Lecture notes

My new Webador website is via ailooksatai.uk 

Page under construction

from 04 04 2025

Dear Dr Gunasekera   

Would you mind please provide material or set up an answer via your own means relating to the following notes.  I may well have missed important concepts etc.  Many notes have been extended.

AFH

--------------------------------------------------------------------------------

Applications of AI

1. Alphago - chess etc  protein structure prediction.
2. 
   
3. Chatgpt.com -  generative - Ode to Oxford in new text out.
4. 
   
5. Dall Energy oil and gas energy.
6. 
   
7. IKEA Case Study - AI integrated retail study.
8. 
   
9. Artificial General Intelligence Risk  which if any?  societal impact of AI?
10. 
    
11. Risks NIST risk management.
12. 
    
13. Chart?  How general AI could affect UK work force?
14. 
    
15. Book - Standard Text? - machine learning?
16. 
    
17. Graphic on screen - large circle - top - AI then ML, then deep learning.  Message?
18. 
    
19. Artificial Neural networks - neuron model - activation function.
20. 
    
21. In artificial neural networks, the neuron model utilizes an activation function to determine a neuron's output. This function takes the neuron's weighted sum of inputs and a bias, then transforms it into a single output value. The activation function decides whether the neuron is "fired" or not, effectively controlling the strength of the signal passed to subsequent layers.  source
22. 
    
23. Neuron network model - input value - output.
24. 
    
25. Forward propulsion and back propagation.
26. 
    
27. In the context of neural networks, forward propagation is the process of feeding input data through the network to generate an output, while backpropagation is the process of calculating and propagating errors backward through the network to adjust weights and biases, enabling the network to learn and improve its predictions. source
28. 
    
29. Supervised learning - features- labels.
30. AI Overview     

    Learn more
    In supervised learning, features are input variables used to make predictions, while labels are the known or correct output values that the model aims to predict.            source
31.   source
32. 
    
33. Unsupervised learning -
34. 
    
35.  input data produces algorithm output.
36. 
    
37. In contrast, unsupervised learning algorithms work independently to learn the data's inherent structure without any specific guidance or instruction. You simply provide unlabeled input data and let the algorithm identify any naturally occurring patterns in the dataset. source
38. 
    
39. It determines the similarities between a set of unlabeled input data by clustering sample data into different groups based on the similarities between them. Contrary to supervised learning, unsupervised learning has no output associated with its inputs and no supervisors. source
40. 
    
41. Reinforcement learning  -  input?         action 1      environment?
42. 
    
43.                                    ^ -----------------------^
44.                                                observation 
45. 
    
46.  
47. Language model        input layer - hidden layers?
48. 
    
49. Transformer - pairs of words
50. 
    
51. ?  nvidia gpu hardware  ?
52. 
    
53. Machine learning software
54. 
    
55. Trump - removing barriers to USA leadership in USA  -- ??
56. 
    
57. Wrap up - level of energy.

My new Webador website in draft format is via https://temp-xmaknfvdxllaxakvoaku.webadorsite.com/

Oxford AI

pagetop