Reproduce “Scenic: A Language for Scenario Specification and Data Generation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Daniel J. Fremont, Edward Kim, Tommaso Dreossi, Shromona Ghosh, Xiangyu Yue, Alberto L. Sangiovanni-Vincentelli, Sanjit A. Seshia
Scenic is a probabilistic programming language for generating specified physical scenes and agent behaviors. It combines feature distributions and constraints to sample scenarios for simulation, testing, or training. In a fixed-size SqueezeDet experiment, replacing 5% of training images with targeted partial-occlusion images raised occlusion-set precision from 62.8 ± 6.1% to 68.9 ± 3.2%, with Matrix-set precision nearly unchanged; both test sets used synthetic GTAV images.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.