Chimple v2.1.12

Updates:

• Added ChimpGamma invoked as chimpGamma(name, alpha, beta)
• Updated ChimpRand with a random walk option invoked as chimpRand(name,jump_parameter)
• MonkeyCage implements a MonkeyMap rather than a HashMap, which allows prefix trees. All interfaces with the MonkeyCage are backwards-compatible. The MonkeyCage parses a prefix as the string preceding the delimiter -, allowing Monkeys with that prefix to be requested from the MonkeyCage using the appropriate methods. This can be useful for tagging sets of variables in customs solvers (e.g. name = hyperparameters-hierachy)

Demos Quick Start:
Clone the repository and find the releases/chimple-(version).jar file, or download the JAR from below. Then simply run the JAR file (on many operating systems, this can be done by double-clicking) and the demo chooser window will appear.

Java Developer Quick Start:
Clone the repository and write a class that extends ChimpleProgram. You will be required to implement the run(Object... args) function. Write your probabilistic program here (for your first program, try a simple coin model). For help, see the demos.

Then, add a cost function, either inline using addEnergy() inside your probabilistic program, or by extending CostFunction. This cost function will represent data that you are conditioning on.

Finally, write a public static void main(String[] args) function that instantiates your program class and calls MHQuery(), and run the class.

MATLAB Developer Quick Start:
Clone the repository. If you wish to avoid compilation, download the chimple2 binary and place it in your external-libs folder. Equivalently, import the project into an IDE such as Eclipse that automatically compiles your project. In either case, now run startup.m in the base directory. Write a MATLAB function that implements your probabilistic program, calling addEnergy() to condition on data. Then write a MATLAB script that calls chimplify on your program, passing in arguments, burn-in iterations, sample iterations, and spacing iterations, and run that script.

Chimple v2.1.11

The Monkey interface has changed significantly since 2.1.10, breaking compatability.

Java Developer Quick Start:
Clone the repository and write a class that extends ChimpleProgram. You will be required to implement the run(Object... args) function. Write your probabilistic program here (for your first program, try a simple coin model). For help, see the demos.

Then, add a cost function, either inline using addEnergy() inside your probabilistic program, or by extending CostFunction. This cost function will represent data that you are conditioning on.

Finally, write a public static void main(String[] args) function that instantiates your program class and calls MHQuery(), and run the class.

MATLAB Developer Quick Start:
Clone the repository and run startup.m in the base directory. Write a MATLAB function that implements your probabilistic program, calling addEnergy() to condition on data. Then write a MATLAB script that calls chimplify on your program, passing in arguments, burn-in iterations, sample iterations, and spacing iterations, and run that script.

Chimple v2.1.10

Demos Quick Start:
Clone the repository and find the releases/chimple-(version).jar file, or download the JAR from below. Then simply run the JAR file (on many operating systems, this can be done by double-clicking) and the demo chooser window will appear.

Java Developer Quick Start:
Clone the repository and write a class that extends ChimpleProgram. You will be required to implement the run(Object... args) function. Write your probabilistic program here (for your first program, try a simple coin model). For help, see the demos.

Then, add a cost function, either inline using addEnergy() inside your probabilistic program, or by extending CostFunction. This cost function will represent data that you are conditioning on.

Finally, write a public static void main(String[] args) function that instantiates your program class and calls MHQuery(), and run the class.

MATLAB Developer Quick Start:
Clone the repository and run startup.m in the base directory. Write a MATLAB function that implements your probabilistic program, calling addEnergy() to condition on data. Then write a MATLAB script that calls chimplify on your program, passing in arguments, burn-in iterations, sample iterations, and spacing iterations, and run that script.

Chimple v2.1.9

Demos Quick Start:
Clone the repository and find the releases/chimple-(version).jar file, or download the JAR from below. Then simply run the JAR file (on many operating systems, this can be done by double-clicking) and the demo chooser window will appear.

Java Developer Quick Start:
Clone the repository and write a class that extends ChimpleProgram. You will be required to implement the run(Object... args) function. Write your probabilistic program here (for your first program, try a simple coin model). For help, see the demos.

Then, add a cost function, either inline using addEnergy() inside your probabilistic program, or by extending CostFunction. This cost function will represent data that you are conditioning on.

Finally, write a public static void main(String[] args) function that instantiates your program class and calls MHQuery(), and run the class.

MATLAB Developer Quick Start:
Clone the repository and run startup.m in the base directory. Write a MATLAB function that implements your probabilistic program, calling addEnergy() to condition on data. Then write a MATLAB script that calls chimplify on your program, passing in arguments, burn-in iterations, sample iterations, and spacing iterations, and run that script.

Chimple v2.1.8

Demos Quick Start:
Clone the repository and find the releases/chimple-(version).jar file, or download the JAR from below. Then simply run the JAR file (on many operating systems, this can be done by double-clicking) and the demo chooser window will appear.

Java Developer Quick Start:
Clone the repository and write a class that extends ChimpleProgram. You will be required to implement the run(Object... args) function. Write your probabilistic program here (for your first program, try a simple coin model). For help, see the demos.

Then, add a cost function, either inline using addEnergy() inside your probabilistic program, or by extending CostFunction. This cost function will represent data that you are conditioning on.

Finally, write a public static void main(String[] args) function that instantiates your program class and calls MHQuery(), and run the class.

MATLAB Developer Quick Start:
Clone the repository and run startup.m in the base directory. Write a MATLAB function that implements your probabilistic program, calling addEnergy() to condition on data. Then write a MATLAB script that calls chimplify on your program, passing in arguments, burn-in iterations, sample iterations, and spacing iterations, and run that script.