SlideShare une entreprise Scribd logo
1  sur  23
Télécharger pour lire hors ligne
Artificial Learning
Ultra-efficient integrated circuits for machine learning
Hardware for deep learning
and mobile autonomy
Business model & market strategy
www.artificiallearning.com
@ArtificialLearn
Copyright © 2014 Artificial Learning Ltd All rights reserved2014-08-06 1
OPPORTUNITY
Artificial Learning
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 2
Sensed environment Affected environment
Learn individuals’
features
Recognise and
classify
individuals • Resident?
• Stranger?
Act on what is
recognised
Autonomous deep learning
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 3
Learn Recognise Act
Four problems we address
1. Deep learning today is (very) inefficient
• Cannot tie autonomous system to the cloud
2. Deep learning today is unscalable
• Need 2 x Google global just to learn YouTube
• Torrent from IoT sensors will swamp the cloud
3. Biology beats technology by »108 - how?
• Deep learning algorithms unsuited to standard
processors
• Moore and Dennard scaling have stopped
4. Tough route to market for disruptive tech
• Silicon economics requires mass markets
• No established autonomous deep-learning market
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 4
versus
Our solutions
Ultra-efficient integrated circuits for machine learning
• Fast, cool, low mass, low volume hybrid ASICs
• Novel highly-scalable architecture
• Inevitable: “Everything good becomes hardware.”
–Nat Torkington, O’Reilly Radar, Machine Learning on a Board review, May 2014
A roadmap to mass markets for our IP
• Agreed R&D plan
• Market development before commitment to new silicon
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 5
MLoaB
MLoaC
MLbD
Unique value propositions
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 6
• For ultra-high efficiency deep machine learning our
ground-breaking chip designs can be 10,000x more
efficient than conventional CPUs
• Unlike general purpose computers, our dedicated chip
designs can help you deploy powerful machine learning
in autonomous mobile apps
• Machine Learning on a Board lets you easily create
products able to learn, recognise and act
Unfair advantage
• Our unique technology gives us orders of magnitude advantage
in scalable efficiency
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 7
Business model
Lean start-up model
evolved rapidly through
hypothesis testing
leanlaunchlab.com
Our market strategy
made front page of
most-voted list in 2014
Stanford Technology
Entrepreneurship class
bit.ly/TPE2MostVotedList
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 8
Lessons learned
We need early adopters to
create novel mass-market
products
We gain advantage if we put
plug-compatible interfaces in
front of disruptive IP
We can grow through several
market stages and product
generations
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 9
MARKET STRATEGY
Artificial Learning
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 10
Market segments
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 11
Market
segment
Products
Outcomes
Markets
Innovators & early
adopters
Plug-compatible circuit
board and application
software
Novel applications of
autonomous machine
learning
Proof of market before
heavy investment in silicon
10’s of channel partners x
10,000s of machine-learning
enhanced products / year
AND 10,000s of enthusiasts
x 1-2 boards / year
Early majority
Packaged chips
Our chips on customers’
own circuit boards for
specific products
100s of companies buying
millions of chips / year
Mass market
Licenced IP
Our IP in customers’
systems-on-chips
100s of companies using IP
in 10s of millions of devices
/ year
Machine Learning on a Board
Machine Learning on a Chip
Machine Learning by Design
Unitvolume(logscale)
Unit price (log scale)
Personal
mobile
Makers &
Hobbyists
Industrial
IoT
Product classes which can be enhanced
by autonomous machine learning
Industrial
R&D
Academic
research
Channels for embedded ML
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 12
Target
market
Early
majority /
low volume
