Technical standards and communication protocols.

Communication Protocols

• Model Context Protocol - Standard for AI systems to access external data
• OpenAI API - REST API format used by many AI providers
• gRPC - High-performance protocol for AI service communication
• WebSocket - Real-time communication for AI chat interfaces

Model Standards

• ONNX - Open standard for representing AI models
• MLflow - Standard for tracking ML experiments and models
• Model Cards - Documentation standard for AI models
• Hugging Face Hub - De facto standard for sharing AI models

Data Formats

• JSON-LD - Structured format for AI training data
• Common Crawl - Standard web crawl format used for training
• OpenImages - Standard format for computer vision datasets
• CoNLL - Standard format for natural language processing data

Evaluation Standards

• BLEU Score - Standard metric for translation quality
• ROUGE Score - Standard metric for text summarization
• Perplexity - Standard metric for language model quality
• Benchmarks - Standard tests for comparing AI performance

Security Protocols

• Differential Privacy - Standard for protecting individual data
• Federated Learning - Standard for distributed AI training
• Secure Aggregation - Standard for combining model updates safely
• Homomorphic Encryption - Standard for computing on encrypted data

← Back to AI Glossary