Autonomous Vehicle Corner Case Verification With Virtual Datasets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in handling corner cases due to limited datasets, leading to potential errors and anomalies in AI/ML models.
Innovation Solution
A method and system for implementing formal verification of corner cases in AI/ML mechanisms for autonomous vehicles, involving the creation of a virtual dataset through energetic neural processes that separate and recombine context and content data to synthesize new samples, and subsequent formal verification to enhance model performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If rigorous testing and data collection are performed to address corner cases, then the robustness and reliability of AI models improve, but the time and resources required for development increase
Solution Approach 1:
The patent applies preliminary action by performing formal verification of AI models before deployment to autonomous vehicles. The system verifies models against synthesized corner case scenarios in advance, ensuring robustness and reliability are established prior to real-world deployment, thereby reducing the need for extensive post-deployment testing and iteration.
2Productivity
If AI models are trained on limited datasets, then training time and computational resources are reduced, but the model's ability to handle corner cases deteriorates
Solution Approach 1:
The patent applies copying by creating synthesized copies of corner case scenarios through formal verification processes. The system generates virtual corner case data by transforming and augmenting limited real-world corner case examples, creating multiple synthetic variations that expand the training dataset without requiring proportional increases in real data collection.
3Ease of operation
If traditional testing methods are used to identify corner cases, then the testing process is simple to implement, but the coverage and detection capability of rare scenarios are insufficient
Solution Approach 1:
The patent introduces an intermediary formal verification layer between traditional testing and AI model deployment. This intermediary system uses synthesized corner case scenarios and formal methods to bridge the gap between simple traditional testing and comprehensive corner case coverage, enabling systematic detection of rare scenarios that traditional methods would miss.
Data Source
AI summary
Systems and methods are provided that implement virtual dataset creation and formal verification for artificial Intelligence (AI)/Machine Learning (ML) models in a manner that improves the performance of AI/ML models when encountering corner cases. For example, a corner case correction system is configured to synthesize new samples by recontextualizing samples related to corner cases, in new environments. The corner case correction system can implement an energetic neural process which separates input data into contextual features and content, and then recombines the context and content to synthesize new samples, generating a virtual dataset. The energetic neural process also performs a formal verification of the AI/ML models to address noise in the virtual dataset. The AI/ML model is trained using the virtual dataset, and an updated AI/ML model is created to execute predictive analysis for the corner cases associated with driving environments of autonomous vehicles.