Maximilian Igl's Avatar

Maximilian Igl

@maxigl.bsky.social

RS at Nvidia focussing on autonomous vehicles, simulation and RL. Opinions my own and do not represent those of my employer, Nvidia.

82 Followers  |  97 Following  |  2 Posts  |  Joined: 22.12.2024
Posts Following

Posts by Maximilian Igl (@maxigl.bsky.social)

I guess it also depends on the domain and action space. E.g. even in AV I'd expect Gaussians on yaw/vel to be enough, but for juke/jerk we probably want multimodal distributions.

03.01.2025 13:52 — 👍 0    🔁 0    💬 0    📌 0

In my (very limited) experience: i) You can get multi-modality on trajectory level even with unimodal Gaussians, so actions being MM isn't as important. ii) When I was using a GMM, nearly all modes collapsed and it was basically a Gaussian + slightly more robust to outliers.

03.01.2025 07:26 — 👍 1    🔁 0    💬 2    📌 0