DecideAI
DecideAI Price Converter
DecideAI Information
DecideAI Markets
DecideAI Supported Platforms
| DCD | ICRC1 | ICP | xsi2v-cyaaa-aaaaq-aabfq-cai | 2024-04-12 |
About DecideAI
DecideAI (DCD) is an artificial intelligence (AI) ecosystem designed to facilitate the development, refinement, and deployment of high-performance, specialised Large Language Models (LLMs). The platform addresses the limitations of general-purpose, centralised AI models by focusing on targeted, domain-specific solutions across industries such as healthcare, finance, and media.
The DecideAI ecosystem consists of three core components:
- Decide Protocol: A platform that coordinates human and artificial intelligence to annotate, train, and continuously refine specialised LLMs using Reinforcement Learning with Human Feedback (RLHF).
- Decide ID: A verification system employing Proof of Personhood (PoP) to ensure data contributors are authentic and credentialed, enhancing the quality and traceability of data.
- Decide Cortex: A platform providing access to pre-trained LLMs and curated datasets for developers and organisations via direct purchase or API.
The DCD token functions as the native utility token within the DecideAI ecosystem, supporting a range of activities:
Incentivising Participation: Contributors are rewarded in DCD for providing high-quality data annotations, participating in model refinement, and enhancing ecosystem infrastructure.
Funding Innovation: Developers and researchers receive DCD incentives for creating new AI models, improving datasets, and developing external applications using DecideAI infrastructure.
Ecosystem Transactions: DCD is used for transactions within the ecosystem, including accessing AI models, datasets, and computation resources.
Collaboration and Governance: DCD supports decentralised collaboration, allowing stakeholders to contribute to ecosystem growth and sustainability.
Decide Protocol:
- Facilitates the collection, annotation, and refinement of datasets for LLM training.
- Employs RLHF, integrating human feedback for model accuracy and domain-specific customisation.
- Incentivises contributors via a reward system using DCD tokens.
Decide ID:
- Utilises a Proof of Personhood (PoP) methodology to verify contributors' identities and credentials.
- Ensures that data contributors are unique, credentialed, and traceable.
- Employs Zero-Knowledge Proofs (ZKPs) for privacy-preserving verification.
Decide Cortex:
- Provides access to pre-trained LLMs and curated datasets for direct use or integration via APIs.
- Supports model customisation and post-deployment management.