Future of Artificial Intelligence - IFTF @ Fujitsu Labs 2015
55 likes7,386 views
The document discusses trends in artificial intelligence investment and patent development. Private investment in AI has grown rapidly over the past five years, with key areas being machine learning, natural language processing, and medical technologies. While startups have received most funding, large companies like IBM, Microsoft and Google hold the majority of AI patents. The number of AI patent filings has increased over sevenfold in the past decade. Going forward, the author argues that further innovation will come from combinations of advances across different AI fields, rather than single technologies, creating new opportunities.
Future of Artificial Intelligence - IFTF @ Fujitsu Labs 2015
- 1. 1 FUTURE OF ARTIFICIAL INTELLIGENCE Andrew Trabulsi June 2015
- 2. 2 What are the major investment trends? What’s the IP landscape look like? What’s coming next?
- 4. 4 Source: Capital IQ, Quid, iftf Funding—where are we? Private investment growing rapidly, M&A peaked in 2012 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 2010 2011 2012 2013 2014 2015* Past five years of funding into artificial intelligence ($mn) M&A Private Investment
- 5. 5 Source: Capital IQ, Quid, iftf How’s the market shaping up? Key investment areas since 2010 Cybersecurity! Medical 10% MapReduce 9% NLP 9% Ads 8% ERP 8% Robotics 7% Shopping 7% Computer Vision 6% Market Composition MapReduce! Natural Language! Processing! Medical! Cancer Diagnostics! Recruiting! Robotics! Computer Vision! Music! Personal Assistant! News Analysis! Ads! Shopping! Financial services! Enterprise Resource! Planning!
- 6. 6 Source: Capital IQ, Quid, iftf Where’s it going? Notable private investments and M&A deals since 2010 Quid, Inc. ColdLight Solutions, LLC Databricks, Inc. Bright Media Corp. Antuit Pte. Ltd. Equivio, Ltd. Sentient Technologies, LLC Deep Mind Technologies, Ltd. $49mn — — $105mn $54mn — — $120mn $59mn — — $200mn $104mn — $400mn Private M&A
- 7. 7 But what are they doing with all this money?
- 8. 8 Source: Thomson Reuters, Quid, iftf Patent development is taking off Over 7x growth in IP development from 2004 - 2013 0 50 100 150 200 250 300 350 400 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Previous ten years of patent filings Machine Learning Patents
- 9. Source: Thomson Reuters, Quid, iftf The IP ecosystem is diverse Machine Learning patents across use cases Text Classification! Text Classification 11% Customer Analytics 11% Querying Methods 8% Medical Diagnostics 7% Data Mining 7% Cybersecurity 6% MRI Analysis 6% Facial Recognition 4% IP Deconstruction Social Network ! Analysis! Natural Language! Processing! Video Analysis! Molecular Analysis! Querying Method! Querying Method! Personal Assistant! Customer Analytics! Cybersecurity! Data Protection! Semiconductors! Data Mining! Neural Network! 9
- 10. 10 Source: Thomson Reuters, Quid, iftf …But held by only a few players Intellectual property not predominantly held by start-ups Microsoft IBM Yahoo Google Nippon Telegraph & Telephone Corp Hewlett-Packard Behavioral Recognition Systems, Inc. Qualcomm Patents held % 140 9% 90 6% 59 4% 49 3% 34 2% 28 2% 24 1.5% 20 1%
- 11. 11 What’s next?
- 12. 12 Combinatorial Innovation Won’t be just single technology or use case that changes game
- 13. 13 Combinatorial Innovation Won’t be just single technology or use case that changes game
- 14. 14 Source: Thomson Reuters, Quid, iftf Revisiting our patent analysis Over 7x growth in IP development from 2004 - 2013 0 50 100 150 200 250 300 350 400 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Previous ten years of patent filings Machine Learning Patents
- 15. 15 Source: Thomson Reuters, Quid, iftf It’s really a much bigger story… IP development across use cases 2004 - 2013 0 100 200 300 400 500 600 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Previous ten years of patent filings Automated Judgement Patents Machine Learning Patents
- 16. 16 Source: Thomson Reuters, Quid, iftf Innovations build on top of each other IP development across use cases 2004 - 2013 0 200 400 600 800 1000 1200 1400 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Previous ten years of patent filings Computer Vision Patents Automated Judgement Patents Machine Learning Patents
- 17. 17 Source: Thomson Reuters, Quid, iftf To create a wide range of opportunities IP development across use cases 2004 - 2013 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Previous ten years of patent filings Neural Network Patents Computer Vision Patents Automated Judgement Patents Machine Learning Patents
- 18. 18 Innovations overlap with one another. It won’t just be advances in a single field that change the world or imp- act our daily lives. It will be combinations of advances, across fields, that do. Understanding this, helps put to- day’s innovations into perspective for tomorrow.
- 19. 19 Thank you!