Se ha denunciado esta presentación.
Utilizamos tu perfil de LinkedIn y tus datos de actividad para personalizar los anuncios y mostrarte publicidad más relevante. Puedes cambiar tus preferencias de publicidad en cualquier momento.

God’s Eye View: Will global AI empower us or destroy us? | Ramesh Raskar

134 visualizaciones

Publicado el

Video of the talk at https://www.youtube.com/watch?v=x9TCYuMUnco

Friction in data sharing is a large challenge for large scale machine learning. Emerging technologies in domains such as biomedicine, health and finance benefit from distributed deep learning methods which can allow multiple entities to train a deep neural network without requiring data sharing or resource aggregation at one single place. The talk will explore the main challenges in data friction that make capture, analysis and deployment of ML. The challenges include siloed and unstructured data, privacy and regulation of data sharing and incentive models for data transparent ecosystems. The talk will compare distributed deep learning methods of federated learning and split learning. Our team at MIT has pioneered a range of approaches including automated machine learning (AutoML), privacy preserving machine learning (PrivateML) and intrinsic as well as extrinsic data valuation (Data Markets). One of the programs at MIT aims to create a standard for data transparent ecosystems that can simultaneously address the privacy and utility of data.

Bio: Ramesh Raskar is an Associate Professor at MIT Media Lab and directs the Camera Culture research group. His focus is on AI and Imaging for health and sustainability. They span research in physical (e.g., sensors, health-tech), digital (e.g., automated and privacy-aware machine learning) and global (e.g., geomaps, autonomous mobility) domains. He received the Lemelson Award (2016), ACM SIGGRAPH Achievement Award (2017), DARPA Young Faculty Award (2009), Alfred P. Sloan Research Fellowship (2009), TR100 Award from MIT Technology Review (2004) and Global Industry Technovator Award (2003). He has worked on special research projects at Google [X], Apple Privacy Team and Facebook and co-founded/advised several companies. Project page https://splitlearning.github.io/" Ramesh Raskar is an Associate Professor at MIT Media Lab and directs the Camera Culture research group. His focus is on Machine Learning and Imaging for health and sustainability. They span research in physical (e.g., sensors, health-tech), digital (e.g., automated and privacy-aware machine learning) and global (e.g., geomaps, autonomous mobility) domains.

In his recent role at Facebook, he launched and led innovation teams in Digital Health, Health-tech, Satellite Imaging, TV and Bluetooth bandwidth for Connectivity, VR/AR and ‘Emerging Worlds’ initiative for FB.

At MIT, his co-inventions include camera to see around corners, femto-photography, automated machine learning (auto-ML), private ML, low-cost eye care devices (Netra,Catra, EyeSelfie), a novel CAT-Scan machine, motion capture (Prakash), long distance barcodes (Bokode), 3D interaction displays (BiDi screen), new theoretical models to augment light fields (ALF) to represent wave phenomena and algebraic rank constraints for 3D displays(HR3D).

Publicado en: Tecnología
  • Sé el primero en comentar

  • Sé el primero en recomendar esto

God’s Eye View: Will global AI empower us or destroy us? | Ramesh Raskar

  1. 1. Privacy Aware AI God’s Eye View Ramesh Raskar Associate Professor MIT Media Lab
  2. 2. Utility Privacy Private AI ? Health Data Google Maps Traffic
  3. 3. Privacy Consent Regulation Trade Secrets Quality Silos, Data Curation Incentive
  4. 4. Nothing to Hide
  5. 5. I We Ours ..
  6. 6. Strava HeatMap
  7. 7. More Regulation ?
  8. 8. Privacy Consent Regulation Trade Secrets Quality Silos, Data Curation Incentive Split Learning Auto ML Data Markets
  9. 9. Capture Precise Data Learn Act Train from Wisdom Predict for Task What Broker Needs
  10. 10. Human  Data  Statistics  Human(s) Precision x Privacy
  11. 11. Noisy GPS but ok for Traffic View
  12. 12. Differential Privacy, [Dwork 2011] US Census 2020
  13. 13. Learn AI EncryptObfuscate Add Noise e.g. Differential Privacy Homomorphic Encryption Infer Statistics Anonymize Strava Utility Privacy
  14. 14. Split Learning (MIT) ~Federated Learning Share Wisdom. Not Raw Data Learn AI EncryptSmashObfuscate Infer Statistics Anonymize Private AI
  15. 15. Master Algo for Pneumonia, Treatments Hospital1 2 3 .. Chest X-rays Train AI 100
  16. 16. Software Wisdom from Rules AI Wisdom from Examples Cooking Raising a child
  17. 17. Collective Intelligence Partially Trained Interns
  18. 18. Collective Intel Hospital1 2 3 .. Average AI 100 Federated Learning [McMahan17] Partially Trained Interns
  19. 19. Split Learning [Gupta17@ MIT]
  20. 20. Split Learning Master AI from Wisdom Smashed Wisdom [Gupta17@ MIT] .. Intern Notes for every patient Intern Notes for every patient .. 1 2 3 … 100 Chest X-rays
  21. 21. Server Smasher Smashed Wisdom 1 2 3 … 100 Split Learning [Gupta17, Vepakomma, Singh@ MIT]
  22. 22. Privacy Consent Regulation Trade Secrets Quality Silos, Data Curation Incentive Split Learning Auto ML Data Markets
  23. 23. Privacy Aware AI God’s Eye View Ramesh Raskar Associate Professor MIT Media Lab

×