BERT を中心に解説した資料です.BERT に比べると,XLNet と RoBERTa の内容は詳細に追ってないです.
あと,自作の図は上から下ですが,引っ張ってきた図は下から上になっているので注意してください.
もし間違い等あったら修正するので,言ってください.
(特に,RoBERTa の英語を読み間違えがちょっと怖いです.言い訳すいません.)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
XLNet: Generalized Autoregressive Pretraining for Language Understanding
RoBERTa: A Robustly Optimized BERT Pretraining Approach
BERT を中心に解説した資料です.BERT に比べると,XLNet と RoBERTa の内容は詳細に追ってないです.
あと,自作の図は上から下ですが,引っ張ってきた図は下から上になっているので注意してください.
もし間違い等あったら修正するので,言ってください.
(特に,RoBERTa の英語を読み間違えがちょっと怖いです.言い訳すいません.)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
XLNet: Generalized Autoregressive Pretraining for Language Understanding
RoBERTa: A Robustly Optimized BERT Pretraining Approach
Scenario-based Economic Model Approach to evaluate the impact of the Internet...Yasushi Hara
Scenario-based Economic Model Approach to evaluate the impact of the Internet of Things: For the Creation of Policy Options of Science, Technology and Innovation Policy
31. Trends in the number of AI patents granted world-wide
Figure
• The number of patents granted has
rapidly increased.
• It shows the number of artificial
intelligence (AI) patents granted by
application country and reveals that it
has increased more than threefold
(from 708 items in 2012 to 2,888 items
in 2016).’
• In particular, AI patents granted in the
US increased by 1,628 items during this
period (Figure 1a), accounting for
approximately 75% of the increase
worldwide.
31(Source: Fuji and Managi (2017))
32. Trends in the number of AI patents, technology-wise granted worldwide
Figure
• The patent share of each AI technology type
changed from 2012 to 2016.
• In 2012, biological and knowledge-based
models were the leaders in patented AI
technologies.
• However, from 2012 to 2016, the number of
patents granted for specific mathematical
models and other AI technologies rapidly
increased, doubling from 2015 to 2016.
32(Source: Fujii and Managi (2017))
33. AI : Number of Patents/Scientific Papers by Year
Number of Papers Number of Patents
0
50000
100000
150000
200000
250000
USA PEOPLES R CHINA
GERMANY JAPAN
ENGLAND France
Canada ITALY
SPAIN Australia
0
200
400
600
800
1000
1200
1400
United States Japan Europe (Patent Office)
Germany Korea Unite Kingdom
France China Taiwan
Israel India
Source: Web of Science Core Collection Source: PatentsView(USPTO)
34. Robotics : Number of Patents/Scientific
Papers by Year
Number of Papers Number of Patents
Source: Web of Science Core Collection Source: PatentsView(USPTO)
0
50
100
150
200
250
300
350
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
USA PEOPLES R CHINA GERMANY
ENGLAND CANADA ITALY
JAPAN FRANCE SPAIN
AUSTRALIA
0
20
40
60
80
100
120
140
160
180
19761978198019821984198619881990199219941996199820002002200420062008201020122014
United States Japan Germany
Korea France Sweden
United Kingdom Italy China
Europe (Patent Office) Switzerland
70. 成績評価
• 平常レポート
• レポート1; max 15
• レポート2; max 5
• レポート3; max 20
• 最終レポート
• 40点
• 最終レポートの360°評価
• 10点
• 投票システムから集計したデータを按分します
71. 参考文献
• https://www.codexa.net/tensorflow-for-begginer/
• http://tekenuko.hatenablog.com/entry/2016/09/19/214330
• http://www.randpy.tokyo/entry/python_random_forest
• FUJII Hidemichi, MANAGI Shunsuke (2017) «Trends and Priority
Shifts in Artificial Intelligence Technology Invention: A global
patent analysis», RIETI Discussion Paper Series 17-E-066,
https://www.rieti.go.jp/jp/publications/dp/17e066.pdf
• The Economics of Artificial Intelligence: An Agenda,
https://www.nber.org/books/agra-1