[MRG] Add factored coupling (#358)

* add gfactored ot

* pep8 and add doc

* add exmaple for factotred OT

* final number of PR

* correct test on backends

* remove useless loss

* better tests

Author: rflamary

README.md

@@ -305,4 +305,6 @@ Conference on Machine Learning, PMLR 119:4692-4701, 2020
 [38] C. Vincent-Cuaz, T. Vayer, R. Flamary, M. Corneli, N. Courty, [Online Graph 
 Dictionary Learning](https://arxiv.org/pdf/2102.06555.pdf), International Conference on Machine Learning (ICML), 2021.
 
-[39] Gozlan, N., Roberto, C., Samson, P. M., & Tetali, P. (2017). [Kantorovich duality for general transport costs and applications](https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.712.1825&rep=rep1&type=pdf). Journal of Functional Analysis, 273(11), 3327-3405.
+[39] Gozlan, N., Roberto, C., Samson, P. M., & Tetali, P. (2017). [Kantorovich duality for general transport costs and applications](https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.712.1825&rep=rep1&type=pdf). Journal of Functional Analysis, 273(11), 3327-3405.
+
+[40] Forrow, A., Hütter, J. C., Nitzan, M., Rigollet, P., Schiebinger, G., & Weed, J. (2019, April). [Statistical optimal transport via factored couplings](http://proceedings.mlr.press/v89/forrow19a/forrow19a.pdf). In The 22nd International Conference on Artificial Intelligence and Statistics (pp. 2454-2465). PMLR.

RELEASES.md

@@ -5,6 +5,7 @@
 
 #### New features
 
+- Implementation of factored OT with emd and sinkhorn (PR #358).
 - A brand new logo for POT (PR #357)
 - Better list of related examples in quick start guide with `minigallery` (PR #334).
 - Add optional log-domain Sinkhorn implementation in WDA to support smaller values

docs/source/all.rst

@@ -29,6 +29,7 @@ API and modules
    partial
    sliced
    weak
+   factored
 
 .. autosummary::
    :toctree: ../modules/generated/

examples/others/plot_factored_coupling.py

@@ -0,0 +1,86 @@
+# -*- coding: utf-8 -*-
+"""
+==========================================
+Optimal transport with factored couplings
+==========================================
+
+Illustration of the factored coupling OT between 2D empirical distributions
+
+"""
+
+# Author: Remi Flamary <remi.flamary@polytechnique.edu>
+#
+# License: MIT License
+
+# sphinx_gallery_thumbnail_number = 2
+
+import numpy as np
+import matplotlib.pylab as pl
+import ot
+import ot.plot
+
+# %%
+# Generate data an plot it
+# ------------------------
+
+# parameters and data generation
+
+np.random.seed(42)
+
+n = 100  # nb samples
+
+xs = np.random.rand(n, 2) - .5
+
+xs = xs + np.sign(xs)
+
+xt = np.random.rand(n, 2) - .5
+
+a, b = ot.unif(n), ot.unif(n)  # uniform distribution on samples
+
+#%% plot samples
+
+pl.figure(1)
+pl.plot(xs[:, 0], xs[:, 1], '+b', label='Source samples')
+pl.plot(xt[:, 0], xt[:, 1], 'xr', label='Target samples')
+pl.legend(loc=0)
+pl.title('Source and target distributions')
+
+
+# %%
+# Compute Factore OT and exact OT solutions
+# --------------------------------------
+
+#%% EMD
+M = ot.dist(xs, xt)
+G0 = ot.emd(a, b, M)
+
+#%% factored OT OT
+
+Ga, Gb, xb = ot.factored_optimal_transport(xs, xt, a, b, r=4)
+
+
+# %%
+# Plot factored OT and exact OT solutions
+# --------------------------------------
+
+pl.figure(2, (14, 4))
+
+pl.subplot(1, 3, 1)
+ot.plot.plot2D_samples_mat(xs, xt, G0, c=[.2, .2, .2], alpha=0.1)
+pl.plot(xs[:, 0], xs[:, 1], '+b', label='Source samples')
+pl.plot(xt[:, 0], xt[:, 1], 'xr', label='Target samples')
+pl.title('Exact OT with samples')
+
+pl.subplot(1, 3, 2)
+ot.plot.plot2D_samples_mat(xs, xb, Ga, c=[.6, .6, .9], alpha=0.5)
+ot.plot.plot2D_samples_mat(xb, xt, Gb, c=[.9, .6, .6], alpha=0.5)
+pl.plot(xs[:, 0], xs[:, 1], '+b', label='Source samples')
+pl.plot(xt[:, 0], xt[:, 1], 'xr', label='Target samples')
+pl.plot(xb[:, 0], xb[:, 1], 'og', label='Template samples')
+pl.title('Factored OT with template samples')
+
+pl.subplot(1, 3, 3)
+ot.plot.plot2D_samples_mat(xs, xt, Ga.dot(Gb), c=[.2, .2, .2], alpha=0.1)
+pl.plot(xs[:, 0], xs[:, 1], '+b', label='Source samples')
+pl.plot(xt[:, 0], xt[:, 1], 'xr', label='Target samples')
+pl.title('Factored OT low rank OT plan')

ot/__init__.py

@@ -33,6 +33,7 @@
 from . import backend
 from . import regpath
 from . import weak
+from . import factored
 
