Insights
The research output of the Spark AI consortium: executive briefings, working papers, and points of view
Working Papers
Read and download with a free Spark AI account.
SCORE-AI: A Human-Compatible Framework for AI Agent Team Orchestration
James Massa
A framework for the Selection, Chaining, Orchestration and Ranking of expert AI agents: recruiting from a registry, vetting through tryouts, onboarding, and evaluating across eight coordination patterns.
Read with a free accountAI-Native Drug Discovery and Development: How Semantic Intelligence, Foundation Models and Autonomous Research Systems Are Re-shaping Pharmaceutical Drug Discovery and Development
James Short, Suresh Mani, Raju Pusapati and Murali Krishnam
Competitive advantage in pharmaceutical R&D increasingly rests on integrating high-quality data, automated experimentation, and iterative model development into one capability that compounds over time.
Read with a free accountAI-Native Drug Discovery and Development: The Convergence of Semantic Biomedical Intelligence, Foundation Models, and Autonomous Research Platforms
James Short, Raju Pusapati, Suresh Mani and Murali Krishnam
How multimodal biomedical data, foundation models, and closed-loop autonomous laboratories are turning therapeutic discovery from a heuristic process into a predictive, data-driven one.
Read with a free accountBuilding Intelligent Foundations: Rethinking Storage Strategy in the Age of AI
James Short
How organizations are responding to AI-driven data growth, drawn from interviews with technology and data leaders. Covers predictive workload modelling, automated data-tiering, and self-optimizing storage.
Read with a free accountScientifically Automating Data Quality Decisions with AI Explainability Weights
James Massa and James Short
Introduces Explainability Driven Data Quality (EDDQ), a methodology that uses machine learning explainability weights to automate data quality management decisions end to end: catalog creation, Key Data Element identification, lineage, change management, validation, and issue management.
Read with a free accountProject Chainwatch: A User Configurable Anomaly Detection System for Monitoring Blockchain Network Data Flows
Samson Qian, Steve Orrin, Matt Clanton and James Short
The motivation, design, testing and implementation of ChainWatch, an experimental system that detects anomalies in blockchain network data flows using realistic network simulations and machine learning.
Read with a free accountMeeting Agentic AI's Data Quality Needs with Zero Trust Data Quality
James Massa and James Short
Introduces Zero Trust Data Quality (ZTDQ), a framework that applies zero trust principles from cybersecurity to data quality assessment. Data quality should always be verified, never assumed.
Read with a free accountIs AI Governable? Industry Perspectives on the Adoption, Effectiveness and Accountability of Frontier AI
James Short
A cross-sector analysis of how industry and public sector leaders are developing governance policies and accountability for frontier AI, and the infrastructure and standards taking shape around them.
Read with a free accountPoints of View
Practitioner perspectives from consortium members. Open to everyone.
Architecting Enterprise AI for Scale, Control, and Competitive Advantage
John Ottman and Suresh Mani, Solix
How enterprises can move beyond AI hype to deliver measurable value at scale: architecture patterns for control, governance, and competitive advantage in production AI systems.
Read the paperRethinking Enterprise Storage Architecture for AI-Driven Systems
Barry Rudolph, Spectra and Christine Telford Teale, IBM
How AI workloads are reshaping enterprise storage requirements, from data ingestion pipelines to model training infrastructure and inference serving.
Read the paperBlockLAB
Peer-reviewed publications from SDSC's blockchain research group, on consensus, governance, and privacy-preserving data frameworks.
A Framework Proposal for Blockchain-Based Scientific Publishing Using Shared Governance
Frontiers in Blockchain (2019) — View →
Examining the Potential of Blockchain Technology to Meet the Needs of 21st-Century Japanese Health Care
J. Medical Internet Research (2020) — View →
Combating Healthcare Fraud and Abuse: A Technology Framework Leveraging Blockchain
J. Medical Internet Research (2020) — View →
hOCBS: A privacy-preserving blockchain framework for healthcare data
Information Processing & Management (2021) — View →
Establishing a blockchain-enabled Indigenous data sovereignty framework for genomic data
Cell 185 (2022)
A Field Test of a Federated Learning/Federated Analytics Blockchain Network in an HPC Environment
Frontiers in Blockchain (2022) — View →
Publish with Spark
Spark members can publish their own Point of View or Executive Briefing through the consortium — a channel no other university-based AI center offers. Submissions are reviewed as they arrive rather than on a fixed calendar. Send a draft or an outline and we will come back to you.
Submit a POV or Executive BriefingGet the Research
Points of view are open to everyone. A free Spark AI account unlocks the working papers and gets you new research as it publishes. Forum members and sponsors get member-only briefings and early access to work in progress.