for the internet of things
play

for the Internet of Things Shaowei Lin GTC San Jose May 2017 - PowerPoint PPT Presentation

Artificial General Intelligence for the Internet of Things Shaowei Lin GTC San Jose May 2017 Internet of Things Heterogeneous Systems Resource Constraints Higher-Order Intelligence Distributed Intelligence Deep Neural Networks


  1. Artificial General Intelligence for the Internet of Things Shaowei Lin GTC San Jose May 2017

  2. Internet of Things Heterogeneous Systems Resource Constraints Higher-Order Intelligence Distributed Intelligence Deep Neural Networks Reinforcement Learning Machine Reasoning

  3. Deep Learning

  4. WHAT IS DEEP LEARNING? Biologically-inspired multi-layer neural networks Deeper layers learn higher-order features

  5. DEEP LEARNING FOR SENSOR NETWORKS Sensor networks form the nervous system of smart cities. Deep visual cortex Deep learning Sensor networks

  6. STRUCTURED MISSING DATA Joint work with Liangze Wong, Daniel Chen, Huiling Chen (A*STAR) Exp 1 Original Data Exp 2 Estimates Masked Inputs Features

  7. MULTIMODAL SENSOR FUSION Joint work with Wenyu Zhang, Zuozhu Liu, Tony Quek Server Fused Representation Gateway Illuminance Humidity Temperature Sensor Sensor Sensor Node Node Node Measured Measured Measured Illuminance Humidity Temperature

  8. REINFORCEMENT LEARNING

  9. How can the network learn to accomplish given tasks and distribute required steps while managing resources efficiently?

  10. self-programming machine machines? reasoning? How can the network learn to accomplish given tasks and distribute required steps while managing resources efficiently? what objective function what is the should we use? action space?

  11. Machine Reasoning

  12. 12

  13. 1901 RUSSELL’S PARADOX TYPE THEORY SET THEORY 2009 HOMOTOPY TYPES CATEGORIES

  14. 14

  15. CURRY-HOWARD CORRESPONDENCE SPACE POINT TYPE TERM THEOREM PROOF INTENT IMPLEMENTATION

  16. CURRY-HOWARD CORRESPONDENCE SPACE POINT TYPE TERM THEOREM PROOF INTENT IMPLEMENTATION

  17. Traditional Programming Intentional Programming <<print elements of list>> (<<list with numbers 1 to 10>>)

  18. INTENT AS A TYPE Poor Type System sort: list nat → list nat Rich Type System sort: ∀ ℓ: list nat , ℓ ′ : list nat sorted ℓ ′ ⋀ same_elements ℓ ℓ ′ }

  19. FINDING IMPLEMENTATIONS FOR INTENTS Context Intent

  20. EQUIVALENT PROBLEMS THROUGH PATHS Context Intent

  21. INTUITION FOR PROBLEM-SOLVING Problems Equivalences (Paths)

  22. INTERNET OF THINGS Functional Plane (Intents) Publish- Named- Linked Function Subscribe Data Protocols Networking Physical Plane (Implementations)

  23. CRITICAL SERVICES Convert intent into implementation. Compilers. Check that implementation matches intent. Blockchain?

  24. THANK YOU HTTPS://SUTDBRAIN.WORDPRESS.COM/

Recommend


More recommend