い
し
未来
設計
How should we design the future? -
Predictive Analytics -
!～BIG DATA VISUALIZATION～!
[SVIF] 6月定例講演会
2017 6月15日＆木＇
Wilson Sonsini Goodrich & Rosa7 650 Page Mill Rd, Palo Alto, CA 94304
Agenda
•
Introduction of VALUENEX
•
What is Panoramic View Analytics
•
Panoramic View Analytics and Innovation
•
Case Study for M&A, Politics,
Investments…
•
Conclusion
Moore's law will end
•
What is Moore’s law
o
Moore's law is the observation that the number of
transistors in a dense integrated circuit doubles
approximately every two years.
•
Competition and collaboration between Hardware
revolution and Software innovation
•
After ending of Moore’s law, Algorithm will be more
significant field.
o
AI :
AlphaGo beat Ke Jie, the world No.1 ranked player at the time,
in a three-game match.
Keywords
© VALUENEX 2016 4
SILICON
VALLEY
STARTUP
LONG LIFE PRODUCTS
PANORAMIC VIEW ANALYTICS
WHITE SPACE
INNOVATION
SYNERGY EFFECTS OF ALLIANCES
M&A
PREDICTIVE ANALYTICS
DATA SCIENTIST
AI
IoT is not the destination
!
© VALUENEX 2017
intellectual innovator 5
It’s
how
we get there.
!
IoT
Case Study
© VALUENEX 2016 6
•
Google vs Honda
•
Apple Invest huge amount to Startup
company of MicroLED
•
Sharp vs Hon Hai
•
Google's IP strategy
•
SoZ Bank and ARM case
•
Renova7on of Automo7ve industry
PANORAMIC VIEW ANALYSIS
!
WHY VISUALIZE?
Making sense of the flood of Big Data
!
is like finding a needle in a haystack.
!
量
毎
%
増加
到達
単 ＆記号＇
kB) ^
MB) ^
GB) ^
TB) ^
PB) ^
エ サ ^
ZB) ^
YB) ^
Big Data:
増大
情報量
9
エ サ
10"
■人間
脳
容量
何
く
い
→
両
あ
■文字
理解
一目見
理解
方
早い
可視化
文字
観
言葉
変換し
頭
中
辞書
文書
ン
マッ
ン
行い
文脈
組
立
理解し
文書
種類
ワ
置
換え
分類し
い
逐次理解
可視化
集合体
類似性
置関係
一目
情報
し
入
く
面的理解
A picture is worth 1000 words
!
TEXT SOUND VISUAL
PICTURE MOVIE
Amount of data 1 < << <<<<
Pros Easy
Can be done simultaneously with other tasks Can be understood at a glance Feels complete Cons Incremental work, time-consuming Requires signal transfer
May require a caption
Incremental work,
time-consuming
Reading vs. Viewing
!
Which would you rather
navigate with?
!
13
© VALUENEX 2017
Which is more maneuverable?
!
Primitive Map
Accurate Map
© VALUENEX 2017
エ ブサマ
!
増加
電池
素材領域
No.1 代表特許 発明 称
NONAQUEOUS-ELECTROLYTE SECONDARY BATTERY
No.2 代表特許 発明 称
LITHIUM SECONDARY BATTERY AND ITS MANUFACTURING METHOD
No.3 代表特許 発明 称
MOTOR
No.4 代表特許 発明 称
WIPER DEVICE
No.5 代表特許 発明 称
ELECTROMAGNETIC SHIELDING TUBE
抽出条件
全期間＆2005-2015 ＇ け 最小自乗 法 回帰直線 傾 0以
全期間 け 2015 公報件数 最大 全期間 け 2013-2015 公報件数
エ ブサマ
!
VS
欧州自動車産業
ELECTRIC VALVE
POWER SUPPLY UNIT
Steering device
FUEL SUPPLY DEVICE
Vehicle cooling system
CHARGING EQUIPMENT
INFORMATION PROCESSOR
IMAGE PROVIDING DEVICE VEHICULAR LIGHTING FIXTURE
特許
ボッシュ 公報 5％以 占 他
ン 公報 5％以 占 両社 公報 5％以 占
ボッシュ POWER SUPPLY UNIT ELECTRIC VALVE領域 優 性
ン INFORMATION PROCESSOR Vehicle cooling system SEMICONDUCTOR DEVICE 領域 優 性 Fuel Supply Device 共通
主要自動車 注力技術!
