MINI CHALLENGE - "1h+ co-creation session with a conference Speaker" | DREAM KG: AI knowledge for the homeless

The Critical Problem: Dynamic Information Crisis in Homelessness Support Every day, homeless service providers face a devastating reality: the information they rely on to help people experiencing homelessness (PEH) is outdated, incomplete, or simply wrong. A shelter that appears available in the morning may be full by noon. A mental health service listed in the directory may have closed months ago. Emergency housing resources change locations, update eligibility requirements, or modify their operating hours without centralized notification systems. This information crisis creates cascading failures throughout the Continuum of Care. Case workers waste precious time calling disconnected numbers. People experiencing homelessness are sent to non-existent resources, eroding trust in the system. Emergency responders lack real-time data about available beds during critical interventions. The result? Vulnerable individuals fall through cracks that could have been closed with accurate, up-to-date information. The Mini Challenge: Explore, Detect, and Prototype Solutions This DREAM-KG mini challenge focuses on exploration and rapid prototyping of automated error detection and correction systems. Over the course of this focused hackathon, teams will: - Explore and analyze existing knowledge graphs to identify patterns of data inconsistencies, outdated information, and structural weaknesses - Design correction mechanisms that can flag suspicious data points and suggest automated fixes - Bonus: Build proof-of-concept demonstrations showing how these systems could work in real-world scenarios The goal isn't to build production-ready systems, but to explore the problem space and create compelling prototypes that demonstrate innovative approaches to automated knowledge graph maintenance. Focus Areas for Rapid Prototyping Teams can choose to focus on one or more of these specific error detection challenges: - Temporal inconsistencies: Detecting when service hours, availability, or contact information appears outdated - Cross-reference validation: Identifying conflicts between different data sources about the same service - Pattern recognition: Finding anomalies in service capacity, referral success rates, or resource utilization - Relationship mapping: Discovering broken or missing connections between related services and resources Available Resources and Quick Start Teams will have immediate access to: - Sample knowledge graph datasets from the DREAM-KG platform with real homeless service data - Pre-built graph structures to jump-start exploration and prototyping - Documentation and APIs for rapid development without complex setup - Mentorship from the DREAM-KG research team for technical guidance What We're Looking For Successful mini challenge submissions will demonstrate: - Creative exploration of the knowledge graph data revealing insights about information quality issues - Innovative prototypes for automated error detection, even if not fully implemented - Clear demonstrations of how the proposed solutions would work in practice - Technical feasibility showing a path from prototype to real-world application Team Composition for Rapid Development Ideal teams combine: - Data scientists/analysts for exploring and understanding the knowledge graph structure - Software developers (Python preferred) for building detection algorithms and prototypes - Visualization specialists for creating compelling demonstrations of findings and solutions - Domain knowledge in social services or public health for contextual insights Beyond the Mini Challenge While this is a focused exploration and prototyping challenge, strong teams and innovative approaches may have opportunities for: - Extended collaboration with the DREAM-KG research team - Further development of prototypes into full research projects - Real-world testing with partner organizations in the homelessness support ecosystem Getting Started All datasets, technical documentation, and platform access are available at: - DREAM-KG Platform: https://dreamkg.com/ - Technical Resources: https://www.proto-okn.net/wp-content/uploads/2024/05/DreamKG-Datasheet.pdf This mini challenge offers a unique opportunity to dive deep into real-world knowledge graphs, prototype cutting-edge solutions for data quality issues, and contribute to technology that could directly impact homelessness support systems nationwide.