Merge remote-tracking branch 'upstream/master'

.coveragerc

@@ -3,4 +3,8 @@ exclude_lines =
     # Have to re-enable the standard pragma
     pragma: no cover
     @abstract
-    NotImplementedError
+    NotImplementedError
+
+[run]
+omit =
+     *__init__.py

CONTRIBUTING.md

@@ -11,7 +11,7 @@ being contribution of code or documentation to the project. Improving the
 documentation is no less important than improving the library itself. If you
 find a typo in the documentation, or have made improvements, do not hesitate to
 submit a GitHub pull request. Documentation can be found under the
-[doc/](https://github.com/neurodata/progressive-learning/tree/master/docs) directory.
+[docs/](https://github.com/neurodata/progressive-learning/tree/master/docs) directory.
 
 But there are many other ways to help. In particular answering queries on the
 [issue tracker](https://github.com/neurodata/progressive-learning/issues), and
@@ -27,4 +27,4 @@ Code of Conduct
 ---------------
 
 We abide by the principles of openness, respect, and consideration of others
-of the Python Software Foundation: https://www.python.org/psf/codeofconduct/.
+of the Python Software Foundation: https://www.python.org/psf/codeofconduct/.

README.md

@@ -60,7 +60,7 @@ python3 setup.py install
 ```
 
 # Contributing
-We welcome contributions from anyone. Please see our [contribution guidelines](http://docs.neurodata.io/progressive-learning/) before making a pull request. Our 
+We welcome contributions from anyone. Please see our [contribution guidelines](https://github.com/neurodata/progressive-learning/blob/master/CONTRIBUTING.md) before making a pull request. Our 
 [issues](https://github.com/neurodata/progressive-learning/issues) page is full of places we could use help! 
 If you have an idea for an improvement not listed there, please 
 [make an issue](https://github.com/neurodata/progressive-learning/issues/new) first so you can discuss with the 

docs/contributing.rst

@@ -46,7 +46,7 @@ good feedback:
 - If an exception is raised, please **provide the full traceback**.
 
 - Please include your **operating system type and version number**, as well as
-  your **Python and hyppo versions**. This information
+  your **Python and ProgL versions**. This information
   can be found by running the following code snippet::
 
     import platform; print(platform.platform())
@@ -61,7 +61,7 @@ good feedback:
 Contributing Code
 -----------------
 
-The preferred workflow for contributing to `hyppo` is to fork the main
+The preferred workflow for contributing to `ProgL` is to fork the main
 repository on GitHub, clone, and develop on a branch. Steps:
 
 1. Fork the `project repository <https://github.com/neurodata/progressive-learning>`__ by clicking
@@ -70,12 +70,12 @@ repository on GitHub, clone, and develop on a branch. Steps:
    fork a repository see `this
    guide <https://help.github.com/articles/fork-a-repo/>`__.
 
-2. Clone your fork of the ``hyppo`` repo from your GitHub account to your
+2. Clone your fork of the ``ProgL`` repo from your GitHub account to your
    local disk:
 
    .. code:: bash
 
-      $ git clone [email protected]:YourLogin/hyppo.git
+      $ git clone [email protected]:YourLogin/progressive-learning.git
       $ cd progressive-learning
 
 3. Create a ``feature`` branch to hold your development changes:
