A Parallel Mixture of SVMs for Very Large Scale

Problems
Ronan Collobert

Universit de Montral, DIRO

CP 6128, Succ. Centre-Ville
Montral, Qubec, Canada
[email protected]

Samy Bengio

IDIAP
CP 592, rue du Simplon 4
1920 Martigny, Switzerland
[email protected]

Yoshua Bengio

Universit de Montral, DIRO

CP 6128, Succ. Centre-Ville
Montral, Qubec, Canada
[email protected]

Abstract
Support Vector Machines (SVMs) are currently the state-of-the-art models for
many classication problems but they suer from the complexity of their training algorithm which is at least quadratic with respect to the number of examples.
Hence, it is hopeless to try to solve real-life problems having more than a few
hundreds of thousands examples with SVMs. The present paper proposes a
new mixture of SVMs that can be easily implemented in parallel and where
each SVM is trained on a small subset of the whole dataset. Experiments on a
large benchmark dataset (Forest) as well as a dicult speech database, yielded
signicant time improvement (time complexity appears empirically to locally
grow linearly with the number of examples). In addition, and that is a surprise,
a signicant improvement in generalization was observed on Forest.

1 Introduction
Recently a lot of work has been done around Support Vector Machines [9], mainly due to
their impressive generalization performances on classication problems when compared to other
algorithms such as articial neural networks [3, 6]. However, SVMs require to solve a quadratic
optimization problem which needs resources that are at least quadratic in the number of training
examples, and it is thus hopeless to try solving problems having millions of examples using
classical SVMs.
In order to overcome this drawback, we propose in this paper to use a mixture of several SVMs,
each of them trained only on a part of the dataset. The idea of an SVM mixture is not new,
although previous attempts such as Kwok's paper on Support Vector Mixtures [5] did not train
the SVMs on part of the dataset but on the whole dataset and hence could not overcome the

Part of this work has been done while Ronan Collobert was at IDIAP, CP 592, rue du Simplon 4,
1920 Martigny, Switzerland.

time complexity problem for large datasets. We propose here a simple method to train such
a mixture, and we will show that in practice this method is much faster than training only
one SVM, and leads to results that are at least as good as one SVM. We conjecture that the
training time complexity of the proposed approach with respect to the number of examples is
sub-quadratic for large data sets. Moreover this mixture can be easily parallelized, which could
improve again signicantly the training time.
The organization of the paper goes as follows: in the next section, we briey introduce the SVM
model for classication. In section 3 we present our mixture of SVMs, followed in section 4 by
some comparisons to related models. In section 5 we show some experimental results, rst on a
toy dataset, then on two large real-life datasets. A short conclusion then follows.

2 Introduction to Support Vector Machines

Support Vector Machines (SVMs) [9] have been applied to many classication problems, generally yielding good performance compared to other algorithms. The decision function is of the
form
N
!
X
y = sign
yi i K(x, xi ) + b
(1)
i=1

where x Rd is the d-dimensional input vector of a test example, y {1, 1} is a class label, xi
is the input vector for the ith training example, yi is its associated class label, N is the number
of training examples, K(x, xi ) is a positive denite kernel function, and = {1 , . . . , N } and
b are the parameters of the model. Training an SVM consists in nding that minimizes the
objective function
N N
N
X
1 XX
i j yi yj K(xi , xj )
(2)
Q() =
i +
2 i=1 j=1
i=1
subject to the constraints
N
X

i yi = 0

(3)

i=1

and

0 i C i.
The kernel K(x, xi ) can have dierent forms, such as the Radial Basis Function (RBF):

kxi xj k2
K(xi , xj ) = exp
2
with parameter .

(4)
(5)

Therefore, to train an SVM, we need to solve a quadratic optimization problem, where the
number of parameters is N . This makes the use of SVMs for large datasets dicult: computing
K(xi , xj ) for every training pair would require O(N 2 ) computation, and solving may take up
to O(N 3 ). Note however that current state-of-the-art algorithms appear to have training time
complexity scaling much closer to O(N 2 ) than O(N 3 ) [2].

3 A New Conditional Mixture of SVMs

In this section we introduce a new type of mixture of SVMs. The output of the mixture for an
input vector x is computed as follows:
M
!
X
f (x) = h
wm (x)sm (x)
(6)
m=1

where M is the number of experts in the mixture, sm (x) is the output of the mth expert
given input x, wm (x) is the weight for the mth expert given by a gater module taking also
x in input, and h is a transfer function which could be for example the hyperbolic tangent for
classication tasks. Here each expert is an SVM, and we took a neural network for the gater in
our experiments. In the proposed model, the gater is trained to minimize the cost function

C=

N
X

[f (xi ) yi ] .