products
Mass marketMass market
& specialised
niches
Early
adopters
Machine
Learning
on a Chip
Machine
Learning by
Design IP
Machine
Learning
on a Board
Target
product
Defence
Space
Domestic
IoT
Commercial
autonomous
systems
Private
autonomous
systems
Initial estimates - being refined now
MLoaB generations
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 13
Ling
•Simple digital
emulation
•COTS + Open
Source
Yi
•GPU-
enhanced
emulation
•RaspberryPi
minimal
configuration
Er
•FPGA-
enhanced
emulation
•Kickstarter?
San
•ASIC package
Plug-compatible upgrades with million-fold performance increase
Market development
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 14
Present
aims
• Create demand
• Establish channel partnerships
• Increase market knowledge and visibility
• Shape minimum viable products
Present
activity
• Engagement with market influencers
• Meet with potential channel partners
• Interviews with early adopters
• Online surveys / calls for proposals
Market makers
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 15
Hardware platforms
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 16
MLaaS MLoaCMLoaB MLbD
CORPORATE INFORMATION
Artificial Learning
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 17
1980
• Co-
founders
met
2009-2010
• Frameworks
for ultra-
efficient
machine
learning
• Project
formalised
2011-2012
• Pigeon
Consortium
formed
• CMOS
feasibility
study with
Imperial
College
• Artificial
Learning Ltd
founded
2013-
2014
• Proof of
concept
plans laid
• Market
strategy
and
roadmap
Company history
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 18
Pigeon Consortium research group
• Artificial Learning Ltd
• University of Edinburgh
• Scottish Microelectronics Centre
• University of Stirling
Core team
• Artificial Learning Ltd artificiallearning.com @ArtificialLearn
Peter Newman PhD, Michael Bate MPhil
• Key concepts and IP for machine learning hardware
• Mathematical design and simulation
• Commercialisation
• Pigeon Consortium
University of Edinburgh, University of Stirling
• Our research collaboration
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 19
Partnership and allies
• Prototype R&D
– Hardware design: Pigeon Consortium
– Research funding: EPSRC
– Chip fabrication: Europractice members
• Channels – Developing these now
– End-user product manufacturers
– Maker community
• Other alliances
– Software engineers
– London Tech City network
– UK Electronic Systems Community
– Deep learning and machine learning
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 20
Risk reduction
People
Technical
Market
Financial
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 21
Research group
formed and
product
development
team skills
identified
Feasibility study
completed and
R&D plan
agreed
Current focus of
risk reduction
efforts: market
gaps, value,
channels
R&D costs
understood but
production and
sales costs to
be determined
Investment Readiness
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 22
4
9. Validate metrics that matter
8. Validate left side of canvas
7. Prototype high-fidelity MVP
6. Validate right side of canvas
5. Validate product/market fit
4. Prototype low-fidelity MVP
3. Problem-solutions validation
2. Market size/competitive analysis
1. Complete first-pass canvas
Tasks in hand to
move to IRL 5:
• Identify real-world
applications and
quantify markets
• Lean prototyping
minimum viable
product designs
tested with
potential early
adopters
• Develop channel
partnerships
Technology Readiness - MLoaC
2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 23
4
9. Flight-proven in operations
8. Flight-qualified in test/demo
7. Prototype in live environment
6. Prototype in relevant environment
5. Component validation relevant env.
4. Component validation lab environment
3. Analytical/experimental proof of concept
2. Technology concept/applications formed
1. Basic principles observed and reported
Tasks in hand to
move to TRL 5:
• Obtain funding for
Pigeon Consortium
development
programme
• Device functional
simulations under
way
• Identifying IP
alliances for next
steps
• Team building