@@ -150,7 +150,7 @@ before you submit a pull request:
 Coding Guidelines
 -----------------
 
-Uniformly formatted code makes it easier to share code ownership. ``hyppo``
+Uniformly formatted code makes it easier to share code ownership. ``ProgL``
 package closely follows the official Python guidelines detailed in
 `PEP8 <https://www.python.org/dev/peps/pep-0008/>`__ that detail how
 code should be formatted and indented. Please read it and follow it.
@@ -164,4 +164,4 @@ guidelines. Please read and follow the
 `numpydoc <https://numpydoc.readthedocs.io/en/latest/format.html#overview>`__
 guidelines. Refer to the
 `example.py <https://numpydoc.readthedocs.io/en/latest/example.html#example>`__
-provided by numpydoc.
+provided by numpydoc.

experiments/parity_experiment/generate_paper_plot.py

@@ -0,0 +1,270 @@
+#%%
+import random
+import matplotlib.pyplot as plt
+import tensorflow as tf
+import tensorflow.keras as keras
+import seaborn as sns 
+import matplotlib
+import numpy as np
+import pickle
+from proglearn.sims import generate_gaussian_parity
+
+from sklearn.model_selection import StratifiedKFold
+from math import log2, ceil 
+
+
+#%%
+def unpickle(file):
+    with open(file, 'rb') as fo:
+        dict = pickle.load(fo, encoding='bytes')
+    return dict
+
+def get_colors(colors, inds):
+    c = [colors[i] for i in inds]
+    return c
+
+#%%#%% Plotting the result
+#mc_rep = 50
+fontsize=30
+labelsize=28
+
+
+fig = plt.figure(constrained_layout=True,figsize=(21,30))
+gs = fig.add_gridspec(30, 21)
+
+colors = sns.color_palette('Dark2', n_colors=2)
+
+X, Y = generate_gaussian_parity(750)
+Z, W = generate_gaussian_parity(750, angle_params=np.pi/2)
+P, Q = generate_gaussian_parity(750, angle_params=np.pi/4)
+
+ax = fig.add_subplot(gs[:6,:6])
+ax.scatter(X[:, 0], X[:, 1], c=get_colors(colors, Y), s=50)
+
+ax.set_xticks([])
+ax.set_yticks([])
+ax.set_title('Gaussian XOR', fontsize=30)
+
+plt.tight_layout()
+ax.axis('off')
+#plt.savefig('./result/figs/gaussian-xor.pdf')
+
+#####################
+ax = fig.add_subplot(gs[:6,7:13])
+ax.scatter(Z[:, 0], Z[:, 1], c=get_colors(colors, W), s=50)
+
+ax.set_xticks([])
+ax.set_yticks([])
+ax.set_title('Gaussian N-XOR', fontsize=30)
+ax.axis('off')
+
+#####################
+ax = fig.add_subplot(gs[:6,14:20])
+ax.scatter(P[:, 0], P[:, 1], c=get_colors(colors, Q), s=50)
+
+ax.set_xticks([])
+ax.set_yticks([])
+ax.set_title('Gaussian R-XOR', fontsize=30)
+ax.axis('off')
+
+######################
+mean_error = unpickle('plots/mean_xor_nxor.pickle')
+
+n_xor = (100*np.arange(0.5, 7.25, step=0.25)).astype(int)
+n_nxor = (100*np.arange(0.5, 7.50, step=0.25)).astype(int)
+
+n1s = n_xor
+n2s = n_nxor
+
+ns = np.concatenate((n1s, n2s + n1s[-1]))
+ls=['-', '--']
+algorithms = ['XOR Forest', 'N-XOR Forest', 'Lifelong Forest', 'Naive Forest']
+
+
+TASK1='XOR'
+TASK2='N-XOR'
+
+fontsize=30
+labelsize=28
+
+colors = sns.color_palette("Set1", n_colors = 2)
+
+ax1 = fig.add_subplot(gs[7:13,2:9])
+# for i, algo in enumerate(algorithms):
+ax1.plot(n1s, mean_error[0,:len(n1s)], label=algorithms[0], c=colors[1], ls=ls[np.sum(0 > 1).astype(int)], lw=3)
+ax1.plot(ns[len(n1s):], mean_error[2, len(n1s):], label=algorithms[1], c=colors[1], ls=ls[1], lw=3)
+
+ax1.plot(ns, mean_error[1], label=algorithms[2], c=colors[0], ls=ls[np.sum(1 > 1).astype(int)], lw=3)