(7)

i=1

To train this model, we propose a very simple algorithm:

1. Divide the training set into M random subsets of size near N/M .
2. Train each expert separately over one of these subsets.
3. Keeping the experts xed, train the gater to minimize (7) on the whole training set.
4. Reconstruct M subsets: for each example (xi , yi ),

sort the experts in descending order according to the values wm (xi ),

assign the example to the rst expert in the list which has less than (N/M + c)
examples , in order to ensure a balance between the experts.
5. If a termination criterion is not fullled (such as a given number of iterations or a
validation error going up), goto step 2.
Note that step 2 of this algorithm can be easily implemented in parallel as each expert can
be trained separately on a dierent computer. Note also that step 3 can be an approximate
minimization (as usually done when training neural networks).

4 Other Mixtures of SVMs

The idea of mixture models is quite old and has given rise to very popular algorithms, such
as the well-known Mixture of Experts [4] where the cost function is similar to equation (7) but
where the gater and the experts are trained, using gradient descent or EM, on the whole dataset
(and not subsets) and their parameters are trained simultaneously. Hence such an algorithm
is quite demanding in terms of resources when the dataset is large, if training time scales like
O(N p ) with p > 1.
In the more recent Support Vector Mixture model [5], the author shows how to replace the
experts (typically neural networks) by SVMs and gives a learning algorithm for this model.
Once again the resulting mixture is trained jointly on the whole dataset, and hence does not
solve the quadratic barrier when the dataset is large.
In another divide-and-conquer approach [7], the authors propose to rst divide the training set
using an unsupervised algorithm to cluster the data (typically a mixture of Gaussians), then
train an expert (such as an SVM) on each subset of the data corresponding to a cluster, and
nally recombine the outputs of the experts. Here, the algorithm does indeed train separately the
experts on small datasets, like the present algorithm, but there is no notion of a loop reassigning
the examples to experts according to the prediction made by the gater of how well each expert
performs on each example. Our experiments suggest that this element is essential to the success
of the algorithm.
Finally, the Bayesian Committee Machine [8] is a technique to partition the data into several
subsets, train SVMs on the individual subsets and then use a specic combination scheme based
on the covariance of the test data to combine the predictions. This method scales linearly in the

where c is a small positive constant. In the experiments, c = 1.

number of training data, but is in fact a transductive method as it cannot operate on a single
test example. Like in the previous case, this algorithm assigns the examples randomly to the
experts (however the Bayesian framework would in principle allow to nd better assignments).
Regarding our proposed mixture of SVMs, if the number of experts grows with the number
of examples, and the number of outer loop iterations is a constant, then the total training
time of the experts scales linearly with the number of examples. Indeed, given N the total
N
number of examples, choose the number of expert M such that the ratio M
is a constant r;
Then, if k is the number of outer loop iterations, and if the training time for an SVM with r
examples is O(r ) (empirically is slightly above 2), the total training time of the experts is
O(kr M ) = O(kr1 N ), where k , r and are constants, which gives a total training time
of O(N ). In particular for = 2 that gives O(krN ). The actual total training time should
however also include k times the training time of the gater, which may potentially grow more
rapidly than O(N ). However, it did not appear to be the case in our experiments, thus yielding
apparent linear training time. Future work will focus on methods to reduce the gater training
time and guarantee linear training time per outer loop iteration.

5 Experiments
In this section, we present three sets of experiments comparing the new mixture of SVMs to
other machine learning algorithms. Note that all the SVMs in these experiments have been
trained using SVMTorch [2].

5.1 A Toy Problem

In the rst series of experiments, we rst tested the mixture on an articial toy problem for
which we generated 10,000 training examples and 10,000 test examples. The problem had two
non-linearly separable classes and had two input dimensions. On Figure 1 we show the decision
surfaces obtained rst by a linear SVM, then by a Gaussian SVM, and nally by the proposed
mixture of SVMs. Moreover, in the latter, the gater was a simple linear function and there were
two linear SVMs in the mixture . This articial problem thus shows clearly that the algorithm
seems to work, and is able to combine, even linearly, very simple models in order to produce a
non-linear decision surface.

5.2 A Large-Scale Realistic Problem: Forest

For a more realistic problem, we did a series of experiments on part of the UCI Forest dataset .
We modied the 7-class classication problem into a binary classication problem where the
goal was to separate class 2 from the other 6 classes. Each example was described by 54 input
features, each normalized by dividing by the maximum found on the training set. The dataset
had more than 500,000 examples and this allowed us to prepare a series of experiments as follows:

We kept a separate test set of 50,000 examples to compare the best mixture of SVMs
to other learning algorithms.
We used a validation set of 10,000 examples to select the best mixture of SVMs, varying
the number of experts and the number of hidden units in the gater.
We trained our models on dierent training sets, using from 100,000 to 400,000 examples.
The mixtures had from 10 to 50 expert SVMs with Gaussian kernel and the gater was
an MLP with between 25 and 500 hidden units.

