Palantir, Quid, RecordedFuture: Augmented Intelligence Frontier
73 likes9,285 views
The document explores the evolution of augmented intelligence technologies, highlighting the capabilities of systems like Palantir and Quid in processing vast amounts of data to enhance decision-making. It details the integration of human input with machine analysis to generate insights, track events, and predict future occurrences through advanced data extraction and visualization techniques. The text emphasizes the importance of adaptive systems that learn from user interactions and preferences, ultimately becoming more attuned to individual user needs.
![• Tracks myriads of data points [series of events]
from the public Web and private data sources
• Computes links and predicts the future [series
of events]
• The user than interacts with the data directly
and gets insights about what might happen](https://image.slidesharecdn.com/palantirquidrecordedfuture-aifrontier-150824204953-lva1-app6892/85/Palantir-Quid-RecordedFuture-Augmented-Intelligence-Frontier-11-320.jpg)
Palantir, Quid, RecordedFuture: Augmented Intelligence Frontier
- 1. FRONTIER OF AUGMENTED INTELLIGENCE What’s next after Palantir, Quid, and Recorded Future
- 6. UNEXPECTED RESULT STRONG HUMAN + MACHINE + INFERIOR PROCESS WEAK HUMAN + MACHINE + BETTER PROCESS >
- 9. AUGMENTED INTELLIGENCE 1.0 WEAK HUMAN + MACHINE + BETTER PROCESS
- 10. • Allows enterprise to define a set of things • Computes links between these things by analyzing text, metadata, relational data, etc. • The user then interacts with the graph directly
- 11. • Tracks myriads of data points [series of events] from the public Web and private data sources • Computes links and predicts the future [series of events] • The user than interacts with the data directly and gets insights about what might happen
- 12. • Allows the user to define a set of things • Computes links between these things by analyzing text • The user then explores the graph directly
- 13. BETTER PROCESS • Keyword/key phrase extraction • Concept extraction • Entity extraction: people | events | orgs | etc. • Sentiment analysis • Dynamic ontologies • Spatio-temporal analysis • Rich visualizations: graph | map | trends | etc.
- 14. SOME NUMBERS Private/Public sources 694’040 sources 250’000 sources
- 15. SPECIALIZATION Corporate & Government Knowledge Mgmt and Analysis Public & Private Data Threats Prediction Public & Commercial News Market Analysis
- 16. WHAT’S IN COMMON? • Work at the Big Data Scale • Data Scientists • Customer-focused special teams (“forward engineers” – Palantir) • Enterprise customers • Graphs • Data Visualization • Live Data
- 18. WHAT IS GRAPH?
- 19. External Network DMZ Internal Network Dispatch Server Rev DB JDBC 3.0 w/ SSL Oracle Database Storage Raptor Server Lucene Index Storage HTTPS Shared Storage HTTPS Job Server Job Data and Specs Job Logs and Results HTTPS Client PALANTIR GOTHAM
- 20. INTEGRATES WITH EXISTING IT INFRASTRUCTURE • Your existing IT infrastructure • Authentication • Information Extractors • Legacy data stores • Rapidly changing data sources
- 21. INFORMATION EXTRACTORS • Large repositories of unstructured text • Multiple information extractors have been run across the text • Provide different types of extraction • Entities • Relationships • Metadata • Geotagging • Siloed view of each entity extractors output • Want to combine these views alongside structured data into one interface
- 22. • Objects • Latin taxonomy of animals • Objects and Properties • Periodic Table (has implicit relationships) • Objects and Relationships • Properties can be modeled as relationships to ‘data’ objects • Objects and Properties and Relationships • How information can be modeled in Palantir DYNAMIC ONTOLOGY
- 23. WHY SOFT-CODE THE ONTOLOGY? • A hard-coded Ontology is inherently limiting • Forces an organization into one of two extremes General Ontology Specific Ontology No Semantics Over-Defined Semantics
- 24. PALANTIR GOTHAM UI: SEARCH • Data Scale • 100 million row Netflix dataset • 10 million document usenet corpus • 1.5 million entity extracted Wikipedia corpus • Indexing Performance • 1m rows/hour structured indexing • 500k docs/hour unstructured document indexing • 100k docs/hour entity-extracted document indexing • Searching Performance • Sub-second search processing
- 28. CONSUMERS WILL WORK WITH AUGMENTED INTELLIGENT SYSTEMS • Consumer-focused PIAs are inherently limiting • Forces a user into one of two extremes Siri, Google Now Palantir Gotham Too- General Too-Enterprise Focused
- 29. AUGMENTED INTELLIGENT SYSTEMS WILL LEARN FROM THEIR USERS • They will learn user’s own dynamic ontology (as opposed to the corporate ontology) by using Semantic Steering • They will learn end user’s priorities (as opposed to the corporate priorities)
- 30. AUGMENTED INTELLIGENT SYSTEMS WILL WORK ON BEHALF OF USERS • Gather data on user’s demand (e.g., prepare reports) • Check teammates’ work progress
- 31. AUGMENTED INTELLIGENT SYSTEMS WILL PREDICT & ALERT • They will use knowledge about their user context (interests, goals, priorities, etc.) • They will combine it with data about the non- user context • To predict what’s next and alert user if necessary
- 32. AUGMENTED INTELLIGENT SYSTEMS WILL BE EMBEDDED INTO THE BRAIN CORTICAL MODEM ENABLES CYBER PROJECTIONS