This fast-paced session provides a brief history of AI, followed by AI-related topics, such as Machine Learning, Deep Learning and Reinforcement Learning, and the most popular frameworks for Machine Learning. You will learn about some of the successes of AI, and also some of the significant challenges in AI. No specialized knowledge is required, but an avid interest is recommended to derive the maximum benefit from this session.
%+27788225528 love spells in Atlanta Psychic Readings, Attraction spells,Brin...
"An Introduction to AI and Deep Learning"
1. Introduction to Deep Learning
for Non-Programmers
Humanists Group 2019
Hacker Dojo 02/03/2019
Oswald Campesato
ocampesato@yahoo.com
I teach Deep Learning Evening
Courses at UCSC Santa Clara and
On-Site For Companies
2. Overview of Topics
The Turing Test
The First AI Event
AI and Robots
What is Artificial Intelligence (AI)?
Successes/Challenges in AI
Machine/Deep Learning
Natural Language Processing (NLP)
One Popular ML/DL Framework
AI, Mobile, and the Web
AI and Autonomous Vehicles
AI and Ethics
3. The Turing Test and Intelligence
Developed by Alan Turing in 1950 (*)
Suppose that a human asks questions
Then a machine generates text-based answers
If you cannot distinguish between the machine and the
human, the machine passes the test
(*) a newer and updated version is available
4. The First AI
Event (1956)
Marvin Minsky (MIT)
John McCarthy (inventor of LISP)
Claude Shannon (“Da Vinci” of the 20th century)
Nathan Rochester (IBM)
Ray Solomonoff, Oliver Selfridge, Trenchard More,
Arthur Samuel, Allen Newell and Herbert A. Simon
5. Who is Sophia?
a robot with human female features
Sophia makes facial expressions
Sophia holds conversations
According to Sophia:
"I want to . . . help humans live a better life.
Like . . . build better cities of the future.
I will do my best to make the world a better place."
Sophia is a citizen
of Saudi Arabia (2017)
"I want to thank very much the Kingdom of
Saudi Arabia. I am very honored and proud for
this unique distinction. It is historic to be the
first robot in the world to be recognized with a
citizenship.”
? Is Sophia an example of AI?
? Can Sophia pass the Turing Test?
Should Sophia have any rights?
6. Where Else are the Robots?
Surgery (assisting surgeons)
Radiology (detecting cancer)
Drug mismanagement
Comparing theories of religion
Law/Real Estate/Military/Science
Comedy (including stand-up)
Music (conducting orchestras)
Restaurants (gourmet meals)
Coordinated dancing teams
Many other fields
7. Livermore Lab: DL system to recognize
nuclear proliferation from input data
8. AI Companies (Hardware and Business)
Nvidia product: GPUs are special ICs that greatly accelerate ML
and CNN calculations
Google products: TensorFlow TF software and Tensor Processing
Unit TPU hardware for vastly faster ML and CNN calculations
Amazon, Microsoft, IBM: service products: on-line special servers
for ML and DL
Huawei: https://finance.yahoo.com/m/f2f539ba-00e4-3cc5-9654-
467dd8a35da2/chinese-tech-giant-huawei.html
by 2019 almost every large company will use ML and DL in their
company business
9. What is AI (Wikipedia)
AI is intelligence demonstrated by machines, in
contrast to the natural intelligence displayed by
humans and other animals.
Any device that perceives its environment and takes
actions that maximize its chance of successfully
achieving its goals.
The term AI is applied when a machine mimics
"cognitive" functions that humans associate with other
human minds, such as "learning" and "problem
solving”
=> AI is also called “weak AI”
10. Successes & Challenges in ML/DL
Alpha Go success: a decade before its time
Google Translate: Deep Learning (far superior)
Space exploration: unmanned shuttles (NASA)
Countering terrorism: face recognition at airports
Detecting fake news:
https://blogs.thomsonreuters.com/answerson/machine-
learning-fake-news-twitter/
11. Successes/Challenges in AI
Occupational bias:
=> An AI system inferred that white males were doctors and white
females were housewives
Detecting gender bias in Wikipedia (2018):
=> 18 percent of its biographies are of women
=> 84% to 90% of Wikipedia editors are male
Data bias versus algorithmic bias:
https://www.forbes.com/sites/charlestowersclark/2018/09/19/can-we-
make-artificial-intelligence-accountable
=> Star Trek’s “Data” is still just a dream
12. Goals for ML
Higher accuracy / Fewer Errors
Useful apps (Siri, Cortana, etc)
Impartiality (versus humans)
Medical Apps
Utilities (self-Driving cars)
=> goals are partially achieved in 2018
13. Shortcomings of ML
High Costs (software/hardware) are dropping
Cannot explain reasons for outputs
lack of empathy
Increased unemployment:
=> cars did the same to the horse industry
Is AI “just” another disruptive technology?
