The workshop "ROAM" aims to gather researchers and engineers from academia and industry to discuss the latest advances in Out-of-Distribution (OOD) perception for autonomous driving. The motivation behind this workshop is threefold:
New Benchmark Driving Dataset: Introducing a new large-scale benchmark driving dataset acquired from diverse cities in South America, particularly Peru. This dataset provides a wide range of dangerous driving situations and O.O.D objects that are not found in existing datasets.
Challenges for OOD Object Detection and Classification: Hosting several challenges for OOD object detection and classification using the dataset and crowd-sourced data.
Exploring Multi-Modal Models: Hosting invited speakers and a workshop paper submission track encouraging the exploration of the opportunities and risks of using foundation multi-modal models for ODD detection in the context of autonomous driving across under-represented cities.
The workshop invites researchers from various fields such as NeuroAI, Robustness, Philosophy/Morality of Self-Driving (with focus on Vision + Decision-Making), and NLP(LLM), to foster the development of new ideas for the Self-Driving Community. The workshop will feature presentations of the top scoring papers as both posters and short talks (Oral), with regular accepted papers presented as Posters. Papers must adhere to a maximum submission length of 5 pages with one extra page-post rebuttal to incorporate reviewer feedback. Acceptance criteria for papers is an average score of 5.5 or above (from 1-10). No rebuttal period will be done during the submission process.