> For the complete documentation index, see [llms.txt](https://azen.gitbook.io/whitepaper/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://azen.gitbook.io/whitepaper/azen-protocol/key-features.md).

# Key features

🔹Dynamic Fractional NFTs (dfNFTs) for Computation, AI & Applications

* Converts computing resources, AI models, and applications into dfNFTs, enabling fractional ownership, leasing, and efficient execution.
* Provides interoperability across decentralized computing environments, ensuring seamless deployment of AI-powered workloads.

🔹Smart Contract-Driven Resource Orchestration

* Automates the allocation and execution of AI, applications, and computational tasks.
* Uses predictive AI-based scheduling to enhance resource utilization and reduce inefficiencies. 

🔹Modular Computation & AI Integration

* Supports decentralized execution of AI inference, model training, and dApp services.
* Allows applications to access distributed computing resources on a pay-per-use basis.

🔹Cloud-to-Edge AI Computing & Execution

* Enables AI & application workloads to dynamically shift between cloud, fog, and edge layers.
* Reduces latency for real-time AI model execution and decentralized application processing.

🔹AI-Powered Security & Privacy Enhancements

* Use zero-knowledge proof (ZKP) and homomorphic encryption for secure computation.
* AI-driven anomaly detection enhances trust and security in decentralized networks.

🔹Decentralized AI, Application & Computation Marketplace

* Users can monetize computing, AI models, and application services.
* Smart contracts facilitate trustless transactions and execution guarantees.

🔹Decentralized Data Collection & Processing for AI

* Data sources are tokenized and secured via smart contracts, ensuring privacy and trust.
* AI models access real-time, distributed data for enhanced learning and adaptive intelligence.
* Supports privacy-preserving machine learning (PPML), allowing AI models to train without compromising user data security.