 # OT functions
 from .lp import emd, emd2, emd_1d, emd2_1d, wasserstein_1d
@@ -44,6 +45,9 @@
 from .gromov import (gromov_wasserstein, gromov_wasserstein2,
                         gromov_barycenters, fused_gromov_wasserstein, fused_gromov_wasserstein2)
 from .weak import weak_optimal_transport
+from .factored import factored_optimal_transport
+
+
 # utils functions
 from .utils import dist, unif, tic, toc, toq
 
@@ -57,4 +61,5 @@
            'sinkhorn_unbalanced2', 'sliced_wasserstein_distance',
            'gromov_wasserstein', 'gromov_wasserstein2', 'gromov_barycenters', 'fused_gromov_wasserstein', 'fused_gromov_wasserstein2',
             'max_sliced_wasserstein_distance', 'weak_optimal_transport',
+            'factored_optimal_transport', 
            'smooth', 'stochastic', 'unbalanced', 'partial', 'regpath']

ot/factored.py

@@ -0,0 +1,145 @@
+"""
+Factored OT solvers (low rank, cost or OT plan)
+"""
+
+# Author: Remi Flamary <remi.flamary@polytehnique.edu>
+#
+# License: MIT License
+
+from .backend import get_backend
+from .utils import dist
+from .lp import emd
+from .bregman import sinkhorn
+
+__all__ = ['factored_optimal_transport']
+
+
+def factored_optimal_transport(Xa, Xb, a=None, b=None, reg=0.0, r=100, X0=None, stopThr=1e-7, numItermax=100, verbose=False, log=False, **kwargs):
+    r"""Solves factored OT problem and return OT plans and intermediate distribution
+
+    This function solve the following OT problem [40]_
+
+    .. math::
+        \mathop{\arg \min}_\mu \quad  W_2^2(\mu_a,\mu)+ W_2^2(\mu,\mu_b)
+
+    where :
+
+    - :math:`\mu_a` and :math:`\mu_b`  are empirical distributions.
+    - :math:`\mu` is an empirical distribution with r samples
+
+    And returns the two OT plans between
+
+    .. note:: This function is backend-compatible and will work on arrays
+        from all compatible backends. But the algorithm uses the C++ CPU backend
+        which can lead to copy overhead on GPU arrays.
+
+    Uses the conditional gradient algorithm to solve the problem proposed in
+    :ref:`[39] <references-weak>`.
+
+    Parameters
+    ----------
+    Xa : (ns,d) array-like, float
+        Source samples
+    Xb : (nt,d) array-like, float
+        Target samples
+    a : (ns,) array-like, float
+        Source histogram (uniform weight if empty list)
+    b : (nt,) array-like, float
+        Target histogram (uniform weight if empty list))
+    numItermax : int, optional
+        Max number of iterations
+    stopThr : float, optional
+        Stop threshold on the relative variation (>0)
+    verbose : bool, optional
+        Print information along iterations
+    log : bool, optional
+        record log if True
+
+
+    Returns
+    -------
+    Ga: array-like, shape (ns, r)
+        Optimal transportation matrix between source and the intermediate
+        distribution
+    Gb: array-like, shape (r, nt)
+        Optimal transportation matrix between the intermediate and target
+        distribution
+    X: array-like, shape (r, d)
+        Support of the intermediate distribution
+    log: dict, optional
+        If input log is true, a dictionary containing the cost and dual
+        variables and exit status
+
+
+    .. _references-factored:
+    References
+    ----------
+    .. [40] Forrow, A., Hütter, J. C., Nitzan, M., Rigollet, P., Schiebinger,
+        G., & Weed, J. (2019, April). Statistical optimal transport via factored
+        couplings. In The 22nd International Conference on Artificial
+        Intelligence and Statistics (pp. 2454-2465). PMLR.
+
+    See Also
+    --------
+    ot.bregman.sinkhorn : Entropic regularized OT ot.optim.cg : General
+    regularized OT
+    """
+
+    nx = get_backend(Xa, Xb)
+
+    n_a = Xa.shape[0]
+    n_b = Xb.shape[0]
+    d = Xa.shape[1]
+
+    if a is None:
+        a = nx.ones((n_a), type_as=Xa) / n_a
+    if b is None:
+        b = nx.ones((n_b), type_as=Xb) / n_b
+
+    if X0 is None:
+        X = nx.randn(r, d, type_as=Xa)
+    else:
+        X = X0
+
+    w = nx.ones(r, type_as=Xa) / r
+
+    def solve_ot(X1, X2, w1, w2):
+        M = dist(X1, X2)
+        if reg > 0:
+            G, log = sinkhorn(w1, w2, M, reg, log=True, **kwargs)
+            log['cost'] = nx.sum(G * M)
+            return G, log
+        else:
+            return emd(w1, w2, M, log=True, **kwargs)
+
+    norm_delta = []
+
+    # solve the barycenter
+    for i in range(numItermax):
+
+        old_X = X
+
+        # solve OT with template
+        Ga, loga = solve_ot(Xa, X, a, w)
+        Gb, logb = solve_ot(X, Xb, w, b)
+
+        X = 0.5 * (nx.dot(Ga.T, Xa) + nx.dot(Gb, Xb)) * r
+
+        delta = nx.norm(X - old_X)
+        if delta < stopThr:
+            break
+        if log:
+            norm_delta.append(delta)
+
+    if log:
+        log_dic = {'delta_iter': norm_delta,
+                   'ua': loga['u'],
+                   'va': loga['v'],
+                   'ub': logb['u'],
+                   'vb': logb['v'],
+                   'costa': loga['cost'],
+                   'costb': logb['cost'],
+                   }
+        return Ga, Gb, X, log_dic
+
+    return Ga, Gb, X

