I have started this presentation by short explanation of what machine learning is.
Types of machine learning techniques.
Need for machine Learning.
Some Applications of Machine Learning
and Two Algorithm with example.
3. PRESENTATION OVERVIEW
โข INTRODUCTION TO MACHINE LEARNING
โข CLASSIFICATIONS OF MACHINE
LEARNING
โข NEED FOR MACHINE LEARNING
โข APPLICATIONS OF MACHINE LEARNING
4. ๏BEFORE KNOWING THE APPLICATIONS OF ML
๏LET US HAVE A SHORT INTRO ON MACHINE LEARNING.
7. ๏ถ Machine Learning is the way by which machines
are made intelligent by training them with some
datasets.
๏ถ It focuses on development of computer programs
that can learn automatically from experiences and
improve itself.
๏ถ It was coined by Arthur Samuel way back in 1959 at
IBM.
8. Classifications Of Machine Learning
1. Supervised Learning
๏ Machines are trained with well labelled data.
๏ After training, there accuracy is tested.
Classification
๏ Categories:
Regression
2. Unsupervised Learning
๏ถ No labelled data, work on data and fine some patterns.
๏ถ Group unsorted data according to their similarities and pattern
without any training.
๏ถ Categories: Clustering
Abstraction
9. 4. Reinforcement Learning
๏ผ No data is provided,
๏ผ It learns to react to an environment.
3. Semi-Supervised Learning
๏ง Given data mixture of Classified and unclassified
data.
๏ง Labelled is less in number
๏ง Unlabeled data is in abundance.
10. SOME EXAMPLES OF ALGORITHM ARE:
SUPERVISED LEARNING:
1. Regression,
II. DECISION TREE
III. KNN
IV. Logistic Regression etc.
UNSUPERVISED LEARNING:
I. k-means for clustering problems.
II. Apriori algorithm for association rule learning problems.
III. Hierarchical Learning etc.
REINFORMENT LEARNING:
I. Markov Decision Process
II. Q-Learning etc.
11. Need For Machine Learning ?
โข Machine can discover Hidden, non-obvious
patterns.
โข It helps us from writing long programs.
โข It solves big mathematics problem easily.
โข It needs only one time programming, it learns from
experiences.
โข It helps to detect unusual behaviors.
โข It fasten the normal working of life.
Many moreโฆ.
13. 1. Virtual Personal Assistants
๏ Siri, Alexa, Google Now are some examples.
๏ Operated with voice
๏ They help to find information when asked over voice.
๏ Activated with the help of voice.
ex: โHello Googleโ, โHey Alexaโ, and so on.
๏Integrated to variety of platforms, Example: smart
speakers, Mobile Apps, Smartphones, etc.
15. 2. Video Surveillance (uses Computer Vision)
โ It makes possible to detect crime before they happen.
โ Track unusual behavior of people like standing, seating, stumbling or
napping on benches, etc.
โ Sends detected data to human attendants.
โ If reported activities are counted true, they help to improve the
surveillance services.
โ It can provides various info about the vehicles too.
17. 3 . Stocks Predictions
๏ Analyses the previous behavior of the market
๏ Predicts rise and fall of stocks.
๏ Predicts the upcoming trend in markets.
๏ Market price of products in future.
๏ Also Learns from the current trends
๏ Upgrade itself automatically and periodically.
๏ Make itself more accurate
18. 4. Face Detection or Recognition
โฆ Real time face detection can be used for locating homeless
children.
โฆ It will help to stop human trafficking also.
โฆ It is used to safeguard banking and authenticate real
person.
โฆ It is used in many fields such as mobiles, computers, MNCs
to provide security and authentication entry.
โฆ Facebook uses this technique in the image tagging.
20. 5. Product Recommendations
โ Analyses our search from one site
โ Recommend the same products on all other
sites.
โ Used by E-Commerce sites like
Flipkart, Amazon, EBay, Snapdeal, etc.
โ Netflix
21. 6.Email Spam Filtering
โ System detects patterns in email.
โ Patterns like โEarn Cash โ, โ earn from home โ are likely to be
spam.
โ Use detected to make automated detections.
โ Over 3,25,000 malwares are detected everyday
โ Thus protects us from unwanted and hackable emails.
โ Email providers are using them
Such as Gmail. yahoo, Outlook, etc.
22.
23. 7. Weather Forecasting
โข Machines are trained with previous data.
โข After training they predict the weather.
โข Predicts Rain, cyclones.
โข Can even forecast the areas where it will rain
and where cyclone can hit.
24.
25. 8. DRIVERLESS VEHICLES
โข Vehicles sense the environment and move
with little or no human input.
โข Vehicles capable of navigating District
roadways
โข No driver is needed
โข Can overcome traffics successfully
โข It will Avoid accidents, traffic jam, etc.
โข It will reduce labor costs
โข Parking space can be fully utilized.
โข Companies working are: Google, BMW,
Tesla, BOSCH, Nissan, Ford, etc.
27. 9. SPEECH RECOGNITION
โข It is technology that recognizes Speech.
โข Translate spoken language into text.
โข It helps people with disabilities
โข Music composing, sound auto tuning.
Ex: Google translator
28. โข There is no doubt that voice-controlled programs such as
Appleโs Siri, Google Now, Amazonโs Alexa, and Microsoftโs
Cortana do not always understand our speech, but things are
likely to be improved in the near future.
โข With the passage of time, the accuracy of speech recognition
engines is increasing
29. 1o. Medical field
โ Assist Doctors in Diagnosis
โ It can help them to diagnose numbers of health problems such as
๏ง Breast cancer
โพ Lung cancer
โพ Prostate cancer
โพ Bone metastases
โพ Congenital heart defect
โพ Alzheimerโs disease
๏ง It can even determine the stage of cancer.
30. 11. AI Based Computer Vision
Computer vision tasks is method for acquiring, processing,
analysing and understanding digital images, and extraction
of high-dimensional data from the real world in order to
produce numerical or symbolic information.
31. APPLICATIONS OF COMPUTER VISION
โ Interaction, e.g., as the input to a device for computer-human
interaction;
โ Modelling objects or environments, e.g., medical image analysis or
topographical modelling;
โ Navigation, e.g., by an autonomous vehicle or mobile robot; and
โ Organizing information, e.g., for indexing databases of images and
image sequences.
32. SYSTEM MODEL
โ Image Acquisition- Getting images through camera
โ Pre-processing-it includes(Re-sampling, noise-reduction, Sale space etc.)
โ Feature extraction-identifying lines and edges
โ Detection and Segmentation-getting points of interests.
โ High level processing- small amount of data is fed based on certain
assumptions.
โ Decision making- making final decision required for the application
33. ALGORITHMS USED IN COMPUTER VISION
โ BASICS OF CONVOLUTION
โ THE SOBEL AND LAPLACIAN DETECTORS
โ IMAGE MOMENTS
โ PIXELS NEIGHBOURHOODS AND
CONNECTEDNESS
โ CONNECTED COMPONENTS LABELLING
โ K-MEAN CLUSTERING
34. The Sobel Edge Detectors
โ The Sobel edge detector is a gradient based method.
โ It works with first order derivatives.
โ It calculates the first derivatives of the image separately for the X and Y axes.
โ The derivatives are only approximations (because the images are not continuous).