genuine zero-shot generalization across diverse tasks from embodied question ...
1
0
0
1/4 [📣 New SISL preprint] Introducing ARNA: a general-purpose navigation system that handles exploration, goal-reaching, text-following, and more without task-specific training via LVLM-orchestrated perception, reasoning, and dynamic action moduls (instead of a fixed pipeline).
1
1
2
4/4 We show how the distance allows for simple gradient-based solutions in distribution steering and ergodic control. Work by: Alexandros Tzikas, Arec Jamgochian, Nazim Kemal Ure, @aiprof_mykel, Stephen P. Boyd https://t.co/fzdy7iZCXM
0
0
1
3/4 ...functions of random linear one-dimensional projections of the random variables. Our proposed distance is interpretable, computationally simple, and admits a differentiable approximation.
1
0
0
2/4 Computing the similarity between two probability distributions is a recurring theme across control. We introduce a unified family of distances between the probability distributions of two random variables that is based on the discrepancy between the cumulative distribution...
1
0
0
1/4 [New SISL preprint 🚀] How do you measure the difference between two probability distributions? Alexandros Tzikas et al. propose a simple, interpretable method based on random projections that's easy to compute and optimize with applications to control and steering
1
1
4
Check out SISL's @aiprof_mykel on NBC Bay Area talking about Waymo's driverless cars extending their service to highways: https://t.co/bclkhKX44b
0
2
2
Joint publication between Stanford Intelligent Systems Laboratory and MIT Lincoln Laboratory: Sydney Katz, Robert Moss, Dylan Asmar, Wes Olson, Jim Kuchar, and @aiprof_mykel Link:
arxiv.org
Aircraft collision avoidance systems is critical to modern aviation. These systems are designed to predict potential collisions between aircraft and recommend appropriate avoidance actions....
0
0
0
These challenges are present in a variety of other domains as well, so this research can provide valuable insights for a wide range of safety-critical systems!
1
0
0
The article summarizes decades of research in aircraft collision avoidance! The aircraft collision avoidance problem presents technological challenges in surveillance, decision making, and validation, which has resulted in a breadth of interesting research in these areas.
1
0
0
Check out the latest @SISLaboratory work: a deep dive into aircraft collision avoidance systems. "Aircraft Collision Avoidance Systems: Technological Challenges and Solutions on the Path to Regulatory Acceptance"
1
1
1
Good morning Suzhou! @amelia_f_hardy and I will be at @emnlpmeeting to present our work *TODAY, Hall C, 12:30PM; paper number 426* Come learn: ✅ why likelihood is important to simultaneously optimize with attack success ✅ online preference learning tricks for LM falsification
New Paper Day! For EMNLP findings—in LM red-teaming, we show you have to optimize for **both** perplexity and toxicity for high-probability, hard to filter, and natural attacks!
1
9
18