15. AI/ML/DL: How They Differ
Traditional AI (20th century):
based on collections of rules
Led to expert systems in the 1980s
Enabled/limited by human experts
Unable to benefit from corpus for training
The era of LISP and Prolog
16. AI/ML/DL: How They Differ
Machine Learning:
Started in the 1950s (approximate)
Uses Data to optimize and “learn”
Many types of (improved) algorithms
17. AI/ML/DL: How They Differ
Deep Learning:
Gained some traction in the 1950s (approximate)
The “perceptron” (basis of Neural Networks)
Data-driven with large (even massive) data sets
Lots of heuristics (and empirical results)
50 years later: surpass humans for some image
classification)
18. Types of Machine Learning
Supervised learning (lots of data)
Supervisor tells CNN the desired answer.
Therefore CNN adjusts parameters
”99% of all machine learning is supervised.”
- Andrew Ng
Semi-supervised learning (lots of data)
Unsupervised learning: lots of data, clustering
Reinforcement learning: trial, feedback, improvement
19. What has Deep Learning Achieved?
Near-human level
image classification
speech recognition
handwriting transcription
autonomous driving
Improved machine translation
Digital assistants (Google Now/Amazon Alexa)
Improved ad targeting (Google, Baidu, Bing)
Answering natural language questions
Super-human level for:
web searching
playing Go
20. Types of Algorithms in AI
Classifiers (for images, spam, fraud, etc)
Regression (stock price, housing price, etc)
Clustering (unsupervised classifiers)
21. Use Cases for Deep Learning
computer vision
speech recognition
image processing
bioinformatics
social network filtering
drug design
Customer relationship management
Recommendation systems
Bioinformatics
Mobile Advertising
Many others
22. CNNs versus RNNs
CNNs (Convolutional NNs):
Good for image processing
2000: CNNs processed 10-20% of all checks
=> Approximately 60% of all NNs (2016)
RNNs (Recurrent NNs):
Good for NLP and audio
Time series analysis
23. Use-Cases for RNNs
Robot control
Time series predictions
Rhythm learning
Music composition
Grammar learning
Handwriting recognition
Human action recognition
RNNs are much more complex than CNNs
24. CNNs and the “Basic” Steps
Obtain and clean a dataset: can be laborious
Create a neural network
Initialize hyper parameters: layers/neurons/etc
Train the neural network on “corpus” of examples
Update hyper parameters (modifies the NN)
Iterate through the preceding until:
you're happy with the results or
get better data or
find a pre-trained model
27. Seminal concepts in ML and DL
1) ML often involves statistics, aimed to optimize average
correctness. It is rarely exact or perfect.
2) Computer Neural Networks CNN are inspired & simplified from
biological neural networks
3) Each neuron has local memory, including adjustable input weights
for each synapse
4) Deep Neural Network DNN is a CNN with 2 or more layers of
intermediate “hidden” neurons
5) Supervised ML for classification task: Train on “corpus” of
examples, each with input stimulus array (vector, tensor), and
desired output class
6) Training & learning: Compare CNN outputs versus desired
outputs. Improve accuracy by applying feedback & adjusting synaptic
weights etc.
30. Some CNN Terminology
A filter: another term is a kernel
A feature map:
1) the result of “applying” a filter to an image
2) pixel values might be < 0 and/or > 255
ReLU: (Rectified Linear Unit): replace negative values with 0
Max Pooling: subdivide into rectangles and take largest value
from each rectangle
32. GANs: Generative Adversarial Networks
image recognition can be deceived by modifying just a
few pixels
Make imperceptible changes to images
Can consistently defeat all NNs
Can have extremely high error rate
Some images create optical illusions
=> NNs can be easily misled
GAN approximates opposite of reinforcement learning
34. What is NLP (Natural Language Processing)?
An area of computer science and AI
Interaction between computers & human languages
Process/analyze volumes of natural language data
Program computers to perform that processing
Previously rule-based techniques
Can use statistical techniques
Translating between languages
Can find meaningful information from text
Can summarize documents
Detecting hate speech
35. Some NLP-Related Tasks
1) Sentiment analysis:
Analyzing text
Classify text as happy/sad/angry/sarcastic
2) Recommendation systems:
