Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Co-creating Innovation: Transformative Research for Results

166 views

Published on

For Fujitsu Laboratories of Europe, co-creation represents the cornerstone of our innovation activity, working with customers and Fujitsu businesses to solve real-world problems in today’s digital society. As Fujitsu’s global center of excellence for advanced creative information analytics, Fujitsu Laboratories is developing significant innovations within Fujitsu's human-centric AI Zinrai program. In this breakout session, we focus on examples that demonstrate our co-creation innovation activities in action, highlighting the essential role of collaboration with customers, partners, academia and business. Applications include healthcare, finance and manufacturing.
Speaker:
Tsuneo Nakata

Published in: Healthcare
  • Be the first to comment

  • Be the first to like this

Co-creating Innovation: Transformative Research for Results

  1. 1. 0 Copyright 2017 FUJITSU Fujitsu Forum 2017 #FujitsuForum
  2. 2. 1 Copyright 2017 FUJITSU Co-Creating Innovation – Transformative Research for Results Tsuneo Nakata CEO, Fujitsu Laboratories of Europe (FLE) Stories Untold
  3. 3. 2 Copyright 2017 FUJITSU Outcome of Laboratory-Led Digital Co-Creation Fujitsu Technology & Service Vision 2017 http://www.fujitsu.com/uk/microsite/vision/download- center/index.html Technology Innovation Business Research Org.Universit y Biz Units Partner/ Custome r Lab Lab Custome r Business Units La b
  4. 4. 3 Copyright 2017 FUJITSU Creating a new innovative business in Europe based on technologies from FLE and Fujitsu Laboratories Group together with local Fujitsu and local customers. Co-Creating Innovation 2. PoC building by co-creation 3. PoB building 4. Business deployment Objective Local Customer Local FujitsuFujitsu Labs Europe Process PoC: Proof of Concept PoB: Proof of Business 1. Show our capability (at events/shows)
  5. 5. 4 Copyright 2017 FUJITSU Fujitsu Laboratories of Europe, Ltd. (UK & Spain) (Established 2001) Fujitsu Laboratories Ltd. Kawasaki Laboratories (Japan) (Established 1968) Fujitsu Research and Development Center Co., Ltd. (China) (Established 1998) Fujitsu Laboratories of America, Inc. (USA) (Established 1993) Fujitsu Laboratories Group at a Glance  CEO: Dr. Shigeru Sasaki  R&D Budget: 250M EUR  Employees: 1,400 Continuously generating R&D results that amaze the world Quickly deploy R&D results to PoC/PoB on a global scale Generating innovations, including new business models that resonate through global markets FLE Co-Creation
  6. 6. 5 Copyright 2017 FUJITSU Case 1: AI-Based Non-Destructive Inspection Solution http://www.fujitsu.com/fts/about/resources/news/press- releases/2017/emeai-20171107-artificial-intelligence-solution-from.html Press release November 7, 2017 Co-created by: Defect
  7. 7. 6 Copyright 2017 FUJITSU AI Can Optimize Inspection Process  Hybrid solution with Deep Learning, Image/Signal Processing  Inspection time: 83% reduction – 34,000 man hour per year  To be put in production in two countries
  8. 8. 7 Copyright 2017 FUJITSU AI Can Optimize Inspection Process  Hybrid solution with Deep Learning, Image/Signal Processing  Inspection time: 83% reduction – 34,000 man hour per year  To be put in production in two countries Convert to image format Story Untold: Imagification Technology Detect salient features
  9. 9. 8 Copyright 2017 FUJITSU Imagification Framework  Imagification: Converting the input data into image by appropriate imagification function and feed it to a certain learning method Imagification framework Non-image Data Classification Regression Feature extraction Little domain knowledge Relatively small data Image Data Image Data Augmented Image Data Imagification Learning
