AIA Hsinchu Orange3

Dr. Squall Wu
Wjsquallwu@gmail.com
Orange3 Introduction

2Manufacturing Technology & Engineering Corning RestrictedDr. Squall Wu 2019
Dr. Chun-Yi (Squall) Wu
• Massachusetts Institute of Technology
– Smart Manufacturing Program 2019
• Taiwan AI Academy, Hsinchu 2018
• Glass Manufacturing Industry
– Image Classification, Text Mining, Deep Learning
• IBM Research Almaden 2017
– Nature Language Processing (SystemT)
• System Integration Company
– Feature Engineering, Predictive Maintenance, Industry 4.0
• Intellectual Property Service
– Patent Map (Visualization) & Patent Quality (PCA, BPANN)
• NTHU IEEM
– Neural Network, Clustering Algorithm, Genetic Algorithm, etc.
Taiwan AI Academy 校友會國際交流推動小組

ML Trend: Visual Programming

https://orange.biolab.si/

Download

Drag and Drop

Data Type, Data Table

Paint a two-dimensional data set

Preprocess
Distributions

Machine Learning

Hierarchical Clustering supports interactive cluster selection.

Interactive k-Means

Machine Learning

Regression model with Rank
1.人口中地位较低人群的百分数
房價
2.平均每居民房数
3.师生比例

Features Analysis with Scatter Plot
LSTAT 人口中地位较低人群的百分数
越高房價越低
RM 平均每居民房数
越高房價越高

Machine Learning

Tree model and Tree Viewer
哺乳類
羽毛
骨頭
鰭
無脊椎
腳
昆蟲爬蟲類

A classification task

Neural Network
• Neurons per hidden layer: defined as the
ith element represents the number of
neurons in the ith hidden layer.
• E.g. a neural network with 3 layers can be
defined as 2, 3, 2.
• Activation function for the hidden layer:
– Identity: no-op activation, useful to
implement linear bottleneck
– Logistic: the logistic sigmoid function
– tanh: the hyperbolic tan function
– ReLu: the rectified linear unit function
• Solver for weight optimization:
– L-BFGS-B: an optimizer in the family of
quasi-Newton methods
– SGD: stochastic gradient descent
– Adam: stochastic gradient-based
optimizer
• Alpha: L2 penalty (regularization term)
parameter
• Max iterations: maximum number of
iterations

Finding common misclassifications of different models.

Machine Learning
Image
Clustering

Image Embedding

Embedders
• InceptionV3 is Google’s deep neural network for image recognition. It
is trained on the ImageNet data set. The model we are using is
available here. For the embedding, Orange3 use the activations of the
penultimate layer of the model, which represents images with vectors.
• SqueezeNet is a deep model for image recognition that achieves
AlexNet-level accuracy on ImageNet with 50x fewer parameters. The
model is trained on the ImageNet dataset. Orange3 re-implemented the
SqueezeNet by using weights from the author’s pretrained model.
Orange3 use activations from pre-softmax (flatten10) layer as an
embedding.
• VGG16 and VGG19 are deep neural networks for image recognition
proposed by Visual Geometry Group from the University of Oxford.
They are trained on the ImageNet data set. Orange3 use a community
implementation of networks with original weights. As an embedding,
Orange3 use activations of the penultimate layer - fc7.

Different embedding: Squeeze, Inception v3, VGG16

Image for clustering

Image for tree classification

Image for classification model

RF & NN

Image embedding for a prediction task
correct
correct
3/4

Machine Learning
Text Mining
& NLP
Image
Clustering

NY Times

Preprocess Text & Word Cloud

Text Classification

Orange3
• 簡易上手的Data Analytics工具
• 可以處理數值，圖片，文字等多型態資料
• 透過Data pipeline，讓你的老闆知道你在做什麼
• 快速比較不同演算法，找到對的方向，再客製化模型
• Add-ons 好用但不太穩定
• 尚不支援ＣＮＮ，ＲＮＮ，ＲＬ等深度學習

AIA Hsinchu Orange3

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to AIA Hsinchu Orange3

Similar to AIA Hsinchu Orange3 (20)

Recently uploaded

Recently uploaded (20)

AIA Hsinchu Orange3