Contenu connexe

Similaire à Hardware for deep learning and mobile autonomy

Ch7 Strategy and Technology_updated.ppt
Ch7 Strategy and Technology_updated.pptCh7 Strategy and Technology_updated.ppt
Ch7 Strategy and Technology_updated.ppt
ZeeshanZahoorSyed
 
Mike Siegler at INCOSE Minneapolis, 2014
Mike Siegler at INCOSE Minneapolis, 2014Mike Siegler at INCOSE Minneapolis, 2014
Mike Siegler at INCOSE Minneapolis, 2014
Etherios
 
Pivotal data science_data_engineering_secret_weapons_of_the_strategic_enterprise
Pivotal data science_data_engineering_secret_weapons_of_the_strategic_enterprisePivotal data science_data_engineering_secret_weapons_of_the_strategic_enterprise
Pivotal data science_data_engineering_secret_weapons_of_the_strategic_enterprise
EMC
 

Similaire à Hardware for deep learning and mobile autonomy (20)

TRIK robotics
TRIK robotics TRIK robotics
TRIK robotics
 
IP and WTP for digital products
IP and WTP for digital productsIP and WTP for digital products
IP and WTP for digital products
 
Big Data Technologies for Social Media Analytics- Impetus Webinar
Big Data Technologies for Social Media Analytics- Impetus WebinarBig Data Technologies for Social Media Analytics- Impetus Webinar
Big Data Technologies for Social Media Analytics- Impetus Webinar
 
#FiaComit - The Mobicloud Showcase
#FiaComit - The Mobicloud Showcase#FiaComit - The Mobicloud Showcase
#FiaComit - The Mobicloud Showcase
 
Crowdsourcing vs. Technology Scouting in a B2B setting
Crowdsourcing vs. Technology Scouting in a B2B settingCrowdsourcing vs. Technology Scouting in a B2B setting
Crowdsourcing vs. Technology Scouting in a B2B setting
 
What is IoT and how Modulus and Pacific can Help - Featuring Node.js and Roll...
What is IoT and how Modulus and Pacific can Help - Featuring Node.js and Roll...What is IoT and how Modulus and Pacific can Help - Featuring Node.js and Roll...
What is IoT and how Modulus and Pacific can Help - Featuring Node.js and Roll...
 
Breaking the barriers of Internet of Things (IoT)
Breaking the barriers of Internet of Things (IoT)Breaking the barriers of Internet of Things (IoT)
Breaking the barriers of Internet of Things (IoT)
 
Ch7 Strategy and Technology_updated.ppt
Ch7 Strategy and Technology_updated.pptCh7 Strategy and Technology_updated.ppt
Ch7 Strategy and Technology_updated.ppt
 
Building a Digital Products Portfolio for Real Business Results
Building a Digital Products Portfolio for Real Business ResultsBuilding a Digital Products Portfolio for Real Business Results
Building a Digital Products Portfolio for Real Business Results
 
OMA Overview of the Organization & Work
OMA Overview of the Organization & WorkOMA Overview of the Organization & Work
OMA Overview of the Organization & Work
 
Getting to timely insights - how to make it happen?
Getting to timely insights - how to make it happen?Getting to timely insights - how to make it happen?
Getting to timely insights - how to make it happen?
 
IOvents project overview
IOvents project overviewIOvents project overview
IOvents project overview
 
Making advertising personal, 4th NL Recommenders Meetup
Making advertising personal, 4th NL Recommenders MeetupMaking advertising personal, 4th NL Recommenders Meetup
Making advertising personal, 4th NL Recommenders Meetup
 
Mike Siegler at INCOSE Minneapolis, 2014
Mike Siegler at INCOSE Minneapolis, 2014Mike Siegler at INCOSE Minneapolis, 2014
Mike Siegler at INCOSE Minneapolis, 2014
 
The Programmable Telecom Network, Doug Tait, Oracle, Enzo Amorino, Telecom It...
The Programmable Telecom Network, Doug Tait, Oracle, Enzo Amorino, Telecom It...The Programmable Telecom Network, Doug Tait, Oracle, Enzo Amorino, Telecom It...
The Programmable Telecom Network, Doug Tait, Oracle, Enzo Amorino, Telecom It...
 