+
+ax1.plot(ns, mean_error[4], label=algorithms[3], c='g', ls=ls[np.sum(1 > 1).astype(int)], lw=3)
+
+ax1.set_ylabel('Generalization Error (%s)'%(TASK1), fontsize=fontsize)
+ax1.legend(loc='upper left', fontsize=20, frameon=False)
+#ax1.set_ylim(0.09, 0.21)
+ax1.set_xlabel('Total Sample Size', fontsize=fontsize)
+ax1.tick_params(labelsize=labelsize)
+#ax1.set_yticks([0.1,0.15, 0.2])
+ax1.set_xticks([250,750,1500])
+#ax1.axvline(x=750, c='gray', linewidth=1.5, linestyle="dashed")
+ax1.set_title('XOR', fontsize=30)
+
+right_side = ax1.spines["right"]
+right_side.set_visible(False)
+top_side = ax1.spines["top"]
+top_side.set_visible(False)
+
+ax1.text(400, np.mean(ax1.get_ylim()), "%s"%(TASK1), fontsize=26)
+ax1.text(900, np.mean(ax1.get_ylim()), "%s"%(TASK2), fontsize=26)
+
+#######################
+mean_error = unpickle('plots/mean_xor_nxor.pickle')
+
+algorithms = ['XOR Forest', 'N-XOR Forest', 'Lifelong Forest', 'Naive Forest']
+
+TASK1='XOR'
+TASK2='N-XOR'
+
+ax1 = fig.add_subplot(gs[7:13,12:19])
+ax1.plot(n1s, mean_error[0,:len(n1s)], label=algorithms[0], c=colors[1], ls=ls[np.sum(0 > 1).astype(int)], lw=3)
+ax1.plot(ns[len(n1s):], mean_error[2, len(n1s):], label=algorithms[1], c=colors[1], ls=ls[1], lw=3)
+
+ax1.plot(ns[len(n1s):], mean_error[3, len(n1s):], label=algorithms[2], c=colors[0], ls=ls[1], lw=3)
+ax1.plot(ns[len(n1s):], mean_error[5, len(n1s):], label=algorithms[3], c='g', ls=ls[1], lw=3)
+
+ax1.set_ylabel('Generalization Error (%s)'%(TASK2), fontsize=fontsize)
+#ax1.legend(loc='upper left', fontsize=18, frameon=False)
+#         ax1.set_ylim(-0.01, 0.22)
+ax1.set_xlabel('Total Sample Size', fontsize=fontsize)
+ax1.tick_params(labelsize=labelsize)
+# ax1.set_yticks([0.15, 0.25, 0.35])
+#ax1.set_yticks([0.15, 0.2])
+ax1.set_xticks([250,750,1500])
+#ax1.axvline(x=750, c='gray', linewidth=1.5, linestyle="dashed")
+
+#ax1.set_ylim(0.11, 0.21)
+
+right_side = ax1.spines["right"]
+right_side.set_visible(False)
+top_side = ax1.spines["top"]
+top_side.set_visible(False)
+
+# ax1.set_ylim(0.14, 0.36)
+ax1.text(400, np.mean(ax1.get_ylim()), "%s"%(TASK1), fontsize=26)
+ax1.text(900, np.mean(ax1.get_ylim()), "%s"%(TASK2), fontsize=26)
+
+ax1.set_title('N-XOR', fontsize=30)
+
+##################
+mean_te = unpickle('plots/mean_te_xor_nxor.pickle')
+algorithms = ['Lifelong BTE', 'Lifelong FTE', 'Naive BTE', 'Naive FTE']
+
+TASK1='XOR'
+TASK2='N-XOR'
+
+ax1 = fig.add_subplot(gs[15:21,2:9])
+
+ax1.plot(ns, mean_te[0], label=algorithms[0], c=colors[0], ls=ls[0], lw=3)
+
+ax1.plot(ns[len(n1s):], mean_te[1, len(n1s):], label=algorithms[1], c=colors[0], ls=ls[1], lw=3)
+
+ax1.plot(ns, mean_te[2], label=algorithms[2], c='g', ls=ls[0], lw=3)
+ax1.plot(ns[len(n1s):], mean_te[3, len(n1s):], label=algorithms[3], c='g', ls=ls[1], lw=3)
+
+ax1.set_ylabel('Forward/Backward \n Transfer Efficiency (FTE/BTE)', fontsize=fontsize)
+ax1.legend(loc='upper left', fontsize=20, frameon=False)
+#ax1.set_ylim(.99, 1.4)
+ax1.set_xlabel('Total Sample Size', fontsize=fontsize)
+ax1.tick_params(labelsize=labelsize)
+ax1.set_yticks([0,.5,1,1.5])
+ax1.set_xticks([250,750,1500])
+#ax1.axvline(x=750, c='gray', linewidth=1.5, linestyle="dashed")
+right_side = ax1.spines["right"]
+right_side.set_visible(False)
+top_side = ax1.spines["top"]
+top_side.set_visible(False)
