mark my words

athiest father preying for a secular future

tf.contrib.learn

So it looks like Tensor flow has a highlevel API tf.contrib.learn which should be easier than the example I looked at yesterday.

It has a training dataset using IRIS (the flower), whereby data is imported via CSV file.  So I’ve formatted a few CSV files, and I’ll see if I can get my first run through.

I’ve formatted them the same as the IRIS data:

 

120 4 setosa versicolor virginica
6.4 2.8 5.6 2.2 2
5 2.3 3.3 1 1
4.9 2.5 4.5 1.7 2
4.9 3.1 1.5 0.1 0
5.7 3.8 1.7 0.3 0
4.4 3.2 1.3 0.2 0
5.4 3.4 1.5 0.4 0
6.9 3.1 5.1 2.3 2

It looks like 120 cause its got 120 records, 4 because there are 4 datapoints and then 3 because of the output of three different types of flower.

My data looks like this:

200,52,pass,1clubs,2clubs,3clubs,4clubs,5clubs,6clubs,7clubs,1diamonds,2diamonds,

3diamonds,4diamonds,5diamonds,6diamonds,7diamonds,1hearts,2hearts,3hearts,4heart

s,5hearts,6hearts,7hearts,1spades,2spades,3spades,4spades,5spades,6spades,7spade

s,1notrump,2notrump,3notrump,4notrump,5notrump,6notrump,7notrump

0,1,0,0,0,0,0,1,0,1,1,0,0,0,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,1,0,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,1,0,0,0,0,0,8

1,1,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,1,0,0,0,0,1,1,1,0,1,0,0,1,0,0,0,0,1,1,0,0,0,0,0,0,1,0,23

The 8 and the 23 were the highest scoring contracts given that hand for North. (1 diamond and 2spades)

Then I’ll call the sample python script and alter the querry data to see what it things I should make for a contract.

new_samples = np.array(

[[0,0,0,0,0,1,1,0,1,0,0,0,0,0,0,1,0,1,0,0,0,1,1,0,0,0,0,0,0,1,1,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,1,1,0,0,0,0], [0,1,0,1,1,0,0,0,0,0,0,1,1,0,0,1,0,0,0,0,0,0,0,1,1,0,0,0,0,0,1,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,1,0]], dtype=float)

February 1, 2017 AI Bridge Project

Podcast

Last night I listened to this podcast

https://twimlai.com/twiml-talk-8-diogo-almeida-deep-learning-modular-theory-inflexible-practice/

It was a real eye opener how much I have to learn!

AI Bridge Project

Welcome to Tensor

 

 

I’ve got Python installed, and Tensor installed.  I ran the sample program.  From what I gather it goes and gets the data from a website, then iterates through the training data.  And spits out some numbers.

It’ll likely take me awhile to figure out what all the things it is doing is, modify it to look at bridge data, even for a simple case.
tensor

 

January 31, 2017 AI Bridge Project

Using Deal to generate some training data

My plan to generate some training data for the Bridge AI is to use Thomas Andrew’s Deal program.  It has a built in double dummy solver.

It seems to use TCL.  Which has an odd syntax that parses commands using [] and literals as {}, but I think I’ve gotten the hang of it.

My first step was to create a custom format file to spit out a bridge hand.

1000000000000001000000010000001001110101010001001001

I plan to train my AI using input strings like this

With all the bridge cards represented 1 = you have that card.  Starting with the AS and going down to the 2C

Another input will be the vulnerability, and the bidding up to that point.

I think used the tricks function to find out how many tricks and the score.  The tricks function has two modes, double dummy and goal mode.

So if you are in 4S, you might want to give up a trick to give yourself an extra chance of making the contract.

I guess I’ll make my bidding program focus on total points for now.

AI Bridge Project

Bridge AI Project

I’ve decided to start a new project and I plan to spend at least a year on it.

I plan to teach myself deep learning. Already I notice that is is going to be hard.

I found this book online for free, which I plan to read

www.deeplearningbook.org

I also plan to use TensorFlow

AI Bridge Project

Neat hand from Buffalo

Bridge hand: 3NT by East

Bridge hand: 3NT by East

Shona opend 1D in the West. 2S by North and take-out double by me. 3D by Shona, and 3NT closed the auction.

Opening lead 6C, East pitched a heart and I won the 8C in my hand.

So it looks like 4 hearts, 2 spades, ace of diamonds and 1 club = 8 top tricks. After some thought, I realized I could cash 4 hearts and keep playing spades until North wins, having no clubs and no hearts left they will be forced to lead into the AQ of diamonds, for tricks 8 and 9.

October 24, 2016 bridge, buffalo

Quote from You Are Now Less Dumb: How to Conquer Mob Mentality, How to Buy Happiness, and Allthe Other Ways to Outsmart Yourself

Describing soldiers facing combat in Vietnam in The Things They Carried, Tim O’Brien wrote, “They were afraid of dying but even more afraid to show it,” adding later, “They carried the soldier’s greatest fear, which was the fear of blushing. Men killed, and died, because they were embarrassed not to.”

August 6, 2016 Uncategorized

Quote from Wherever I Wind Up

As 2007 goes on, more and more God helps me see and believe that it is time for me to stake out my own knuckleballing turf.

Lol, you think good gives a fuck about your knuckleball?

June 19, 2016 Uncategorized

Quote from Wherever I Wind Up

Maybe I should just end it now. Maybe it’s time for me to stop all this pain once and for all.

I think about options. Carbon monoxide? That’s a possibility. We just built a new garage; the old carport would’ve had too much fresh air.

A knife or a gun? Not in a million years. I don’t like blades or bullets.

Uncategorized

Quote from Chasing the Scream: The First and Last Days of the War on Drugs

We all know deep down it doesn’t make us happy, to be endlessly working to buy shiny consumer objects we have seen in advertisements. But we keep doing it, day after day. It in fact occupies most of our time on earth. We could slow down. We could work less and buy less. It would prevent the environment—our habitat—from being systematically destroyed. But we don’t do it, because we are isolated in our individual cages. In that environment, the idea of consuming less, in fact, fills us with panic. All this stuff, Bruce believes, is filling the hole where normal human connection should be.

Well said!

May 2, 2016 Uncategorized