Statement of Martin M Ferber, Associate Director National Security and International Affairs Division Before the Task Force on Inventory Management Senate Committee on Armed Services on Problems in Accountability and Security of DOD Supply Inventories
Abstract
Scene graph generation models understand the scene through object and predicate recognition, but are prone to mistakes due to the challenges of perception in the wild. Perception errors often lead to non-sensical compositions in the output scene graph, which do not follow real-world rules and patterns, and can be corrected using commonsense knowledge. We propose the first method to acquire visual commonsense such as axB;ordnance and intuitive physics automatically from data, and use that to improve the robustness of scene understanding. To this end, we extend Transformer models to incorporate the structure of scene graphs, and train our Global-Local Attention Transformer on a scene graph corpus. Once trained, our model can be applied on any scene graph generation model and correct its obvious mistakes, resulting in more semantically plausible scene graphs. Through extensive experiments, we show our model learns commonsense better than any alternative, and improves the accuracy of state-of-the-art scene graph generation methods.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 23, 1986
- Accession Number
- AD1122911
Entities
People
- Martin M. Ferber
Organizations
- United States Government Accountability Office