NuRisk: Visual Question Answering for Autonomous Driving Risk Assessment

TUM-AVS created the NuRisk dataset for agent-level risk assessment in autonomous driving. It contains visual question-answering pairs based on Bird's-Eye View images. The dataset was last updated on July 9, 2026.

Use Cases

• Training visual question-answering models for driving scenarios based on Bird's-Eye View images.
• Benchmarking risk assessment algorithms for autonomous vehicles based on image-question pairs.
• Developing multimodal AI agents that can interpret driving scenes and answer safety-related questions.

Strengths

• Focuses specifically on risk assessment for autonomous driving, a critical application area.
• Provides Bird's-Eye View images paired with driving-related questions and answers, creating a structured multimodal task.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

TUM-AVS

Freshness

Last updated 2026-07-09 11:28:34.