Artificial Intelligence
30 views
AI Safety
Quick Definition
Research ensuring AI is developed safely without harm
Full Definition
Research ensuring AI systems are developed and deployed safely without unintended harm.
Examples
adversarial robustness, alignment, containment
Related Terms
ai-alignment
ai-ethics
More Artificial Intelligence Terms
Federated Learning
Distributed ML training without sharing raw data
Speech Synthesis
Converting written text into natural-sounding spoken audio
Self-Attention
Mechanism allowing sequence positions to attend to all others
In-Context Learning
LLMs learning new tasks from prompt examples without weight updates
AI Benchmark
Standardized tests measuring AI model performance
Reward Modeling
Training models to predict human preferences for RLHF