1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
|
{
"nbformat_minor": 0,
"nbformat": 4,
"cells": [
{
"execution_count": null,
"cell_type": "code",
"source": [
"%matplotlib inline"
],
"outputs": [],
"metadata": {
"collapsed": false
}
},
{
"source": [
"\n# OT for domain adaptation\n\n\n\n"
],
"cell_type": "markdown",
"metadata": {}
},
{
"execution_count": null,
"cell_type": "code",
"source": [
"import matplotlib.pylab as pl\nimport ot\n\n\n\n\n#%% parameters\n\nn=150 # nb samples in source and target datasets\n\nxs,ys=ot.datasets.get_data_classif('3gauss',n)\nxt,yt=ot.datasets.get_data_classif('3gauss2',n)\n\n\n\n\n#%% plot samples\n\npl.figure(1)\n\npl.subplot(2,2,1)\npl.scatter(xs[:,0],xs[:,1],c=ys,marker='+',label='Source samples')\npl.legend(loc=0)\npl.title('Source distributions')\n\npl.subplot(2,2,2)\npl.scatter(xt[:,0],xt[:,1],c=yt,marker='o',label='Target samples')\npl.legend(loc=0)\npl.title('target distributions')\n\n\n#%% OT estimation\n\n# LP problem\nda_emd=ot.da.OTDA() # init class\nda_emd.fit(xs,xt) # fit distributions\nxst0=da_emd.interp() # interpolation of source samples\n\n\n# sinkhorn regularization\nlambd=1e-1\nda_entrop=ot.da.OTDA_sinkhorn()\nda_entrop.fit(xs,xt,reg=lambd)\nxsts=da_entrop.interp()\n\n# non-convex Group lasso regularization\nreg=1e-1\neta=1e0\nda_lpl1=ot.da.OTDA_lpl1()\nda_lpl1.fit(xs,ys,xt,reg=reg,eta=eta)\nxstg=da_lpl1.interp()\n\n\n# True Group lasso regularization\nreg=1e-1\neta=2e0\nda_l1l2=ot.da.OTDA_l1l2()\nda_l1l2.fit(xs,ys,xt,reg=reg,eta=eta,numItermax=20,verbose=True)\nxstgl=da_l1l2.interp()\n\n\n#%% plot interpolated source samples\npl.figure(4,(15,8))\n\nparam_img={'interpolation':'nearest','cmap':'jet'}\n\npl.subplot(2,4,1)\npl.imshow(da_emd.G,**param_img)\npl.title('OT matrix')\n\n\npl.subplot(2,4,2)\npl.imshow(da_entrop.G,**param_img)\npl.title('OT matrix sinkhorn')\n\npl.subplot(2,4,3)\npl.imshow(da_lpl1.G,**param_img)\npl.title('OT matrix non-convex Group Lasso')\n\npl.subplot(2,4,4)\npl.imshow(da_l1l2.G,**param_img)\npl.title('OT matrix Group Lasso')\n\n\npl.subplot(2,4,5)\npl.scatter(xt[:,0],xt[:,1],c=yt,marker='o',label='Target samples',alpha=0.3)\npl.scatter(xst0[:,0],xst0[:,1],c=ys,marker='+',label='Transp samples',s=30)\npl.title('Interp samples')\npl.legend(loc=0)\n\npl.subplot(2,4,6)\npl.scatter(xt[:,0],xt[:,1],c=yt,marker='o',label='Target samples',alpha=0.3)\npl.scatter(xsts[:,0],xsts[:,1],c=ys,marker='+',label='Transp samples',s=30)\npl.title('Interp samples Sinkhorn')\n\npl.subplot(2,4,7)\npl.scatter(xt[:,0],xt[:,1],c=yt,marker='o',label='Target samples',alpha=0.3)\npl.scatter(xstg[:,0],xstg[:,1],c=ys,marker='+',label='Transp samples',s=30)\npl.title('Interp samples non-convex Group Lasso')\n\npl.subplot(2,4,8)\npl.scatter(xt[:,0],xt[:,1],c=yt,marker='o',label='Target samples',alpha=0.3)\npl.scatter(xstgl[:,0],xstgl[:,1],c=ys,marker='+',label='Transp samples',s=30)\npl.title('Interp samples Group Lasso')"
],
"outputs": [],
"metadata": {
"collapsed": false
}
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 2",
"name": "python2",
"language": "python"
},
"language_info": {
"mimetype": "text/x-python",
"nbconvert_exporter": "python",
"name": "python",
"file_extension": ".py",
"version": "2.7.12",
"pygments_lexer": "ipython2",
"codemirror_mode": {
"version": 2,
"name": "ipython"
}
}
}
}
|
