Future Technology Trends Explained

Transcription

1 Microsoft Future Decoded Future Technology Trends Explained Michael Chief Technology Officer Microsoft UK Thursday 1 st November 2018

2 Disclaimer! We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten. - Bill Gates, The Road Ahead, 1995

3 We are in the Era of the 4 th Industrial Revolution Historical Definition FIRST SECOND THIRD FOURTH 90 YEARS 70 YEARS 55 YEARS NOW BEYOND THE FOURTH Water and steam power is used to create mechanical production facilities Electricity lets us create a division of labor and mass production IT systems automate production line further IoT & Cloud Technology automate complex tasks Cloud Systems of Intelligence allow humans to work alongside robots and IoT systems ~2014 ~2018+ First mechanical loom First assembly line First programmable logic controller Advanced Analytics, big data and IoT enabled sensors Democratized AI and Industrial IoT, connected factories and Intelligent Edge

4

5

6

7

8 Foundations for AI Data Cloud Models Speech Vision Language Your data + Microsoft data Power of Azure Breakthrough advancements

9 Big Data Smart City 250 PB per day Smart Devices 20B IoT Devices People 1.5 GB per day Smart Home 50 GB per day 1 million/hour new devices coming online by 2020 Connected Factory 1 PB per day Autonomous Vehicle 5 TB per day Stadium 200 TB per game Smart Office 150 GB per day

10 Bytes Percentage of Digital Universe Storage Capacity is growing too slowly 5.0E E % 80% 3.0E+22 60% 2.0E+22 40% 1.0E E+00 Source: IDC Digital Universe Installed Capacity 20% 0% Data that has to be discarded Data that can be stored

11 10TB Credit: Tara Brown Photography/University of Washington

12 DNA data storage basics Simple mapping bases: A C G T Bits Base G data A C T A G C 00 A 01 C 10 G 11 T Store data in synthetic DNA strands

13 DNA data storage in a datacenter

14

15

16 Real-time AI at cloud scale with phenomenal performance and affordable cost Ultra-fast DNN performance with accelerated ResNet50 Models are easy to create and deploy into Azure cloud Write once, deploy anywhere to intelligent cloud or edge Manage and update your models using Azure ML & IoT Edge

17

18

19 Helping computers communicate with us better

20 Azure Cognitive Services Translation enhancements 拜尔斯帮助美国队在奥运会女子体操个人全能上实现了四连冠, 她个人在本届奥运会上共夺得五枚奖牌, 进一步印证了她是这一代, 也可能是有史以来, 最优秀的体操队员 Byers helped the U.S. team achieve a four-run individual allaround in Olympic women's gymnastics, and she personally won five medals at the Games, further confirming that she is the best gymnast of her generation and possibly of all time. Bayer helped team USA win four consecutive individual Olympic gold medals in women's gymnastics. She herself won five medals in this year s Olympics, which further confirms that she is the best gymnast of her generation, and possibly of all time.

21 Azure Cognitive Services Translation enhancements 拜尔斯帮助美国队在奥运会女子体操个人全能上实现了四连冠, 她个人在本届奥运会上共夺得五枚奖牌, 进一步印证了她是这一代, 也可能是有史以来, 最优秀的体操队员 Byers helped the U.S. team achieve a four-run individual allaround in Olympic women's gymnastics, and she personally won five medals at the Games, further confirming that she is the best gymnast of her generation and possibly of all time. Bayer helped team USA win four consecutive individual Olympic gold medals in women's gymnastics. She herself won five medals in this year s Olympics, which further confirms that she is the best gymnast of her generation, and possibly of all time. Coming soon!

22 Azure Cognitive Services Industry leading neural text to speech Byers helped the U.S. team achieve a four-run individual allaround in Olympic women's gymnastics, and she personally won five medals at the Games, further confirming that she is the best gymnast of her generation and possibly of all time. Sample 1 Sample 2 Sign up for access today aka.ms/tryneuraltts

23 Design AI to Earn Trust 23

24 Fairness Understand affect recommendations Attract diverse detect eliminate bias Human review Fairness 24

25 Reliability & Safety Evaluate training data Test extensively (and enable a user feedback loop) Monitor ongoing performance Design for unexpected circumstances including nefarious attacks Human in the loop 25

26 Privacy & Security Existing privacy laws (e.g. the General Data Protection Regulation apply Provide transparency about data collection and use, and good controls so people can make choices about their data Design systems to protect against bad actors Use de-identification techniques to promote both privacy and security 26

27 Inclusiveness Inclusive design practices to address potential barriers that could unintentionally exclude people Enhances opportunities for those with disabilities Build trust through contextual interaction EQ in addition to IQ 27

28 Transparency People should understand how decisions were made Provide contextual explanations Make it easier to raise awareness of potential bias, errors and unintended outcomes 28

29 Accountability People must be accountable for how their systems operate Norms should be observed during system design and in an ongoing manner Role for internal review boards 29

30

31 DIGITAL TWINS: FUSING PHYSICAL AND DIGITAL PHYSICAL WORLD DIGITAL WORLD Data Measure VI Control Insight, optimization Quantified Organization Data Measure Control P EO P L E Quantified Self Insight, optimization Data SPACES Intent, Context Control Assistance, Task completion Artificial Intelligence

32 Wide FOV Dual lasers for close and far range Megapixel Sensor

33

34

35

36

37

38 Superposition

39 Electron in left box bit 0 Electron in right box bit 1

40 Superposition Electron in left and right box simultaneously qubit 0 and 1

41 Schrödinger's cat

42 Calculation Time Classical Computing Age of the Universe Number of Atoms

43 RSA-2048 Challenge Problem Classical Quantum billion years seconds

44 Post-Quantum Cryptography - Cryptography approaches that appear to be resistant to an attacker who has access to a quantum computer NIST Post-Quantum Project - Standardisation approach for quantum-resistant public-key cryptographic algorithms Microsoft is working on four proposals: FrodoKEM is based upon the Learning with Errors problem, which is, in turn, based upon lattices. SIKE (Supersingular Isogeny Key Encapsulation) uses arithmetic operations of elliptic curves over finite fields to build a key exchange. Picnic is a public-key digital signature algorithm, based on a zero-knowledge proof system and symmetric key primitives. qtesla is a post-quantum signature scheme based upon the Ring Learning With Errors (R-LWE) problem. More Information:

45 Things to do next Session Feedback Please rate this session in the Future Decoded app! Microsoft UK AI Research Report Download the AI Report at Enhance your Digital Skills Find great guidance and resources at

46 Thank You