Pivotal data science_data_engineering_secret_weapons_of_the_strategic_enterprise
Pivotal data science_data_engineering_secret_weapons_of_the_strategic_enterprisePivotal data science_data_engineering_secret_weapons_of_the_strategic_enterprise
Pivotal data science_data_engineering_secret_weapons_of_the_strategic_enterprise
 
Invention to Innovation - Journey of a High Tech Product
Invention to Innovation - Journey of a High Tech Product Invention to Innovation - Journey of a High Tech Product
Invention to Innovation - Journey of a High Tech Product
 
Introducing Gradient
Introducing Gradient Introducing Gradient
Introducing Gradient
 
Cutting Through the Disruption
Cutting Through the DisruptionCutting Through the Disruption
Cutting Through the Disruption
 
141113 tektronix _meet the experts_final
141113 tektronix _meet the experts_final141113 tektronix _meet the experts_final
141113 tektronix _meet the experts_final
 

Dernier

一比一定(购)坎特伯雷大学毕业证(UC毕业证)成绩单学位证
一比一定(购)坎特伯雷大学毕业证(UC毕业证)成绩单学位证一比一定(购)坎特伯雷大学毕业证(UC毕业证)成绩单学位证
一比一定(购)坎特伯雷大学毕业证(UC毕业证)成绩单学位证
wpkuukw
 
Call Now ≽ 9953056974 ≼🔝 Call Girls In Yusuf Sarai ≼🔝 Delhi door step delevry≼🔝
Call Now ≽ 9953056974 ≼🔝 Call Girls In Yusuf Sarai ≼🔝 Delhi door step delevry≼🔝Call Now ≽ 9953056974 ≼🔝 Call Girls In Yusuf Sarai ≼🔝 Delhi door step delevry≼🔝
Call Now ≽ 9953056974 ≼🔝 Call Girls In Yusuf Sarai ≼🔝 Delhi door step delevry≼🔝
9953056974 Low Rate Call Girls In Saket, Delhi NCR
 
Makarba ( Call Girls ) Ahmedabad ✔ 6297143586 ✔ Hot Model With Sexy Bhabi Rea...
Makarba ( Call Girls ) Ahmedabad ✔ 6297143586 ✔ Hot Model With Sexy Bhabi Rea...Makarba ( Call Girls ) Ahmedabad ✔ 6297143586 ✔ Hot Model With Sexy Bhabi Rea...
Makarba ( Call Girls ) Ahmedabad ✔ 6297143586 ✔ Hot Model With Sexy Bhabi Rea...
Naicy mandal
 
➥🔝 7737669865 🔝▻ kakinada Call-girls in Women Seeking Men 🔝kakinada🔝 Escor...
➥🔝 7737669865 🔝▻ kakinada Call-girls in Women Seeking Men  🔝kakinada🔝   Escor...➥🔝 7737669865 🔝▻ kakinada Call-girls in Women Seeking Men  🔝kakinada🔝   Escor...
➥🔝 7737669865 🔝▻ kakinada Call-girls in Women Seeking Men 🔝kakinada🔝 Escor...
amitlee9823
 
Vip Mumbai Call Girls Kalyan Call On 9920725232 With Body to body massage wit...
Vip Mumbai Call Girls Kalyan Call On 9920725232 With Body to body massage wit...Vip Mumbai Call Girls Kalyan Call On 9920725232 With Body to body massage wit...
Vip Mumbai Call Girls Kalyan Call On 9920725232 With Body to body massage wit...
amitlee9823
 
一比一定(购)新西兰林肯大学毕业证(Lincoln毕业证)成绩单学位证
一比一定(购)新西兰林肯大学毕业证(Lincoln毕业证)成绩单学位证一比一定(购)新西兰林肯大学毕业证(Lincoln毕业证)成绩单学位证
一比一定(购)新西兰林肯大学毕业证(Lincoln毕业证)成绩单学位证
wpkuukw
 
Call Girls In RT Nagar ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In RT Nagar ☎ 7737669865 🥵 Book Your One night StandCall Girls In RT Nagar ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In RT Nagar ☎ 7737669865 🥵 Book Your One night Stand
amitlee9823
 