ot/plot.py

@@ -85,8 +85,13 @@ def plot2D_samples_mat(xs, xt, G, thr=1e-8, **kwargs):
     if ('color' not in kwargs) and ('c' not in kwargs):
         kwargs['color'] = 'k'
     mx = G.max()
+    if 'alpha' in kwargs:
+        scale = kwargs['alpha']
+        del kwargs['alpha']
+    else:
+        scale = 1
     for i in range(xs.shape[0]):
         for j in range(xt.shape[0]):
             if G[i, j] / mx > thr:
                 pl.plot([xs[i, 0], xt[j, 0]], [xs[i, 1], xt[j, 1]],
-                        alpha=G[i, j] / mx, **kwargs)
+                        alpha=G[i, j] / mx * scale, **kwargs)

test/test_factored.py

@@ -0,0 +1,56 @@
+"""Tests for main module ot.weak """
+
+# Author: Remi Flamary <remi.flamary@unice.fr>
+#
+# License: MIT License
+
+import ot
+import numpy as np
+
+
+def test_factored_ot():
+    # test weak ot solver and identity stationary point
+    n = 50
+    rng = np.random.RandomState(0)
+
+    xs = rng.randn(n, 2)
+    xt = rng.randn(n, 2)
+    u = ot.utils.unif(n)
+
+    Ga, Gb, X, log = ot.factored_optimal_transport(xs, xt, u, u, r=10, log=True)
+
+    # check constraints
+    np.testing.assert_allclose(u, Ga.sum(1))
+    np.testing.assert_allclose(u, Gb.sum(0))
+
+    Ga, Gb, X, log = ot.factored_optimal_transport(xs, xt, u, u, reg=1, r=10, log=True)
+
+    # check constraints
+    np.testing.assert_allclose(u, Ga.sum(1))
+    np.testing.assert_allclose(u, Gb.sum(0))
+
+
+def test_factored_ot_backends(nx):
+    # test weak ot solver for different backends
+    n = 50
+    rng = np.random.RandomState(0)
+
+    xs = rng.randn(n, 2)
+    xt = rng.randn(n, 2)
+    u = ot.utils.unif(n)
+
+    xs2 = nx.from_numpy(xs)
+    xt2 = nx.from_numpy(xt)
+    u2 = nx.from_numpy(u)
+
+    Ga2, Gb2, X2 = ot.factored_optimal_transport(xs2, xt2, u2, u2, r=10)
+
+    # check constraints
+    np.testing.assert_allclose(u, nx.to_numpy(Ga2).sum(1))
+    np.testing.assert_allclose(u, nx.to_numpy(Gb2).sum(0))
+
+    Ga2, Gb2, X2 = ot.factored_optimal_transport(xs2, xt2, reg=1, r=10, X0=X2)
+
+    # check constraints
+    np.testing.assert_allclose(u, nx.to_numpy(Ga2).sum(1))
+    np.testing.assert_allclose(u, nx.to_numpy(Gb2).sum(0))