Note that the transfer function h() was still a tanh().

The Forest dataset is available on the UCI website at the following
ftp://ftp.ics.uci.edu/pub/machine-learning-databases/covtype/covtype.info.

address:

1.5

1.5

1.5

0.5

0.5

0.5

0.5

0.5

0.5

1.5

1.5

1.5

0.5

0.5

(a) Linear SVM

1.5

1.5

1.5

0.5

0.5

(b) Gaussian SVM

1.5

1.5

0.5

0.5

1.5

(c) Mixture of two linear

SVMs

Figure 1: Comparison of the decision surfaces obtained by (a) a linear SVM, (b) a Gaussian
SVM, and (c) a linear mixture of two linear SVMs, on a two-dimensional classication toy
problem.
Note that since the number of examples was quite large, we selected the internal training parameters such as the of the Gaussian kernel of the SVMs or the learning rate of the gater
using a held-out portion of the training set. We compared our models to

a single MLP, where the number of hidden units was selected by cross-validation between
25 and 250 units,
a single SVM, where the parameter of the kernel was also selected by cross-validation,
a mixture of SVMs where the gater was replaced by a constant vector, assigning the
same weight value to every expert.
Table 1 gives the results of a rst series of experiments with a xed training set of 100,000
examples. To select among the variants of the gated SVM mixture we considered performance
over the validation set as well as training time. All the SVMs used = 1.7. The selected model
had 50 experts and a gater with 150 hidden units. A model with 500 hidden units would have
given a performance of 8.1% over the test set but would have taken 621 minutes on one machine
(and 388 minutes on 50 machines).

one MLP
one SVM
uniform SVM mixture
gated SVM mixture

Train
Error
17.56
16.03
19.69
5.91

Test
(%)
18.15
16.76
20.31
9.28

Time (minutes)
(1 cpu) (50 cpu)
12
3231
85
2
237
73

Table 1: Comparison of performance between an MLP (100 hidden units), a single SVM, a
uniform SVM mixture where the gater always output the same value for each expert, and nally
a mixture of SVMs as proposed in this paper.
As it can be seen, the gated SVM outperformed all models in terms of training and test error.
Note that the training error of the single SVM is high because its hyper-parameters were selected
to minimize error on the validation set (other values could yield to much lower training error but
larger test error). It was also much faster, even on one machine, than the SVM and since the
mixture could easily be parallelized (each expert can be trained separately), we also reported

the time it took to train on 50 machines. In a rst attempt to understand these results, one
can at least say that the power of the model does not lie only in the MLP gater, since a single
MLP was pretty bad, it is neither only because we used SVMs, since a single SVM was not
as good as the gated mixture, and it was not only because we divided the problem into many
sub-problems since the uniform mixture also performed badly. It seems to be a combination of
all these elements.
We also did a series of experiments in order to see the inuence of the number of hidden units
of the gater as well as the number of experts in the mixture. Figure 2 shows the validation error
of dierent mixtures of SVMs, where the number of hidden units varied from 25 to 500 and the
number of experts varied from 10 to 50. There is a clear performance improvement when the
number of hidden units is increased, while the improvement with additional experts exists but
is not as strong. Note however that the training time increases also rapidly with the number of
hidden units while it slightly decreases with the number of experts if one uses one computer per
expert.
Validation error as a function of the number of hidden units
of the gater and the number of experts

Validation error (%)

14
13
12
11
10
9
8
2550

100

150

50
200

250

Number of hidden
units of the gater

500

10

15

20

25

Number of experts

Figure 2: Comparison of the validation error of dierent mixtures of SVMs with various number
of hidden units and experts.
In order to nd how the algorithm scaled with respect to the number of examples, we then
compared the same mixture of experts (50 experts, 150 hidden units in the gater) on dierent
training set sizes. Table 3 shows the validation error of the mixture of SVMs trained on training
sets of sizes from 100,000 to 400,000. It seems that, at least in this range and for this particular
dataset, the mixture of SVMs scales linearly with respect to the number of examples, and not
quadratically as a classical SVM. It is interesting to see for instance that the mixture of SVMs
was able to solve a problem of 400,000 examples in less than 7 hours (on 50 computers) while it
would have taken more than one month to solve the same problem with a single SVM.
Finally, gure 4 shows the evolution of the training and validation errors of a mixture of 50
SVMs gated by an MLP with 150 hidden units, during 5 iterations of the algorithm. This
should convince that the loop of the algorithm is essential in order to obtain good performance.
It is also clear that the empirical convergence of the outer loop is extremely rapid.

