PwC Austria & Alpen Adria Universität Klagenfurt
Klagenfurt 2020
Backgrounds
History of Artificial Intelligence
Definitions
Terminologies
Machine Learning Tasks
Machine Learning Branches
Syllabus and Assessments
Text
Deep Learning Theory
Natural Language Processing
DL based Machine Vision
Reinforcement Learning
Machine Learning fundamentals
Statistical Machine Learning
Conventional Machine Vision
Text
Linear Regression
Logistics Regression
Shallow/Deep Neural Network
Back propogation algorithms
Convolutional Neural Network
Recurrent Neural Network
Semantic Segmentation
Generative Adversarial Network
Text
Machine learning is a type of artificial intelligence (AI)
that provides computers with the ability to learn without being explicitly
programmed.
Machine learning is a computer program is said to learn from experience E with respect to some class of tasks T and performance measure P , if its performance at tasks in T , as measured by P , improves with experience E .
Margaret Rouse
Tom Mitchell
A handwriting recognition learning problem:
Learning to recognize spoken words:
All of the most successful speech recognition systems employ machine learning in some form.
Alexa
Google Assistant
Zenbo Robot
Self Driving Cars
Nvidia, Uber, and Tesla all use deep learning as one of the main model in their self driving cars
Medical Imaging
Deep Learning is playing a key role in the future of medical imaging dignoses
Data
Features / Attributes
In case of vision system: colors, edges, motion filed, depth are examples of low level features
In case of acoustic system: frequency, pitch, amplitude are examples of low level features
In case of financial trading: the historical trends, statistical moments (mean, variance)
Features / Attributes
| House Size in | House Price in $ |
|---|---|
| 80 | 100k |
| 100 | 120k |
| 120 | 130k |
| 56 | 56k |
| 64 | 70k |
Decision Function/Hypothesis
In case of self driving cars: the output of the decision function is the steering angle
In case of voice recognition: the output selects a predicted action of a set of actions
Decision Function/Hypothesis
Model Training/Learning
Performance/Cost/Loss
where \(e_t = ( y^{\prime}- y_t) \)
1. Classification
Machine learning task are categorized by the target of the studies problems as follows:
is the problem of identifying to which of a set of categories (sub-populations) a new observation
1. Regression
Machine learning task are categorized by the target of the studies problems as follows:
is the problem of identifying the relationship (mathematical model) among different variables
3. Clustering
Machine learning task are categorized by the target of the studies problems as follows:
is the task of grouping a set of objects in such a way that objects in the same group (called a cluster) are more similar (in some sense or another) to each other than to those in other groups (clusters)
4. Time Series Forecast
Machine learning task are categorized by the target of the studies problems as follows:
Time series forecasting is the use of a model to predict future values based on previously observed values
The computer is presented with example inputs and their desired outputs, and the goal is to learn a decision function that maps inputs to outputs.
No labels are given to the learning algorithm, leaving it on its own to find structure in its input. Unsupervised learning can be a goal in itself (discovering hidden patterns in data) or a means towards an end (feature learning).
where the teacher gives an incomplete training signal: a training set with some (often many) of the target outputs missing.
A computer program interacts with a dynamic environment in which it must perform a certain goal (such as driving a vehicle or playing a game against an opponent[4]:3). The program is provided feedback in terms of rewards and punishments as it navigates its problem space
is a branch of machine learning that attempt to model high level abstractions. DL can be supervised, Unsupervised, Semi Supervised and Reinforcement learning. The common model used in DL is Neural Networks