Notes on codes, projects and everything
We were asked to develop a program that demonstrate Artificial Intelligence and apply what we had learned in class into the program. Then we came out with an idea to develop a program that is able to do tarot reading. At first the program did not have a GUI and is operated through Swi-Prolog interpreter but in the end we included a GUI-frontend for the program using XPCE.
First the program will require 3 random numbers from the user, then a random number will be generated as the seed. Then the card will be “shuffled” (in fact, we did some calculations to the user provided random numbers) and then determine the orientation of the card (normal and reversed). Then, we use the computed number to pick a tarot card and get the reading based on the orientation.
To shuffle the cards, we use a random number and place the random number into the formula such that $$c=x^y\text{ mod }23$$ where c is the corresponding card number, x is the user provided number and y is the system generated random number. Then the orientation of the card is determine by calculating the mod 2 of the random number.
To implement simple search strategy, we used a hash function to divide the total of 22 major arcana cards into 5 groups. Each group will have one individual database file storing the card characteristics and reading. Then, depending on the calculated number, the system will choose to query which database file to decrease the inclusion of database files. We may not be able to observe the performance enhancement as we only have 5 database files but it should be more obvious for system with more database files.
I am the developer for the GUI front-end. At first we planned to write the GUI using other languages/platforms but because of the time constraint we decided to give up on GUI front-end. However, I managed to pick up basic XPCE GUI construction just in time and made a GUI for the program. However, the GUI is not very friendly and is not able to display images. Because XPCE is a multi-platform supported toolkit therefore the appearance of the GUI is consistent in various supported operating system.
The program is reading input through prolog interpreter.
The program displays the result after computing the input.
The program is reading input through GUI
The program displays the result through the GUI
Recently I switched my search code to Annoy because the input dataset is huge (7.5mil records with 20k dictionary count). It wasn’t without issues though, however I would probably talk about it next time. In order to figure out what each parameters meant, I spent some time watching through the talk given by the author @fulhack.
Should have done this earlier, I was just being lazy to go through all the steps to publish it properly. So here it is, the full source is published to bitbucket. Feel free to fork the project if you are interested. I have not attach a licence to it but it will most probably be BSD licence. I have also uploaded the latest 0.0.2 release to bitbucket and would update the download link posted previously soon.
Although my supervisor strongly recommend using JENA for RDF related work, but as I really don’t like Java (just personal preference), and wouldn’t want to install JRE/JVM (whatever it is called) at my shared server account, so I went to look for an alternative. After spending some time searching, I found this library called Redland and it provides binding for my current favorite language — PHP, so I decided to use this for my RDF work.
A friend of mine recently posted a screenshot containing a code snippet for a fairly straight forward problem. So after reading the solution I suddenly had the itch to propose another solution that I initially thought would be better (SPOILER: Turns out it isn’t). Then mysteriously I stuck myself to my seat and started coding an alternative solution to it instead of playing Diablo 3 just now.
While working on a text classification task, I spent quite some time preparing the training set for a given document collection. The project is supposed to be a pure golang implementation, so after some quick searching I found some libraries that are either a wrapper to libsvm, or a re-implementation. So I happily started to prepare my training set in the libsvm format.