Over-justification Effect

I am watching an online course on gamification. It mentioned explicit motivation and implicit motivation and described interesting examples of over-justification effect that subjects can be demotivated by rewards. The first one is about drawing by kids. Many kids are very interested in drawing and are inherently motivated to draw. However, when they are given…

TF-IDF

Here are some notes of TF-IDF as a measure for identifying keywords in a document. To efficiently indexing documents for future search, it is important to identify the keywords in a document. For a word to be a good keyword, first, it should be relatively rare. This can be measured by the inverse document frequency (IDF), which…

Locality Sensitivity Hashing (LSH)

I am studying an online class on Web Intelligence and Big Data. Dr. Shroff described a very nice example of LSH. Below are some quick notes. The idea of LSH is essentially the same as dimension reduction. We have some data vector that can be in very high dimension. By processing hash data instead, we…

Computer graphic course

I accidentally found a very interesting course from Berkeley today. The title is Mesh Generation and Geometry Processing in Graphics, Engineering, and Modeling by Jonathan Shewchuk.

Useful Tools

Just found a list of good tools suggested in Nuit Blanche. The list may be a bit dated but many of the tools are still very useful.

A not so rigorous introduction to finite fields

Field is meant to extend the most basic set of mathematical operations: addition, subtraction, multiplication, and division beyond real number. As subtraction and division are really inverse operations of addition and multiplication, we actually only need to consider the latter two. The set of all real numbers are closed under addition and multiplication. In the…

About method of types

Question: I was confused on the proof of Theorem 11.1.3 in yersterday’s class. And I am still confused after review the book and your notes many many times. I think that one of key points of Theorem 11.1.3 is that, under any source distribution , the upper bound and the lower bound is true for…