5.3 Verication on Another Large-Scale Problem

In order to verify that the results obtained on Forest were replicable on other large-scale problems, we tested the SVM mixture on a speech task. We used the Numbers95 dataset [1] and

Training time as a function of the number of train examples

450

Error as a function of the number of training iterations

14
Train error
Validation Error

400

13
12

300

11
Error (%)

Time (min)

350

250
200

10
9

150

100

50
1

6
1.5

2
2.5
3
Number of train examples

3.5

4
5

5
1

x 10

Figure 3: Comparison of the training time

of the same mixture of SVMs (50 experts,
150 hidden units in the gater) trained on
dierent training set sizes, from 100,000 to
400,000.

2
3
4
Number of training iterations

Figure 4: Comparison of the training and

validation errors of the mixture of SVMs as
a function of the number of training iterations.

turned it into a binary classication problem where the task was to separate silence frames from
non-silence frames. The total number of frames was around 540,000 frames. The training set
contained 100,000 randomly chosen frames out of the rst 400,000 frames. The disjoint validation set contained 10,000 randomly chosen frames out of the rst 400,000 frames also. Finally,
the test set contained 50,000 randomly chosen frames out of the last 140,000 frames. Note that
the validation set was used here to select the number of experts in the mixture, the number of
hidden units in the gater, and . Each frame was parameterized using standard methods used
in speech recognition (j-rasta coecients, with rst and second temporal derivatives) and was
thus described by 45 coecients, but we used in fact an input window of three frames, yielding
135 input features per examples.
Table 2 shows a comparison between a single SVM and a mixture of SVMs on this dataset. The
number of experts in the mixture was set to 50, the number of hidden units of the gater was set
to 50, and the of the SVMs was set to 3.0. As it can be seen, the mixture of SVMs was again
many times faster than the single SVM (even on 1 cpu only) but yielded similar generalization
performance.

one SVM
gated SVM mixture

Train
Error
0.98
4.41

Test
(%)
7.57
7.32

Time (minutes)
(1 cpu) (50 cpu)
6787
851
65

Table 2: Comparison of performance between a single SVM and a mixture of SVMs on the
speech dataset.

6 Conclusion
In this paper we have presented a new algorithm to train a mixture of SVMs that gave very good
results compared to classical SVMs either in terms of training time or generalization performance
on two large scale dicult databases. Moreover, the algorithm appears to scale linearly with
the number of examples, at least between 100,000 and 400,000 examples.

These results are extremely encouraging and suggest that the proposed method could allow
training SVM-like models for very large multi-million data sets in a reasonable time. If training
of the neural network gater with stochastic gradient takes time that grows much less than
quadratically, as we conjecture it to be the case for very large data sets (to reach a good enough
solution), then the whole method is clearly sub-quadratic in training time with respect to the
number of training examples. Future work will address several questions: how to guarantee
linear training time for the gater as well as for the experts? can better results be obtained by
tuning the hyper-parameters of each expert separately? Does the approach work well for other
types of experts?

Acknowledgments
RC would like to thank the Swiss NSF for nancial support (project FN2100-061234.00). YB
would like to thank the NSERC funding agency and NCM2 network for support.

References
[1] R.A. Cole, M. Noel, T. Lander, and T. Durham. New telephone speech corpora at CSLU.
Proceedings of the European Conference on Speech Communication and Technology, EUROSPEECH, 1:821824, 1995.
[2] R. Collobert and S. Bengio. SVMTorch: Support vector machines for large-scale regression
problems. Journal of Machine Learning Research, 1:143160, 2001.
[3] C. Cortes and V. Vapnik. Support vector networks. Machine Learning, 20:273297, 1995.
[4] Robert A. Jacobs, Michael I. Jordan, Steven J. Nowlan, and Georey E. Hinton. Adaptive
mixtures of local experts. Neural Computation, 3(1):7987, 1991.
[5] J. T. Kwok. Support vector mixture for classication and regression problems. In Proceedings
of the International Conference on Pattern Recognition (ICPR), pages 255258, Brisbane,
Queensland, Australia, 1998.
[6] E. Osuna, R. Freund, and F. Girosi. Training support vector machines: an application to
face detection. In IEEE conference on Computer Vision and Pattern Recognition, pages
130136, San Juan, Puerto Rico, 1997.
[7] A. Rida, A. Labbi, and C. Pellegrini. Local experts combination trough density decomposition. In International Workshop on AI and Statistics (Uncertainty'99). Morgan Kaufmann,
1999.
[8] V. Tresp. A bayesian committee machine. Neural Computation, 12:27192741, 2000.
[9] V. N. Vapnik. The nature of statistical learning theory. Springer, second edition, 1995.