+ax1.hlines(1, 50,1500, colors='gray', linestyles='dashed',linewidth=1.5)
+
+ax1.text(400, np.mean(ax1.get_ylim()), "%s"%(TASK1), fontsize=26)
+ax1.text(900, np.mean(ax1.get_ylim()), "%s"%(TASK2), fontsize=26)
+
+######################
+mean_te = unpickle('plots/mean_te_xor_rxor.pickle')
+algorithms = ['Lifelong BTE', 'Lifelong FTE', 'Naive BTE', 'Naive FTE']
+
+TASK1='XOR'
+TASK2='R-XOR'
+
+ax1 = fig.add_subplot(gs[15:21,12:19])
+
+ax1.plot(ns, mean_te[0], label=algorithms[0], c=colors[0], ls=ls[0], lw=3)
+
+ax1.plot(ns[len(n1s):], mean_te[1, len(n1s):], label=algorithms[1], c=colors[0], ls=ls[1], lw=3)
+
+ax1.plot(ns, mean_te[2], label=algorithms[2], c='g', ls=ls[0], lw=3)
+ax1.plot(ns[len(n1s):], mean_te[3, len(n1s):], label=algorithms[3], c='g', ls=ls[1], lw=3)
+
+ax1.set_ylabel('Forward/Backward \n Transfer Efficiency (FTE/BTE)', fontsize=fontsize)
+#ax1.legend(loc='upper left', fontsize=20, frameon=False)
+#ax1.set_ylim(.99, 1.4)
+ax1.set_xlabel('Total Sample Size', fontsize=fontsize)
+ax1.tick_params(labelsize=labelsize)
+ax1.set_yticks([0,.5,1])
+ax1.set_xticks([250,750,1500])
+#ax1.axvline(x=750, c='gray', linewidth=1.5, linestyle="dashed")
+right_side = ax1.spines["right"]
+right_side.set_visible(False)
+top_side = ax1.spines["top"]
+top_side.set_visible(False)
+ax1.hlines(1, 50,1500, colors='gray', linestyles='dashed',linewidth=1.5)
+
+ax1.text(400, np.mean(ax1.get_ylim()), "%s"%(TASK1), fontsize=26)
+ax1.text(900, np.mean(ax1.get_ylim()), "%s"%(TASK2), fontsize=26)
+
+########################################################
+ax = fig.add_subplot(gs[23:29,2:9])
+with open('plots/mean_angle_te.pickle','rb') as f:
+    te = pickle.load(f)
+angle_sweep = range(0,90,1)
+
+ax.plot(angle_sweep,te,c='r',linewidth = 3)
+ax.set_xticks(range(0,91,15))
+ax.tick_params(labelsize=labelsize)
+ax.set_xlabel('Angle of Rotation (Degrees)', fontsize=fontsize)
+ax.set_ylabel('Backward Transfer Efficiency (XOR)', fontsize=fontsize)
+#ax.set_title("XOR vs. Rotated-XOR", fontsize = fontsize)
+ax.hlines(1,0,90, colors='grey', linestyles='dashed',linewidth=1.5)
+
+right_side = ax.spines["right"]
+right_side.set_visible(False)
+top_side = ax.spines["top"]
+top_side.set_visible(False)
+
+#####################################
+ax = fig.add_subplot(gs[23:29,12:19])
+
+with open('plots/mean_sample_te.pickle','rb') as f:
+    te = pickle.load(f)
+task2_sample_sweep = (2**np.arange(np.log2(60), np.log2(5010)+1, .25)).astype('int')
+
+ax.plot(task2_sample_sweep,te,c='r',linewidth = 3)
+ax.hlines(1, 60,5200, colors='gray', linestyles='dashed',linewidth=1.5)
+ax.set_xscale('log')
+ax.set_xticks([])
+ax.set_yticks([0.98,1,1.02,1.04])
+ax.tick_params(labelsize=26)
+ax.get_xaxis().set_major_formatter(matplotlib.ticker.ScalarFormatter())
+ax.text(50, np.mean(ax.get_ylim())-.042, "50", fontsize=labelsize)
+ax.text(500, np.mean(ax.get_ylim())-.042, "500", fontsize=labelsize)
+ax.text(5000, np.mean(ax.get_ylim())-.042, "5000", fontsize=labelsize)
+
+ax.text(50, np.mean(ax.get_ylim())-.047, "Number of $25^\circ$-RXOR Training Samples", fontsize=fontsize-4)
+ax.set_ylabel('Backward Transfer Efficiency (XOR)',fontsize=24)
+
+right_side = ax.spines["right"]
+right_side.set_visible(False)
+top_side = ax.spines["top"]
+top_side.set_visible(False)
+
+
+plt.savefig('./plots/parity_exp.pdf')
+# %%