1. 利用事例にみる
AI 技術活用のポイントと
Microsoft AI 最新動向
畠山 大有 | Daiyu Hatakeyama | dahatake
Architect && Software Engineer && Applied Data Scientist (目指している)
Microsoft Japan
次世代産業フォーラム in KOBE 2019【AI
編】
2.
3. MSR Beijing
MSR Cambridge
MSR Redmond MSR Montreal
MSR New EnglandMSR New York
96%
RESNET vision テスト
94.9%
Switchboard テスト
89.4%
Stanford CoQA テスト
69.9%
機械翻訳
Research system
39.5 Tera FLOPS
Intel Stratix 10
MSR India
MSR Shanghai
人と同等の
機械翻訳
人と同等の
物体認識
Switchboar
d
Switchbo
ard
cellular
Meeting
speech
IBM
Switchboard
Broadcast
speech
人と同等の
音声認識
人と同等の
会話形式のQ&A
初めて FPGA を
データセンターに展開
5. • データを基にした統計に、機械
学習の推論結果も迅速に追加し
たい
(人員配置、価格設定、在庫設定、
リモート検査など)
• アプリケーションに
迅速に ML を組み込みたい
• ML 開発の生産性を上げたい
• Model の説明性や解釈性を上げた
い
• Model のライフサイクルを管理し
たい
• Automated machine learning
(Auto ML)
• Model explainability
and interpretability
• ML Ops
• Visual Interface
• Auto ML in GUI
• Cognitive Services
• ML Ops
• Auto ML in Power BI
Model
On-
premises
Cloud
Edge
22. Mileage
Condition
Car brand
Year of make
Regulations
…
Parameter 1
Parameter 2
Parameter 3
Parameter 4
…
Gradient Boosted
Nearest Neighbors
SVM
Bayesian Regression
LGBM
…
Mileage Gradient Boosted Criterion
Loss
Min Samples Split
Min Samples Leaf
Others Model
Which algorithm? Which parameters?Which features?
Car brand
Year of make
試行錯誤
23. Criterion
Loss
Min Samples Split
Min Samples Leaf
Others
N Neighbors
Weights
Metric
P
Others
Which algorithm? Which parameters?Which features?
Mileage
Condition
Car brand
Year of make
Regulations
…
Gradient Boosted
Nearest Neighbors
SVM
Bayesian Regression
LGBM
…
Nearest Neighbors
Model
繰り返し
Gradient BoostedMileage
Car brand
Year of make
Car brand
Year of make
Condition
24. Mileage
Condition
Car brand
Year of make
Regulations
…
Gradient Boosted
Nearest Neighbors
SVM
Bayesian Regression
LGBM
…
Gradient Boosted
SVM
Bayesian Regression
LGBM
Nearest Neighbors
Which algorithm? Which parameters?Which features?
繰り返し
Regulations
Condition
Mileage
Car brand
Year of make