Escorts Service Sanjay Nagar ☎ 7737669865☎ Book Your One night Stand (Bangalore)
Escorts Service Sanjay Nagar ☎ 7737669865☎ Book Your One night Stand (Bangalore)Escorts Service Sanjay Nagar ☎ 7737669865☎ Book Your One night Stand (Bangalore)
Escorts Service Sanjay Nagar ☎ 7737669865☎ Book Your One night Stand (Bangalore)
amitlee9823
 
(👉Ridhima)👉VIP Model Call Girls Mulund ( Mumbai) Call ON 9967824496 Starting ...
(👉Ridhima)👉VIP Model Call Girls Mulund ( Mumbai) Call ON 9967824496 Starting ...(👉Ridhima)👉VIP Model Call Girls Mulund ( Mumbai) Call ON 9967824496 Starting ...
(👉Ridhima)👉VIP Model Call Girls Mulund ( Mumbai) Call ON 9967824496 Starting ...
motiram463
 
CHEAP Call Girls in Hauz Quazi (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Hauz Quazi  (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICECHEAP Call Girls in Hauz Quazi  (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Hauz Quazi (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
9953056974 Low Rate Call Girls In Saket, Delhi NCR
 
CHEAP Call Girls in Ashok Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Ashok Nagar  (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICECHEAP Call Girls in Ashok Nagar  (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Ashok Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
9953056974 Low Rate Call Girls In Saket, Delhi NCR
 
CHEAP Call Girls in Vinay Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Vinay Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICECHEAP Call Girls in Vinay Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Vinay Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
9953056974 Low Rate Call Girls In Saket, Delhi NCR
 

Dernier (20)

一比一定(购)坎特伯雷大学毕业证(UC毕业证)成绩单学位证
一比一定(购)坎特伯雷大学毕业证(UC毕业证)成绩单学位证一比一定(购)坎特伯雷大学毕业证(UC毕业证)成绩单学位证
一比一定(购)坎特伯雷大学毕业证(UC毕业证)成绩单学位证
 
(ISHITA) Call Girls Service Aurangabad Call Now 8617697112 Aurangabad Escorts...
(ISHITA) Call Girls Service Aurangabad Call Now 8617697112 Aurangabad Escorts...(ISHITA) Call Girls Service Aurangabad Call Now 8617697112 Aurangabad Escorts...
(ISHITA) Call Girls Service Aurangabad Call Now 8617697112 Aurangabad Escorts...
 
Call Now ≽ 9953056974 ≼🔝 Call Girls In Yusuf Sarai ≼🔝 Delhi door step delevry≼🔝
Call Now ≽ 9953056974 ≼🔝 Call Girls In Yusuf Sarai ≼🔝 Delhi door step delevry≼🔝Call Now ≽ 9953056974 ≼🔝 Call Girls In Yusuf Sarai ≼🔝 Delhi door step delevry≼🔝
Call Now ≽ 9953056974 ≼🔝 Call Girls In Yusuf Sarai ≼🔝 Delhi door step delevry≼🔝
 
Get Premium Pimple Saudagar Call Girls (8005736733) 24x7 Rate 15999 with A/c ...
Get Premium Pimple Saudagar Call Girls (8005736733) 24x7 Rate 15999 with A/c ...Get Premium Pimple Saudagar Call Girls (8005736733) 24x7 Rate 15999 with A/c ...
Get Premium Pimple Saudagar Call Girls (8005736733) 24x7 Rate 15999 with A/c ...
 
Top Rated Pune Call Girls Chakan ⟟ 6297143586 ⟟ Call Me For Genuine Sex Serv...
Top Rated  Pune Call Girls Chakan ⟟ 6297143586 ⟟ Call Me For Genuine Sex Serv...Top Rated  Pune Call Girls Chakan ⟟ 6297143586 ⟟ Call Me For Genuine Sex Serv...
Top Rated Pune Call Girls Chakan ⟟ 6297143586 ⟟ Call Me For Genuine Sex Serv...
 
