# Bart the bartender

I just released a new chatbot called Bart on the Facebook Messenger platform. It can help you find cocktail recipes, tells jokes, and engages in smalltalk.

Check out the video below or chat with Bart here http://m.me/bart.the.bartender. Be sure to like its Facebook page https://www.facebook.com/bart.the.bartender!

The dialogflow.com platform takes care of intent classification and named-entity recognition. The chatbot uses TheCocktailDB to search for recipes. The dialogue manager, a Prolog program that runs on my VPS, handles the state and flow of the conversation. It’s the same dialogue manager that I also used in other chatbots like Dr. Stat.

# Dr. Stat

This is a demo of my chatbot called Dr. Stat.

It can answer questions about statistics and help a user select an appropriate statistical technique.

I’ve been working on this for quite a while now. Chatbots certainly aren’t anything new, but chatbots that are useful and display a more ‘natural’ way of conversing are rare.

I chose the domain of statistics for two reasons. Firstly, many of my students (and researchers I know) struggle with statistical concepts and I wanted to create something that would make their lives easier. Secondly, this domain has proven to be suitable for trying out different ideas about dialogue management.

Read on for a transcript with annotations.

# Skyline puzzle

One of my coworkers brought the following puzzle to work:

The puzzle is called Skyline and it’s a packing puzzle. The objective is to place the metal rod in one of the holes in the base and place the nine wooden pieces around it. It was designed by Jean Claude Constantin. When solved, the puzzle looks something like this:

Sometimes with these kinds of puzzles it’s quicker to write a program that finds a solution than trying to solve it by hand. Check out this github repository for a Prolog program that finds solutions for a given rod location.

To use this program open the file skyline.pl in your favorite Prolog interpreter (e.g. SWI-Prolog) and execute the following:

```?- print_solution(pos(4,4)).
+-------+
|ggeeeeh|
|gggdehh|
|ggiddhh|
|iii ddh|
|iiicfff|
|bicccff|
|baaaaaf|
+-------+
true```

You can press ; to find alternative solutions. The pos(X,Y) part refers to the location of the metal rod.

# MLP in C++11

In this post I present my Christmas gift to the world: A multilayer perceptron written in C++11.

I mainly wrote this to get some practice with some of the new C++11 features such as variadic templates and lambda functions. It uses template metaprogramming to construct (but not train) the neural network at compile time. You can download the code from its github repository. It’s lacking proper documentation, but I’ve included two examples that should get you started: the xor problem and Fisher’s Iris data set.

Happy Holidays.

# Bridge and torch puzzle

Four people need to cross a bridge at night which only supports two people at the same time. Person A needs 1 minute to cross the bridge, B needs 2 minutes, C needs 5 minutes and D needs 10 minutes. When two people cross the bridge they move at the slowest person’s pace.  They have a torch which has battery left for only 17 minutes. They can’t cross the bridge without light. How can they manage to cross the bridge?

One might guess that an obvious solution would be to let the fastest person (A) shuttle each other person over the bridge and return alone with the torch. This would give the following schedule:

 A, B -> 2 A <- 1 A,C -> 5 A <- 1 A,D -> 10

The total duration of this schedule would be 19 minutes, so the torch would run out of battery while person A and D are still on the bridge.

The optimal solution consists of letting the two slowest people (C and D) cross the bridge together, giving the following schedule:

 A, B -> 2 B <- 2 C,D -> 10 A <- 1 A,B -> 2

Which gives a total crossing time of exactly 17 minutes.

Writing a prolog program to solve this kind of river crossing problems is a walk in the park. Check it out if you want to know an alternative solution.

# Checkers game

Long, long time ago, when I was still a full time student, I became first in a checkers tournament with a program I wrote. Alas, those days of glory are over.

Looking back, the program was pretty simple. Choosing an efficient board representation made generating the moves a piece of cake. My program didn’t make use of any opening libraries or endgame strategies, it just implemented minimax search with alpha-beta pruning.

Although my program wasn’t very sophisticated, I still think it was pretty sweet and felt disappointed when I realized I had lost the source code. That’s why I decided to implement this game again only this time in Java (the original was written in C++). Check out my code here.

# Einstein’s riddle

The following puzzle is said to be invented by Einstein. Supposedly, he also claimed that only 2% of the world’s population would be smart enough to solve it.

There are 5 houses in 5 different colors in a row. In each house lives a person with a different nationality. These 5 owners drink a certain drink, smoke a certain brand of cigar, and keep a certain pet. No owners have the same pet, smoke the same brand of cigar or drink the same drink.

The question is: WHO OWNS THE FISH?

HINTS:

• the Brit lives in the red house
• the Swede keeps dogs as pets
• the Dane drinks tea
• the green house is on the immediate left of the white house
• the green house owner drinks coffee
• the person who smokes Pall Mall rears birds
• the owner of the yellow house smokes Dunhill
• the man living in the house right in the center drinks milk
• the Norwegian lives in the first house
• the man who smokes blends lives next to the one who keeps cats
• the man who keeps horses lives next to the one who smokes Dunhill
• the owner who smokes Bluemaster drinks beer
• the German smokes prince
• the Norwegian lives next to the blue house
• the man who smokes blends has a neighbor who drinks water

Working out the solution with nothing more that a pen and some paper is certainly doable by, I suspect hope, a larger percentage of people than the 2 % mentioned above. But as an example of how to solve these kinds of logic puzzles using Prolog, I wrote this code.

# Sudoku solver

It’s pretty straightforward to make a Sudoku solver in Prolog especially when applying CLP (Constraint Logic Programming).

Here is how to use my program:

`:- solve_sudoku.`

Then you can enter the known numbers one by one.

```+-+-+-+-+-+-+
|9 8| 5 |  7|
|72 |1```

When complete, the program determines and prints the solution.

```+-+-+-+-+-+-+
|938|254|617|
|726|193|854|
|514|876|329|
+-+-+-+-+-+-+
|692|387|145|
|187|465|293|
|453|921|768|
+-+-+-+-+-+-+
|345|712|986|
|871|649|532|
|269|538|471|
+-+-+-+-+-+-+```

Typing

`:- sudoku(L), write_sol(L).`

Gives

```+-+-+-+-+-+-+
|123|456|789|
|456|789|123|
|789|123|456|
+-+-+-+-+-+-+
|214|365|897|
|365|897|214|
|897|214|365|
+-+-+-+-+-+-+
|531|642|978|
|642|978|531|
|978|531|642|
+-+-+-+-+-+-+```

By pressing ; over and over again, you could enumerate all 6,670,903,752,021,072,936,960 possible Sudoku solution grids, but this might take a while..

It shouldn’t be too hard to extend this program to actually create new puzzles. If anyone does, let me know.

The prolog environment I used here is SWI-Prolog.