Week 2 Flashcards
(20 cards)
What are the 4 steps Polya’s heuristic strategies divided into?
- Understand the problem
- Devise a plan
- Carry out the plan
- Look back and reflect
What does (1) Understand the problem mean (Polya’s)?
Simply understanding the problem. A failure to understand problem explains a failure to solve it
What does (2) Devise a plan mean (Polya’s)?
Identifying the calculations, computations or constructions that must be performed in order to obtain a solution
What is (3) Carry out the plan mean (Polya’s)?
Performing steps in sequence to give the solution.
What does (4) Look back and reflect mean (Polya’s)?
Taking time to consider what worked and what did not. This makes it easier to find solutions in future problems
What is a Design Pattern?
Captures idea of a good solution to a common design problem
By using design patterns, novices behave like experts
What design pattern for functional programs are we going to use?
Generate-and-select
How does Generate-and-select work?
The generator constructs large number of values that might be solutions to problem
The Selector filters out values that are solutions to the problem
What is the Productivity Paradox?
Better technology should make things faster and easier, so productivity should increase
However in reality it does not increase as impressively as we may think
This is due to many reasons, some being:
- Takes time to learn new technology
2.Measuring productivity is hard
- Benefits are uneven (Technology might help one group (like online stores) while hurting another (like regular shops)
What is the Uneven/ concentrated gains explanation for the Productivity Paradox?
The gains are in few productive firms and sectors with limited weight in the overall economy
What is the Implementation lags explanation for the Productivity Paradox?
It takes considerable time for new technologies to achieve critical mass, or for necessary complementary ones to appear
What is the Mismeasurement explanation for the Productivity Paradox?
Adopting new technologies can lead workers to move from more productive adopting sectors to less productive ones, and so to negligible aggregate productivity growth
How do Artificial Neural Networks (ANNs) work behind the scenes?
- The core ANN is trained on text from internet to respond to a prompt with a list of most probable next words after prompt
- Core ANN is tweaked by scoring its responses to sample queries.
Second ANN is trained on these scores to predict most likely one to be assigned to a prompt
- Second ANN is used in reinforcement learning to adjust weights in the core ANN so that its outputs are even more likely to satisfy humans
- Sometimes, data from user responses to LLM responses is def back to fine-tune for still better results
What are the main concerns when using Large Language Models (LLMS)?
- Copyright
- Education
- Code Quality
- Code Security
- Not an expert
- What is the LLM Copyright concern?
Although the code produced is the meldingof that from many sources, there may still be licensing problems
- What is the LLM Education concern?
Good solutions to many assignments can be completed directly, reducing their challenge and worth
- What is the LLM Code Quality concern?
No guarantee of code quality when using a large language model, and a code review will need to be conducted before code is deployed
- What is the Code Security LLM concern?
There is no guarantee of code security when using a large language model, and a security audit will need to be conducted before the code is deployed
- What is the Not An Expert LLM concern?
A LLM is not an expert - it does not know that is does not know, and will waffle if necessary, mixing high-quality text with low-quality rubbish