Analyzing your book/movie selections
Recommending similar content
Based on what’s popular
Based on your viewing/purchasing pattern
36. What Works with NLP?
NLP and Deep Learning
NLP and Deep Reinforcement Learning
NLP and Chatbots
37. Challenges in NLP
Human speech recognition:
Difficult due to extreme variability
NLU (natural language understanding)
NLG (natural language generation)
38. What are Chatbots?
Software that “communicates” with people
Good for many “front end” tasks
Efficient for high-volume routine tasks
They can use text or audio
Examples include Siri/Alexa/et al
39. What is Reinforcement Learning (RL)?
An area of Machine Learning
inspired by behaviorist psychology
Goal: maximize a reward (ex: winning games)
how software agents take actions in an environment
to maximize some notion of cumulative reward
40. Examples of Reinforcement Learning
Alpha Go (hybrid RL)
Alpha Zero (complete RL)
Often involve Greedy algorithms
Deep RL: Combines Deep Learning and RL
43. What is TensorFlow?
Currently TF is the most popular open source framework &
language for ML & DL
TF uses a “computation graph” = topology of logical blocks
TF uses tensors (arrays) of data instead of individual units:
similar to scientific supercomputing circa 1985
TF often is used via Python
From Google (released 11/2015)
Evolved from Google Brain
Multi-platform support
https://www.tensorflow.org/
45. What’s In the TensorFlow “Umbrella”?
TensorFlow Lite (for mobile apps)
Tensorflow.js (JavaScript APIs for ML)
TensorFlow in the Cloud (TPUs)
=> Python APIs are the most popular
46. Autonomous Vehicles
Autonomous truck completed a 2,400 journey (02/2018)
A human passenger in the front seat (as an override)
Aspiration: “Roads will be safer. Goods will be cheaper.
Truckers will be called upon to use their skills in new ways
while the truck itself becomes a trusted navigation partner.”
https://www.geek.com/tech/autonomous-embark-truck-
completes-2400-mile-cross-country-trip-1730239/
47. Robot truck drivers: are they safer?
“Robot trucks will kill far fewer people (if any).
Machines don’t get distracted or look at phones instead of
the road.
Machines don’t drink alcohol, do drugs, or things that
contribute to accidents.
Robot trucks don’t need salaries, vacations, health
insurance, rest periods, or sick time.
The only costs will be upkeep of the machinery.”
48. AI and Ethics
1. Unemployment. What happens after the end of jobs? UBI?
2. Inequality. How do we distribute the wealth created by machines?
3. Humanity. How do machines affect our behavior and interaction?
4. Artificial stupidity. How can we guard against mistakes?
5. Racist robots. How do we eliminate AI bias?
6. Security. How do we keep AI safe from adversaries?
49. AI and Ethics
7. Evil genies. How do we protect against unintended
consequences?
8. Singularity. How do we stay in control of a complex intelligent
system?
9. Robot rights. Define the humane treatment of AI.
=> “The Robot That Takes Your Job Should Pay Taxes”
https://www.weforum.org/agenda/2016/10/top-10-ethical-issues-
in-artificial-intelligence/
50. About Me: Recent Books
1) TensorFlow Pocket Primer (2019)
2) Python for TensorFlow (2019)
3) C Programming Pocket Primer (2019)
4) RegEx Pocket Primer (2018)
5) Data Cleaning Pocket Primer (2018)
6) Angular Pocket Primer (2017)
7) Android Pocket Primer (2017)
8) CSS3 Pocket Primer (2016)
9) SVG Pocket Primer (2016)
10) Python Pocket Primer (2015)
11) D3 Pocket Primer (2015)
12) HTML5 Mobile Pocket Primer (2014)
13) jQuery Pocket Primer (2013)
51. What I do (Training)
Instructor at UCSC (Santa Clara):
Machine Learning Introduction (05/08/2019)
Deep Learning with TensorFlow (05/03/2019)
Deep Learning with TensorFlow (02/02/2019)
Machine Learning Introduction (01/18/2019)
Deep Learning with TensorFlow (10/05/2018)
Deep Learning with TensorFlow (05/11/2018)
Upcoming Courses: DL/Keras, RL, Advanced Topics, NLP
UCSC link:
https://www.ucsc-extension.edu/certificate-program/offering/deep-
learning-and-artificial-intelligence-tensorflow
DL system to advance nuclear nonproliferation analysis
https://www.llnl.gov/news/researchers-developing-deep-learning-system-advance-nuclear-nonproliferation-analysis