Escorts Service Daryaganj - 9899900591 College Girls & Models 24/7
Escorts Service Daryaganj - 9899900591 College Girls & Models 24/7Escorts Service Daryaganj - 9899900591 College Girls & Models 24/7
Escorts Service Daryaganj - 9899900591 College Girls & Models 24/7
 
Deira Dubai Escorts +0561951007 Escort Service in Dubai by Dubai Escort Girls
Deira Dubai Escorts +0561951007 Escort Service in Dubai by Dubai Escort GirlsDeira Dubai Escorts +0561951007 Escort Service in Dubai by Dubai Escort Girls
Deira Dubai Escorts +0561951007 Escort Service in Dubai by Dubai Escort Girls
 
Makarba ( Call Girls ) Ahmedabad ✔ 6297143586 ✔ Hot Model With Sexy Bhabi Rea...
Makarba ( Call Girls ) Ahmedabad ✔ 6297143586 ✔ Hot Model With Sexy Bhabi Rea...Makarba ( Call Girls ) Ahmedabad ✔ 6297143586 ✔ Hot Model With Sexy Bhabi Rea...
Makarba ( Call Girls ) Ahmedabad ✔ 6297143586 ✔ Hot Model With Sexy Bhabi Rea...
 
➥🔝 7737669865 🔝▻ kakinada Call-girls in Women Seeking Men 🔝kakinada🔝 Escor...
➥🔝 7737669865 🔝▻ kakinada Call-girls in Women Seeking Men  🔝kakinada🔝   Escor...➥🔝 7737669865 🔝▻ kakinada Call-girls in Women Seeking Men  🔝kakinada🔝   Escor...
➥🔝 7737669865 🔝▻ kakinada Call-girls in Women Seeking Men 🔝kakinada🔝 Escor...
 
Vip Mumbai Call Girls Kalyan Call On 9920725232 With Body to body massage wit...
Vip Mumbai Call Girls Kalyan Call On 9920725232 With Body to body massage wit...Vip Mumbai Call Girls Kalyan Call On 9920725232 With Body to body massage wit...
Vip Mumbai Call Girls Kalyan Call On 9920725232 With Body to body massage wit...
 
一比一定(购)新西兰林肯大学毕业证(Lincoln毕业证)成绩单学位证
一比一定(购)新西兰林肯大学毕业证(Lincoln毕业证)成绩单学位证一比一定(购)新西兰林肯大学毕业证(Lincoln毕业证)成绩单学位证
一比一定(购)新西兰林肯大学毕业证(Lincoln毕业证)成绩单学位证
 
Call Girls in Vashi Escorts Services - 7738631006
Call Girls in Vashi Escorts Services - 7738631006Call Girls in Vashi Escorts Services - 7738631006
Call Girls in Vashi Escorts Services - 7738631006
 
Call Girls In RT Nagar ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In RT Nagar ☎ 7737669865 🥵 Book Your One night StandCall Girls In RT Nagar ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In RT Nagar ☎ 7737669865 🥵 Book Your One night Stand
 
Escorts Service Sanjay Nagar ☎ 7737669865☎ Book Your One night Stand (Bangalore)
Escorts Service Sanjay Nagar ☎ 7737669865☎ Book Your One night Stand (Bangalore)Escorts Service Sanjay Nagar ☎ 7737669865☎ Book Your One night Stand (Bangalore)
Escorts Service Sanjay Nagar ☎ 7737669865☎ Book Your One night Stand (Bangalore)
 
(👉Ridhima)👉VIP Model Call Girls Mulund ( Mumbai) Call ON 9967824496 Starting ...
(👉Ridhima)👉VIP Model Call Girls Mulund ( Mumbai) Call ON 9967824496 Starting ...(👉Ridhima)👉VIP Model Call Girls Mulund ( Mumbai) Call ON 9967824496 Starting ...
(👉Ridhima)👉VIP Model Call Girls Mulund ( Mumbai) Call ON 9967824496 Starting ...
 
