What should I learn?

AI and psychology is a complicated field and it is easy to get lost in technical details. Below we summarize broad questions that we students are expected be able to answer following the course and possible in the examination:

 

  • Ethics
    • Describe ethical concerns, or risks, regarding AI in the past and the presence
    • Describe advantages with AI in the past and the presence
    • Discuss possible ethical concerns, or risks, regarding AI in the future
    • Discuss possible advantages with AI in the future 
  • Abstract neural networks (ANN)
    • What are the basic concepts in neutral networks and why do we need them?
    • Compare fundamental concepts in abstract neural networks with biological neural networks
    • How is information represented in neural networks? What are the advantage and disadvantages with this representation?
    • How is retrieval conducted in neural networks?
    • What is an activation function and why do we need it?
    • What is a learning rule? Describe hebbian learning rule and why the weights are changed as they are.
  • Human memory and neural networks
    • Describe how basic human memory phenomena can be accounted for in neural network terms. 
    • How does forgetting occur in neural networks?
    • How can reaction times be simulated in neural networks?
    • List basic neural articitures and what they are used for, for example auto-associative, heteroassociative networks, single versus multiple layer networks. 
    • How can recall and recognition of high high and low frequency words be simulated in neural networks? 
  • Deep neural networks (DNN)
    • What is a deep learning? Why do we need several layers in neural networks?
    • Describe how learning occurs in multiple layer networks? 
    • Describe some technical applications of neural networks.
    • What are the arguments for claiming that neural networks intelligent, or not intelligent? 
    • What is transformer model and what can it be used for?
    • Describe how convolution neural network (CNN) model works and what it can be used for.
  • Natural language processing (NLP)
    • What is NLP?
    • How does Latent Semantic Analysis (LSA) work and what are the disadvantages and advantages with it?
    • How does transformer based NLP model work and what are the advantages with this model?
    • Discuss the advantages and disadvantages of using NLP methods to analysis language data compared qualitative methods!
    • What is wrong with this statement: "I am using qualitative text data". 
    • What are the advantages and disadvantage with questions based computational language assessment (QCLA) compared to rating scales?
    • Describe a method for quantifying a psychological construct based on language NLP/ML and compared it to quantify the construct with rating scales. 
    • What is LIWC and how has it been used?
    • How well can QCLA measure psychological constructs? Give example on data showing that QCLA has higher validity than rating scales
  • AI based text and picture generation