Course talk:CPSC522/Self Improving Machines

Hi Wenyi. I find your content interesting. There are a few notes on it. you may want to expand you abstract by giving more details on the discussions you will have. It seems a bit short to me. Although it is a theoretical page, adding some images can be done in some cases. It would also be nice to talk about the future works.

• The topic is relevant for the course. 5
• The writing is clear and the English is good. 5
• The page is written at an appropriate level for CPSC 522 students (where the students have diverse backgrounds). 5
• The formalism (definitions, mathematics) was well chosen to make the page easier to understand. 5
• The abstract is a concise and clear summary. 3 (you can give more details on what you are going to talk about and what will be the discussions)
• There were appropriate (original) examples that helped make the topic clear. 4
• There was appropriate use of (pseudo-) code. 4 (maybe it wasn't needed)
• It had a good coverage of representations, semantics, inference and learning (as appropriate for the topic). 5
• It is correct. 5
• It was neither too short nor too long for the topic. 3 (it could be a bit shorter on the paper review and then added your point of view in more details)
• It was an appropriate unit for a page (it shouldn't be split into different topics or merged with another page). 5
• It links to appropriate other pages in the wiki. 5
• The references and links to external pages are well chosen. 5
• I would recommend this page to someone who wanted to find out about the topic. 4
• This page should be highlighted as an exemplary page for others to emulate. 5

I would give it 17 in total.

Overall, I think it's a good page. I have some comments that maybe useful. You may want to explain a bit about the relation between the papers in a discussion section. Also, I think it would be a good idea to add some practical examples/figures to the page to make it more interesting.

I a scale of 1 to 5, where 1 means "strongly disagree" and 5 means "strongly agree" please rate and comment on the following:

• The topic is relevant for the course. 5
• The writing is clear and the English is good. 5
• The page is written at an appropriate level for CPSC 522 students (where the students have diverse backgrounds). 5
• The formalism (definitions, mathematics) was well chosen to make the page easier to understand. 4
• The abstract is a concise and clear summary. 5
• There were appropriate (original) examples that helped make the topic clear. 4
• There was appropriate use of (pseudo-) code. 4
• It had a good coverage of representations, semantics, inference and learning (as appropriate for the topic). 5
• It is correct. 5
• It was neither too short nor too long for the topic. 5
• It was an appropriate unit for a page (it shouldn't be split into different topics or merged with another page). 5
• It links to appropriate other pages in the wiki. 5
• The references and links to external pages are well chosen. 5
• I would recommend this page to someone who wanted to find out about the topic. 5
• This page should be highlighted as an exemplary page for others to emulate. 5

If I was grading it out of 20, I would give it: 18

Thanks for your advise. I'll add a discussion section. The problem of adding examples is that these concepts are just theoretical at current stage.

Comments

The chosen topic is very interesting and thought-provoking, and overall the article is enjoyable. Since the article deals with a rather abstract concept, it is understandable that most of the discussions are conceptual. Keeping that in mind, certain parts still need some clarification and context:

1. The explanation accompanying the utility function and the target theorem (for paper 1) only describe the mechanics of the equations, but do not provide enough insight into the meaning of the given equations.
   1. It is unclear what the term $\mathbb {E} _{\mu }$ is
   2. How do we determine an appropriate reward $r$?
   3. What does "environment distribution" mean? probability distribution of possible environment states?
   4. While the target theorem equation is straightforward, it should be described in plain English. e.g. utility of applying the rewrite is greater than then utility of continuing with the same program.

1. It is mentioned that paper 2 provides classifications for different types of RSI systems and their limits, but the distinction between different classes is a little fuzzy. In particular, "Definitions and Types of Self-Improving" section describes several concepts one after another without connecting the dots: relation between non-recursive optimization and law of diminishing returns; and Chalmer's proportionality thesis and number of possible improvements. Perhaps the section could be reorganized?

Minor comments

• The references made throughout the page should be cited (as you mention in "To Add" section)
• There are multiple grammatical and typographical errors; the page should be proof-read for the final draft.

Scheme

• The topic is relevant for the course. 5
• The writing is clear and the English is good. 3
• The page is written at an appropriate level for CPSC 522 students (where the students have diverse backgrounds). 5
• The formalism (definitions, mathematics) was well chosen to make the page easier to understand. 4
• The abstract is a concise and clear summary. 4
• There were appropriate (original) examples that helped make the topic clear. -
• There was appropriate use of (pseudo-) code. -
• It had a good coverage of representations, semantics, inference and learning (as appropriate for the topic). 3
• It is correct. 5
• It was neither too short nor too long for the topic. 4
• It was an appropriate unit for a page (it shouldn't be split into different topics or merged with another page). 5
• It links to appropriate other pages in the wiki. 3
• The references and links to external pages are well chosen. -
• I would recommend this page to someone who wanted to find out about the topic. 5
• This page should be highlighted as an exemplary page for others to emulate. 3
• If I was grading it out of 20, I would give it: 15