1 -1 Introduction to machine learning.pptx

Machine learning presentation

1. Introduction to Machine Learning Dr Mehrdad Ghaziasgar Senior Lecturer in Computer Science Department of Computer Science University of the Western Cape [email protected] 1

2. Machine learning is concerned with software, algorithms and systems that learn from data Machine Learning Introduction

3. Machine Learning Introduction Definition: Samuel (1959): Machine learning is the field of study that gives computers the ability to learn without being explicitly programmed. Mitchell (1998): A well-posed learning problem: A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E. E.g. Training a computer program to detect spam / non-spam emails: T: detecting whether an email is spam / non-spam E: the program takes note of which emails you mark as spam (and which you don’t) and uses this knowledge. P: the probability of the program marking your next email correctly as spam / non-spam

4. Machine Learning Introduction Mitchell (1998): A well-posed learning problem: A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E. E.g. Training a computer program to detect whether an image of a tumour is malignant or benign: T? detecting whether a tumour in an image is malignant or benign. E? a “bunch” of images of benign and malignant tumours . P? the probability with which the program correctly detects the next image it is given

5. Machine Learning Introduction

6. Machine Learning Introduction Why has it become “hot” / important now? Faster hardware Improved algorithms Explosion of data Excellent and well-built frameworks

7. Spam detecti on Fraud detecti on Video surveillance Access control Malware detecti on Crowd Abnormal Behaviour Detection Use Cases – Data and Personal Security Introduction

8. Use Cases – Object Recognition Introduction

9. Use Cases – Financial Trading Introduction

10. Google reduces electricity consumption by 15% Equivalent to 55,000 households Saving 100s of millions of dollars Use Cases – Energy Optimization Introduction

11. Medical diagnoses (oncology) Personalized treatments Remote patient monitoring Epidemic Outbreak Prediction Research into medicines. Follow-up of patients. Radiology and radiotherapy. Use Cases – Healthcare Introduction

12. Tech Support Agents Chatbots Task-Oriented Dialog Agents (TODA) Cross-domain agents & general purpose AI 11 https://donotpay-search-master.herokuapp.com Use Cases – Conversational AI Introduction

13. Insight into customers. Personalized emails. Smart bidding ( ads ). Dynamic pricing. Lead scoring Predictive customer service 12 Use Cases – Sales and Marketing Introduction

14. Personalized recommendations.  Website personalization Real-time notifications. Collaborative filtering Content Filtering: i f you like this, you might also like this. Item-Item Collaborative Filtering : customers who liked this also like this. User-Item Collaborative Filtering : customers who are similar to you, liked this. 13 Use Cases – Recommender Systems Introduction

15. 14 Use Cases – Natural Language Processing Introduction

16. 15 Use Cases – Automatic Navigation Introduction

17. Style transfer Emotion Analysis Design Music compositions Pose estimation Image filtering 16 Use Cases – Multimedia Support Introduction

18. Machine Learning Introduction

19. Machine Learning – Context in AI Introduction Source:  SAS Institute - A Venn diagram that shows how machine learning and statistics are related

20. Machine Learning - Categories Introduction Supervised Learning Inputs with desired outputs are given Learn to mimic patterns in the data Unsupervised Learning R ein f or c ement Learning Decision process based on rewards Algorithm learns to respond to its environment The desired outputs are not given Discover patterns in data

21. Machine Learning - Categories Introduction

22. Machine Learning - Categories Introduction

23. Machine Learning - Approach Introduction

24. Machine Learning - Approach Introduction

25. THE END Of The Beginning