CHEAP Call Girls in Hauz Quazi (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Hauz Quazi  (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICECHEAP Call Girls in Hauz Quazi  (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Hauz Quazi (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
 
Call Girls Pimple Saudagar Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Pimple Saudagar Call Me 7737669865 Budget Friendly No Advance BookingCall Girls Pimple Saudagar Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Pimple Saudagar Call Me 7737669865 Budget Friendly No Advance Booking
 
SM-N975F esquematico completo - reparación.pdf
SM-N975F esquematico completo - reparación.pdfSM-N975F esquematico completo - reparación.pdf
SM-N975F esquematico completo - reparación.pdf
 
CHEAP Call Girls in Ashok Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Ashok Nagar  (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICECHEAP Call Girls in Ashok Nagar  (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Ashok Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
 
CHEAP Call Girls in Vinay Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Vinay Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICECHEAP Call Girls in Vinay Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Vinay Nagar (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
 

Hardware for deep learning and mobile autonomy

  • 1. Artificial Learning Ultra-efficient integrated circuits for machine learning Hardware for deep learning and mobile autonomy Business model & market strategy www.artificiallearning.com @ArtificialLearn Copyright © 2014 Artificial Learning Ltd All rights reserved2014-08-06 1
  • 2. OPPORTUNITY Artificial Learning 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 2
  • 3. Sensed environment Affected environment Learn individuals’ features Recognise and classify individuals • Resident? • Stranger? Act on what is recognised Autonomous deep learning 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 3 Learn Recognise Act
  • 4. Four problems we address 1. Deep learning today is (very) inefficient • Cannot tie autonomous system to the cloud 2. Deep learning today is unscalable • Need 2 x Google global just to learn YouTube • Torrent from IoT sensors will swamp the cloud 3. Biology beats technology by »108 - how? • Deep learning algorithms unsuited to standard processors • Moore and Dennard scaling have stopped 4. Tough route to market for disruptive tech • Silicon economics requires mass markets • No established autonomous deep-learning market 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 4 versus
  • 5. Our solutions Ultra-efficient integrated circuits for machine learning • Fast, cool, low mass, low volume hybrid ASICs • Novel highly-scalable architecture • Inevitable: “Everything good becomes hardware.” –Nat Torkington, O’Reilly Radar, Machine Learning on a Board review, May 2014 A roadmap to mass markets for our IP • Agreed R&D plan • Market development before commitment to new silicon 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 5 MLoaB MLoaC MLbD
  • 6. Unique value propositions 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 6 • For ultra-high efficiency deep machine learning our ground-breaking chip designs can be 10,000x more efficient than conventional CPUs • Unlike general purpose computers, our dedicated chip designs can help you deploy powerful machine learning in autonomous mobile apps • Machine Learning on a Board lets you easily create products able to learn, recognise and act
  • 7. Unfair advantage • Our unique technology gives us orders of magnitude advantage in scalable efficiency 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 7
  • 8. Business model Lean start-up model evolved rapidly through hypothesis testing leanlaunchlab.com Our market strategy made front page of most-voted list in 2014 Stanford Technology Entrepreneurship class bit.ly/TPE2MostVotedList 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 8
  • 9. Lessons learned We need early adopters to create novel mass-market products We gain advantage if we put plug-compatible interfaces in front of disruptive IP We can grow through several market stages and product generations 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 9
  • 10. MARKET STRATEGY Artificial Learning 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 10
  • 11. Market segments 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 11 Market segment Products Outcomes Markets Innovators & early adopters Plug-compatible circuit board and application software Novel applications of autonomous machine learning Proof of market before heavy investment in silicon 10’s of channel partners x 10,000s of machine-learning enhanced products / year AND 10,000s of enthusiasts x 1-2 boards / year Early majority Packaged chips Our chips on customers’ own circuit boards for specific products 100s of companies buying millions of chips / year Mass market Licenced IP Our IP in customers’ systems-on-chips 100s of companies using IP in 10s of millions of devices / year Machine Learning on a Board Machine Learning on a Chip Machine Learning by Design