主要自動車
別注力領域
比較
TOYOTA MOTOR HONDA MOTOR NISSAN MOTOR
密度 高密度
集計 範 ：1 準化範 ：5 距離減衰：2 最大値：32 最小値：1
FUEL SUPPLY DEVICE
CHARGING EQUIPMENT
VEHICLE SEAT
Steering device
FUEL CELL SYSTEM
TOYOTA
網羅的
HONDA : ICT
制御以外
コン
ンシ
領域
NISSAN
車体
内装等
比較的手薄
DISCOVERING M&A
OPPORTUNITIES
!
CASE STUDY: Hon Hai and SHARP
SHARP and Hon Hai in 2012
!
© VALUENEX 2016 19
Hon Hai Sharp Both
Until 2012, there were few
common technical fields
between SHARP and Hon
Hai.
Sharp asked Hon Hai to
invest but failed.
Hon Hai proposed a
US$13m investment in an
LCD company.
LCD
SHARP and Hon Hai in 2016
!
© VALUENEX 2016 20
Hon Hai Sharp Both
By 2016, Hon Hai
(Foxconn) had moved
closer to SHARP’s core
technology areas.
SHARP lost its technical
advantage in 4 years.
Was it ever worth $6bn?
(Final price was near $3.5bn
)
LCD
Solar Amplifier
Technology
Touch panel technology
“Competitive on “Internet of Things” (IoT) platform has
already begun
21
Google bought Nest
for $3.2 billion in
January 2014.
Android@Home
Automatic Summary Report
22
Summary Report
Can Nest provide a bridge
technology?
© VALUENEX 2017 23
Technological Positioning of
!
Soft Bank, ARM, SPRINT
25
Arm and Sprint will Move to Mobile Video
Communication Area including Medical
入力
R&D directions Become Closing
26
I
Case Study :
!
サ
ン掃除機
ン
特許
2012
出現 ＆
from 1993 to 2002
＇
28
1993‐2002 applica7on year
Cyclone System
Conven7onal Cleaner
Dyson’s patents appear
2002
以降
離
小島
大陸
繋
＆
from 2003 to
2013
＇
29
Cyclone System
Conven7onal Cleaner
Electric Cyclone Cleaner
Related Technology of
Electric Cyclone Cleaner
2003‐2013 applica7on year
2002
以降
離
小島
大陸
繋
＆
from 1993 to
2013
＇
30
Cyclone System
Conven7onal Cleaner
Electric Cyclone Cleaner
Related Technology of
Electric Cyclone Cleaner
Japanese Major Electric Companies rush here
小島
旗
立
！
Case Study :
ボッ
掃除機
ボッ
掃除機
俯瞰
33
従来型掃除機
ボッ
掃除機
主要各社
公開件数推移
34
2004
iRbot
社
ン
日本国内
販売開始
基本的特許
死滅し
い
35
• 国内企業 高死滅率
M&A &
!
INVESTING IN STARTUPS
!
CASE STUDY: LuxVue
Google and Honda
37
Should
compete or collaborate with
Two companies are in different industries.
!
But, …
38
HONDA motor corp. Google’s patents
Can Nest provide a bridge
technology?
© VALUENEX 2017 39
Patents Gravity are moving to Automobile area
2001‐2017
US Applica7on Patent : 2001‐2017
Synergy Cases
•
Synergy can be measured by
o
Distance of core technology between Both
companies
o
Width of Technology Fields
Our approach:
!
© VALUENEX 2016 41
SIMILARITY
ASSESSMENT
The algorithm scans data for similarities
CLUSTERING
Similar data is arranged into clusters
VISUALIZATION
Information is presented in a dashboard with a
3D map
IMPLEMENTATION
Strategic
decision-making based on analysis results
Our proprietary algorithm allows the accurate processing of
massive quantities of information.
Key Features
!
SPEED
cloud computing and
parallel processing allow
for high-speed analyses
© VALUENEX 2016 42
PANORAMIC
ANALYSIS
ACCURACY
uses 60 million dimensions
to analyze data
QUANTITY
analyzes and visualizes
100,000s of documents
Inspiring Growth
!
