There are a number of demo projects included with the software, which are typically installed in
/usr/local/share/Emergent/demo (Linux, Mac) and live in the Emergent\demo application directory under Windows.
These can be useful for providing working examples of various types of models and procedures. Each contains its own more detailed documentation that should auto-open when you open the project, so this is just a brief listing of what is available.
See PublishedProjectList for a list of all demos and other projects that have been uploaded to this wiki, and can be directly downloaded here.
- 1 demo/programs (Programs)
- 2 demo/data_proc (DataProcessing)
- 3 demo/leabra (Leabra)
- 4 demo/network_misc
- 5 demo/actr (ACT-R)
- 6 demo/bp (Backpropagation)
- 7 demo/cs (Constraint Satisfaction)
- 8 demo/so Self Organizing
- 9 demo/virt_env Virtual Environment
- 10 demo/audioproc AuditoryProc
- 11 Computational Explorations in Cognitive Neuroscience Projects
- demos/program_examples -- simple examples of all manner of Program code, great for copy-and paste to do things you want!
- demos/random_numbers -- test of threaded random number generator, with histogram output.
- tutorials/data_tutorial -- takes you through the DataProcTutorial -- see tutorials/data_tutorial_final for the final completed version for comparison.
- demos/ca_eq_sim -- calcium equation simulation -- shows how to use a DataTable to graph equations (e.g., like gnuplot or similar), using a DataCalcLoop object and various other important data processing techniques.
- demos/canvas_drawing -- example of using the taCanvas object for using basic graphic rendering operations on an image
- demos/data_calc_loop -- focused simple example of the powerful DataCalcLoop program element.
- demos/epoch_log_analysis -- program that performs various standard analysis operations on epoch data logs generated as a network is trained -- tells you how many epochs it too to reach various criteria, etc (associated shell script makes a simple executable system for analyzing your data!)
- demos/matrix_demo -- how to use Matrix objects in Programs -- creates matrices from scratch, sets values, runs Math functions, etc.
- demos/fft_image_xform -- how to use fast fourier transforms on image data to achieve various forms of filtering etc.
- tutorials/ax_tutorial and tutorials/ax_tutorial_final -- see AX_Tutorial -- build a network from scratch to perform the CPT-AX task.
- demos/error_driven_hidden -- an project from the CECN book (see link below), demonstrating the power of error-driven learning.
- demos/param_search -- simple demo of Param Search using ClusterRun infrastructure.
- demos/scalar_val_test -- a test of the ScalarValLayerSpec, trained with different probabilities of a 0 or 1, showing that they accurately represent these probabilities as a graded scalar value.
This has misc projects for exploring various general neural network functionality:
- demos/projection_sampler -- demonstrates how to use various ProjectionSpec's to connect units according to various patterns.
- demos/std_wizards -- standard wizards constructed in Programs using taGui programmable gui -- good starting point for making your own custom wizards.
- See ACT-R page for more info about these demos, which are largely taken from the official ACT-R demos that go along with the original lisp version.
- demos/bp_xor -- classic XOR problem -- also includes separate testing and training programs, and a grid view display of the network's performance.
- demos/bp_softmax -- demonstrates use of SoftMax output layer activation -- now much easier and fully self-contained in version 8.0
demo/cs (Constraint Satisfaction)
- demos/cs_424 -- classic 4-2-4 auto-encoder using stochastic units, learning via the Boltzmann machine learning rule
- demos/cs_424 -- deterministic version of the above, learning using CHL (deterministic Boltzmann machine learning rule)
- demos/cs_ra -- random associator (deterministic) -- learns to associate random bit patterns and is a good test of general learning abilities.
demo/so Self Organizing
- demos/som -- self-organizing map model -- learns topographic representations of input data in a self-organizing manner
- demos/cl_lines -- competitive learning self-organizing model, applied to horizontal and vertical lines inputs
demo/virt_env Virtual Environment
- demos/ve_arm -- A simple simulated robot that can move a 2 jointed arm to reach a target, all under control of the network.
- demos/cereb_arm -- Model of how the cerebellum learns to control a biophysically-realistic arm with 12 muscles, based on internally-generated error signals.
- demos/vis_test -- Allows testing of V1 visual filters on well-controlled visual stimuli rendered using a canvas and in a 3D world.
- demos/vis_test_shadow -- Visual test on 3D world with shadows enabled
- demos/vis_color_test -- Test of color processing using opponent-coding color channels in the DoG (difference of Gaussian) filter system
- demos/vis_motion_test -- Test of visual motion filters using stimuli that move in reliable ways (also demonstrates drawing of SVG images on a taCanvas)
- demos/sound_object -- Test of loading and processing sound files using filters based on the auditory pathways of the brain
- demos/vocal_tract -- Generates sound output based on a simulation of the human vocal tract configuration, and can produce reasonable speech sounds
Computational Explorations in Cognitive Neuroscience Projects
- https://grey.colorado.edu/CompCogNeuro/index.php/CCNBook/Sims/All -- large collection of well-documented projects built in emergent, exploring a wide range of cognitive phenomena.
Back to Using emergent