from retrospective to continuous deep analytics
play

FROM RETROSPECTIVE TO CONTINUOUS DEEP ANALYTICS Seif Haridi KTH - PowerPoint PPT Presentation

FROM RETROSPECTIVE TO CONTINUOUS DEEP ANALYTICS Seif Haridi KTH SICS Why most Data Analysis today is Retrospective From OLAP Databases Data Cube LOCATION TIME To Deep Analysis on Batches Past Data Sets Map Map Map Map Map Map


  1. FROM RETROSPECTIVE TO CONTINUOUS DEEP ANALYTICS Seif Haridi KTH ‐ SICS

  2. Why most Data Analysis today is Retrospective

  3. From OLAP Databases Data Cube LOCATION TIME

  4. To Deep Analysis on Batches Past Data Sets Map Map Map Map Map Map Reduce Reduce Reduce Reduce Reduce Reduce Past Models

  5. And Machine Learning Pipelines Feature Extraction Yesterday’s Dataset Model Training

  6. Today’s Retrospective Data Processing Knowledge Data STORE LOAD PROCESS STORE

  7. Today’s Retrospective Data Processing Knowledge Data time Takes Long to Extract Knowledge

  8. Models and patterns on older data personalized ad winter jacket discount! are often irrelevant today

  9. Critical Decisions demand Continuous Analysis

  10. We propose a continuous processing architecture Knowledge Data STORE LOAD PROCESS STORE time

  11. We propose a continuous processing architecture Knowledge Data PROCESS time

  12. We propose a deep processing architecture PROCESS arbitrarily iterative computation

  13. We propose a fast processing architecture PROCESS TPUs

  14. for Dynamic Graph analysis virus outbreaks social network trends

  15. for Online Machine Learning feature learning tensor programming

  16. and Relational Data Streaming σ θ π σ θ π σ θ σ θ dynamic tables

  17. Unified and Optimised on a Common Representation Dynamic Online Relational Graphs ML Streams Intermediate Representation

  18. A Runtime Designed for Continuous Deep Analysis Intermediate Representation Distributed Runtime metrics config constraints Self‐Reconfiguration

  19. A Full Stack for Continuous Deep Analytics Dynamic Online Relational Graphs ML Streams Intermediate Representation Distributed Runtime metrics config constraints Interpretable, Online Models of the world

Recommend


More recommend