140+
clients
30%
" Fortune500 Japanese co.42%
"Thomson Global Innovator 100 Japanese co.
30
+
industries
automotive, electronic, chemical, precision equipment, financial…
© VALUENEX 2016 43
10
+
years
Case Studies: Unmet Needs
!
© VALUENEX 2016 44
Needs
Status Issue
Subject
clear?
Action
taken? Solution? Satisfactory?
Radar
Benefits
Case
Study
UNMET
Unknown ✗ ✗ ✗ ✗
Extract potential needs using characteristic words - Known, but not a high
priority ✓ ✗ ✗ ✗
Focus on emerging areas based on grant projects - Known, but no known solution ✓ ✓ ✗ ✗
Focus on different science & technology subjects using papers and patents Regenerative Medicine Known, but unsatisfactory solution
✓ ✓ ✓ ✗ Find emerging areas Google’s Nest technology
MET
Known, with satisfactory solution
Memorizing vs. Speculating
45
Memorizing Type
Specula7ng Type
Predictive Analytics is key
!
Big Data Analytics Predictive Analytics
ICT Cloud Data Scientist
Artificial
Intelligence
PAST PRESENT FUTURE
PREDICTIVE ANALYTICS &
INNOVATION
!
WHY DO WE NEED PREDICTIVE ANALYTICS?
Time frame comparison
!
© VALUENEX 2016 48
1968
1980
1983
33 YEARS
36 YEARS
33 YEARS
34 YEARS
2001 2016 2030 2050
Discover new services…
!
© VALUENEX 2016 49
ON THE
SURFACE
IN DEPTH
SERVICE 1 SERVICE 2 SERVICE 3
TECHNOLOGY A TECHNOLOGY B TECHNOLOGY C
ELEMENT TECHNOLOGY α
ELEMENT
TECHNOLOGY γ
…with Panoramic View
!
© VALUENEX 2016 50
ON THE
SURFACE
IN DEPTH
SERVICE 1 SERVICE 2 SERVICE 3
TECHNOLOGY A TECHNOLOGY B TECHNOLOGY C
ELEMENT TECHNOLOGY α
ELEMENT
TECHNOLOGY γ
1. Dig deep
2. Find common technologies 3. Reach new service areas
Reach new areas
!
© VALUENEX 2016 51
You are here.
All these areas are related at a deeper
level.
The birth of Great Innovations
!
In the beginning, there was White Space…
What
is
White Space?
!
•
Patent data is past data.
•
When planning the next product and/or
service, we need to find completely new
technical fields. This is where White Space
comes in.
Informative Accuracy and Reliability
•
We need to consider accuracy and reliability
•
DeNA stopped WELQ
which
is one of curation
site.
•
Brexit and Trump’s won
o
People have rely on Internet information beyond
existing media as News Paper and TV.
o
The Possibility to include Bogus intelligence on
Internet to rise up access number.
In particular Health and Medical Information should
be accurate and reliable.
Finding not only density but also White Space using
!
Accuracy Visualization
55
White Space
第
次
AI
ブ
© VALUENEX 2016 56
時期
主
ロ
チ
第一次
ム
1950
年代後
半～
70
年代
初期
[
推論
探索の時代
][
知能の時代
]
探索期
よる迷路やパ
ルの回
答
数学野定理の証明
第二次
ム
1980
年代
[
知識の時代
]
キ
パ
ム
理論
第三次
ム
2012
年頃～
[
機械学習と特徴表現の時代
][
ン
の時代
]
機械学習
ニュ
ラルネッ
ワ
深層学習
[
ラ
ニン
]
人工知能 金融 支配 日 櫻井豊 著
58
Data Fusion
59
• Patent Documents is very sophis7cated and clumped to handle and seeds
informa7on.
• When we think about future products, service, marke7ng data, review data,
product and service release informa7on are needed to analysis.
• Valuenex‘s DocRadar is the service for the aim to expand data source to
analysis.
• 特許文献 洗練さ 使いや い一群 あ 技術シ
情報
• 将来 商品 サ ビ 考え マ ン ビュ
商品 サ ビ 情報 解析 必要
• Valuenex DocRadar 解析 拡張 目的 サ ビ