  10. 10. 9 Copyright 2017 FUJITSU Validated Applications and Data Types Financial time series Wearable sensor Multi-variate TS IoT Sensor data measurements Smart energy consumption meter Raw data type Imagified data Application Historical trend search & prediction Real-time human activity detection; Driver safety Room occupancy detection Energy disaggregation by electrical appliance Raw data type Imagified data Application 3D CAD Models Ultrasound NDT measurements 3D Geometry and tabular data Self-assessment & benefit claims forms (scanned) Manufacturing Defect detection for quality control Pin quotation system Signature analysis for fraud detection
  11. 11. 10 Copyright 2017 FUJITSU Case 2: HIKARI – Intelligent Healthcare http://www.fujitsu.com/uk/about/resources/new s/press-releases/2015/pr-fle20161110.html Press release November 10, 2016 Co-created by: https://www.youtube.com/watch?v=NIDNmwYMjAE
  12. 12. 11 Copyright 2017 FUJITSU HIKARI Can Support Clinician’s Diagnosis 36,000 patient data Open Data Pubmed, ICD9/10, SNOMED Private Data Diagnosis by 5 senior clinicians 20min. per patient >85% Risk Assessment Accuracy Diagnosis by AI 0.2sec. per patient 33,215 61,491 5,944 1,242,192 339 Diagnoses Drugs Symptoms Sc. papers Concepts
  13. 13. 12 Copyright 2017 FUJITSU HIKARI Can Support Clinician’s Diagnosis 36,000 patient data Open Data Pubmed, ICD9/10, SNOMED Private Data Diagnosis by 5 senior clinicians 20min. per patient >85% Risk Assessment Accuracy Diagnosis by AI 0.2sec. per patient 33,215 61,491 5,944 1,242,192 339 Diagnoses Drugs Symptoms Sc. papers Concepts Data AnonymizationEngine Story Untold: Privacy Protection
  14. 14. 13 Copyright 2017 FUJITSU What is Anonymization?  [Wikipedia] Data anonymization is a type of information sanitization whose intent is privacy protection. It is the process of either encrypting or removing personally identifiable information from the data sets… Name Gender Age Disease Thomas M 103 Diabetes Ellen F 83 Cancer John M 23 Pneumonia Nick M 23 Pneumonia Gender Age Disease - Over 80 Diabetes - Over 80 Cancer M 20-29 Pneumonia M 20-29 Pneumonia Pseudo ID Gender Age Disease 1307 M 103 Diabetes 3931 F 83 Cancer 6494 M 23 Pneumonia 8529 M 23 Pneumonia
  15. 15. 14 Copyright 2017 FUJITSU Hospital Privacy Protection for Co-Creation Patient Data Patient Data GDPR Violation!
  16. 16. 15 Copyright 2017 FUJITSU Hospital Privacy Protection Protocol for Co-Creation Anonymized Patient Data Anonymization Patient Data Anonymized Patient Data Privacy Risk Scan Privacy Protection Tools Privacy Risk Scan GDPR-Ready!
  17. 17. 16 Copyright 2017 FUJITSU Privacy Risk Scan Technology  Estimate compensation for damage when data leakage  Judge conformance to multiple anonymization guidelines http://pr.fujitsu.com/jp/news/2016/07/19.html Summary of risk - Proposed measures - Guideline violations - Amount of damage eg. No measures may cause 660M JPY loss. Detail of guideline violation Personal identification risks
  18. 18. 17 Copyright 2017 FUJITSU Privacy Risk Scan Results  Prior check results if the data from a partner contains privacy info Raw data Pseudo ID Data Minimization Rounding Birthday Safer Actual Data Applying stronger anonymization scheme will reduce the risk almost to zero but may affect analysis results. Note: The data source is NOT the San Carlos Hospital.
  19. 19. 18 Copyright 2017 FUJITSU Conclusions ’Co-Create Innovation’ by Laboratory, business team and partners Two successful cases in manufacturing and healthcare Untold aspects in innovative co-creations: Unique and effective technology – Imagification Privacy protection protocol with the stakeholder(s) and more…
  20. 20. 19 Copyright 2017 FUJITSU

×