  • 12. Unitvolume(logscale) Unit price (log scale) Personal mobile Makers & Hobbyists Industrial IoT Product classes which can be enhanced by autonomous machine learning Industrial R&D Academic research Channels for embedded ML 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 12 Target market Early majority / low volume products Mass marketMass market & specialised niches Early adopters Machine Learning on a Chip Machine Learning by Design IP Machine Learning on a Board Target product Defence Space Domestic IoT Commercial autonomous systems Private autonomous systems Initial estimates - being refined now
  • 13. MLoaB generations 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 13 Ling •Simple digital emulation •COTS + Open Source Yi •GPU- enhanced emulation •RaspberryPi minimal configuration Er •FPGA- enhanced emulation •Kickstarter? San •ASIC package Plug-compatible upgrades with million-fold performance increase
  • 14. Market development 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 14 Present aims • Create demand • Establish channel partnerships • Increase market knowledge and visibility • Shape minimum viable products Present activity • Engagement with market influencers • Meet with potential channel partners • Interviews with early adopters • Online surveys / calls for proposals
  • 15. Market makers 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 15
  • 16. Hardware platforms 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 16 MLaaS MLoaCMLoaB MLbD
  • 17. CORPORATE INFORMATION Artificial Learning 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 17
  • 18. 1980 • Co- founders met 2009-2010 • Frameworks for ultra- efficient machine learning • Project formalised 2011-2012 • Pigeon Consortium formed • CMOS feasibility study with Imperial College • Artificial Learning Ltd founded 2013- 2014 • Proof of concept plans laid • Market strategy and roadmap Company history 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 18 Pigeon Consortium research group • Artificial Learning Ltd • University of Edinburgh • Scottish Microelectronics Centre • University of Stirling
  • 19. Core team • Artificial Learning Ltd artificiallearning.com @ArtificialLearn Peter Newman PhD, Michael Bate MPhil • Key concepts and IP for machine learning hardware • Mathematical design and simulation • Commercialisation • Pigeon Consortium University of Edinburgh, University of Stirling • Our research collaboration 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 19
  • 20. Partnership and allies • Prototype R&D – Hardware design: Pigeon Consortium – Research funding: EPSRC – Chip fabrication: Europractice members • Channels – Developing these now – End-user product manufacturers – Maker community • Other alliances – Software engineers – London Tech City network – UK Electronic Systems Community – Deep learning and machine learning 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 20
  • 21. Risk reduction People Technical Market Financial 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 21 Research group formed and product development team skills identified Feasibility study completed and R&D plan agreed Current focus of risk reduction efforts: market gaps, value, channels R&D costs understood but production and sales costs to be determined
  • 22. Investment Readiness 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 22 4 9. Validate metrics that matter 8. Validate left side of canvas 7. Prototype high-fidelity MVP 6. Validate right side of canvas 5. Validate product/market fit 4. Prototype low-fidelity MVP 3. Problem-solutions validation 2. Market size/competitive analysis 1. Complete first-pass canvas Tasks in hand to move to IRL 5: • Identify real-world applications and quantify markets • Lean prototyping minimum viable product designs tested with potential early adopters • Develop channel partnerships
  • 23. Technology Readiness - MLoaC 2014-08-06 Copyright © 2014 Artificial Learning Ltd All rights reserved 23 4 9. Flight-proven in operations 8. Flight-qualified in test/demo 7. Prototype in live environment 6. Prototype in relevant environment 5. Component validation relevant env. 4. Component validation lab environment 3. Analytical/experimental proof of concept 2. Technology concept/applications formed 1. Basic principles observed and reported Tasks in hand to move to TRL 5: • Obtain funding for Pigeon Consortium development programme • Device functional simulations under way • Identifying IP alliances for next steps • Team building