add examples + tests + warning in entropic solvers + releases

Author: cedricvincentcuaz

RELEASES.md

@@ -1,5 +1,20 @@
 # Releases
 
+## 0.9.1dev
+*Sine April 2023*
+
+
+#### New features
+
+- Added Proximal Point algorithm to solve GW problems via a new parameter `solver="PPA"` in `ot.gromov.entropic_gromov_wasserstein` + examples (PR #455)
+- Added features `warmstart` and `kwargs` in `ot.gromov.entropic_gromov_wasserstein` to respectively perform warmstart on dual potentials and pass parameters to `ot.sinkhorn` (PR #455)
+- Added sinkhorn projection based solvers for FGW `ot.gromov.entropic_fused_gromov_wasserstein` and entropic FGW barycenters + examples (PR #455)
+- Added features `warmstartT` and `kwargs` to all CG and entropic (F)GW barycenter solvers (PR #455)
+- Added entropic semi-relaxed (Fused) Gromov-Wasserstein solvers in `ot.gromov` + examples (PR #455)
+
+#### Closed issues
+
+
 ## 0.9.0
 
 This new release contains so many new features and bug fixes since 0.8.2 that we

examples/gromov/plot_entropic_semirelaxed_fgw.py

@@ -113,7 +113,9 @@
 # ---------------------------------------------
 #
 # We color nodes of the graph on the right - then project its node colors
-# based on the optimal transport plan from the entropic srGW matching
+# based on the optimal transport plan from the entropic srGW matching.
+# We adjust the intensity of links across domains proportionaly to the mass
+# sent, adding a minimal intensity of 0.1 if mass sent is not zero.
 
 
 def draw_graph(G, C, nodes_color_part, Gweights=None,
@@ -187,11 +189,12 @@ def draw_transp_colored_srGW(G1, C1, G2, C2, part_G1,
     pos2 = draw_graph(G2, C2, nodes_color_part2, Gweights=p2, pos=pos2,
                       node_size=node_size, shiftx=shiftx, seed=seed_G2)
     for k1, v1 in pos1.items():
+        max_Tk1 = np.max(T[k1, :])
         for k2, v2 in pos2.items():
             if (T[k1, k2] > 0):
                 pl.plot([pos1[k1][0], pos2[k2][0]],
                         [pos1[k1][1], pos2[k2][1]],
-                        '-', lw=0.5, alpha=0.3,
+                        '-', lw=0.6, alpha=min(T[k1, k2] / max_Tk1 + 0.1, 1.),
                         color=nodes_color_part1[k1])
     return pos1, pos2
 

examples/gromov/plot_fgw.py

@@ -4,13 +4,13 @@
 Plot Fused-gromov-Wasserstein
 ==============================
 
-This example illustrates the computation of FGW for 1D measures [18].
+This example first illustrates the computation of FGW for 1D measures estimated
+using a Conditional Gradient solver [24].
 
-[18] Vayer Titouan, Chapel Laetitia, Flamary Rémi, Tavenard Romain
+[24] Vayer Titouan, Chapel Laetitia, Flamary Rémi, Tavenard Romain
 and Courty Nicolas
 "Optimal Transport for structured data with application on graphs"
 International Conference on Machine Learning (ICML). 2019.
-
 """
 
 # Author: Titouan Vayer <[email protected]>
@@ -24,11 +24,13 @@
 import ot
 from ot.gromov import gromov_wasserstein, fused_gromov_wasserstein
 
+
 ##############################################################################
 # Generate data
 # -------------
 
-#%% parameters
+# parameters
+
 # We create two 1D random measures
 n = 20  # number of points in the first distribution
 n2 = 30  # number of points in the second distribution
@@ -53,10 +55,9 @@
 # Plot data
 # ---------
 
-#%% plot the distributions
+# plot the distributions
 
-pl.close(10)
-pl.figure(10, (7, 7))
+pl.figure(1, (7, 7))
 
 pl.subplot(2, 1, 1)
 
@@ -78,7 +79,7 @@
 # Create structure matrices and across-feature distance matrix
 # ------------------------------------------------------------
 
-#%% Structure matrices and across-features distance matrix
+# Structure matrices and across-features distance matrix
 C1 = ot.dist(xs)
 C2 = ot.dist(xt)
 M = ot.dist(ys, yt)
@@ -90,10 +91,9 @@
 # Plot matrices
 # -------------
 
-#%%
 cmap = 'Reds'
-pl.close(10)
-pl.figure(10, (5, 5))
+
+pl.figure(2, (5, 5))
 fs = 15
 l_x = [0, 5, 10, 15]
 l_y = [0, 5, 10, 15, 20, 25]
@@ -113,7 +113,6 @@
 pl.imshow(C2, cmap=cmap, interpolation='nearest')
 pl.title("$C_2$", fontsize=fs)
 pl.ylabel("$l$", fontsize=fs)
-#pl.ylabel("$l$",fontsize=fs)
 pl.xticks(())
 pl.yticks(l_y)
 ax2.set_aspect('auto')
@@ -133,28 +132,27 @@
 # Compute FGW/GW
 # --------------
 
-#%% Computing FGW and GW
+# Computing FGW and GW
 alpha = 1e-3
 
 ot.tic()
 Gwg, logw = fused_gromov_wasserstein(M, C1, C2, p, q, loss_fun='square_loss', alpha=alpha, verbose=True, log=True)
 ot.toc()
 
-#%reload_ext WGW
+# reload_ext WGW
 Gg, log = gromov_wasserstein(C1, C2, p, q, loss_fun='square_loss', verbose=True, log=True)
 
 ##############################################################################
 # Visualize transport matrices
 # ----------------------------
 
-#%% visu OT matrix
+# visu OT matrix
 cmap = 'Blues'
 fs = 15
-pl.figure(2, (13, 5))
+pl.figure(3, (13, 5))
 pl.clf()
 pl.subplot(1, 3, 1)
 pl.imshow(Got, cmap=cmap, interpolation='nearest')
-#pl.xlabel("$y$",fontsize=fs)
 pl.ylabel("$i$", fontsize=fs)
 pl.xticks(())