BOID smart scheduling tool : CLICK HERE

Our BOID scheduling system implements a novel hybrid architecture that combines logic-based reasoning with embedding models to automatically mine decision rules from user data, eliminating the traditional dependency on domain experts in logic implementations. The system uses E5-large-v2 embeddings to extract obligations from Google Scholar profiles and desires from user-selected keywords, matching them to conference topics via cosine similarity thresholds. Process trees (implemented via the Tweety library) calculate extensions using prioritized default rules, while WebSockets enable real-time bidirectional communication for resolving branch splits through human input.

Key Features:

• No expert required – Embedding models automatically mine BOID rules from existing user data (Scholar profiles, keyword selections)
• Fully interpretable – Every scheduling decision traces back to its originating mental attitude (Belief, Obligation, Intention, or Desire)
• Hybrid intelligence – Combines rule-based reasoning, semantic similarity matching, and human-in-the-loop decision making
• Real-time collaboration – WebSockets let users resolve conflicts interactively when process trees split
• Flexible prioritization – Adjustable priority gaps between mental attitudes enable different agent types (social, realistic, selfish)