Odyssey AI Private Language Model (PLM)
Security and Privacy
For security and privacy, each Odyssey AI instance includes its own privately hosted Private Language Model (PLM) (e.g., Llama3). These models are pre-trained and utilized as is in their general-purpose ("vanilla") form. The default language model can be changed upon request by the instance owner for an additional cost, ensuring flexibility while maintaining strict privacy standards.
Syncing New Data within Odyssey AI
Odyssey AI automatically synchronizes new data into its vectorized knowledge base using an advanced ingestion process. This process leverages proprietary algorithms: DocuTableAI™ extracts data from documents or multi-modal files, Topic Models analyze the content to identify key topics, and InteliCompress™ compresses the data, retaining essential information while discarding noise. This occurs automatically whenever a new file or document is added to the data sources integrated with Odyssey AI, keeping the system current without altering the underlying PLM parameters.
Fine-Tuning for Enterprise Customers
For Enterprise customers seeking enhanced performance within specific domains, Odyssey AI offers the option to fine-tune the PLM based on their unique data and project requirements. This creates a customized, domain-adapted model exclusive to their Odyssey AI instance, improving accuracy and relevance for specialized applications. Fine-tuning requires dedicated compute resources and periodic re-training to ensure the model remains effective as domain-specific data evolves. This premium service is available to large Enterprise customers for an additional fee, with pricing and details tailored during an engagement meeting. Contact us for further information.