on the extension of trace norm to tensors
TRANSCRIPT
-
8/8/2019 On the extension of trace norm to tensors
1/16
On the extension of trace norm to tensors
Ryota Tomioka1, Kohei Hayashi2, Hisashi Kashima1
1The University of Tokyo2
Nara Institute of Science and Technology
2010/12/10
NIPS2010 Workshop: Tensors, Kernels, and Machine Learning
-
8/8/2019 On the extension of trace norm to tensors
2/16
Convex low-ranktensorcompletion
Sensors
Time
Features
Factors
(loadings)
Core
(interactions)
I1
I2I3
I1
I2
I3
r1
r2
r3r1
r2r3
Tucker decomposition
-
8/8/2019 On the extension of trace norm to tensors
3/16
Conventional formulation (nonconvex)
mode-k productobservation
minimizeC,U1,U2,U3
(Y C 1 U1 2 U2 3 U3) 2
F + regularization.
minimizeX
(YX)
2
F
s.t. rank
(X) (
r1, r2, r3
).
Alternate minimization
Have to fix the rank beforehand
-
8/8/2019 On the extension of trace norm to tensors
4/16
Our approach
MatrixEstimation of
low-rank matrix
(hard)
Trace norm
minimization
(tractable)[Fazel, Hindi, Boyd 01]
Tensor
Estimation oflow-rank tensor
(hard)Rank defined in the sense of
Tucker decomposition
Extended
trace norm
minimization
(tractable)
Generalization
-
8/8/2019 On the extension of trace norm to tensors
5/16
Trace norm regularization (for matrices)
X RRC, m = min(R, C)
Linear sum of
singular-values
Roughly speaking, L1 regulariza6on on the singularvalues.
Stronger regulariza6on > more zero singularvalues >
low rank.
Not obvious for tensors (no singularvalues for tensors)
Xtr =
m
j=1
j(X)
-
8/8/2019 On the extension of trace norm to tensors
6/16
Mode-k unfolding (matricization)
I1
I2 I3
I1
I2 I2 I2
I3
Mode-1 unfolding
I1
I2 I3
I2
I3 I3 I3
I1
Mode-2 unfolding
X(1)
X(2)
-
8/8/2019 On the extension of trace norm to tensors
7/16
Elementary facts about Tucker decomposition
X = C 1 U1 2 U2 3 U3
r1
r2r3
C
I1r1
U1 I2
r2U2
I3
r3
U3
Mode-1 unfolding
Mode-2 unfolding
Mode-3 unfolding
X(1) = U1C(1)(U3 U2)
X(2) =
U2C
(2)(U
1U
3)
X(3) = U3C(3)(U2 U1)
rankr1
rankr2
rankr3
-
8/8/2019 On the extension of trace norm to tensors
8/16
What it means
We can use the trace norm of an unfolding of a
tensorXto learn lowrankX.
Matricization
Tensorization
-
8/8/2019 On the extension of trace norm to tensors
9/16
Approach 1: As a matrix
Pick a mode k, and hope that the tensor to be
learned is low rank in mode k.
minimizeXR
I1IK
12
(Y X) 2F + X(k),
Pro: Basically a matrix problem Theoretical guarantee (Candes & Recht 09)
Con: Have to be lucky to pick the right mode.
-
8/8/2019 On the extension of trace norm to tensors
10/16
Approach 2: Constrained optimization
Constrain so that each unfolding ofXis
simultaneously low rank.
minimizeXR
I1IK
12
(Y X) 2F +K
k=1
kX(k).
k: tuning parameter usually set to 1.
Pro: Jointly regularize every modeCon: Strong constraint
See also Marco Signoretto et al.,10
-
8/8/2019 On the extension of trace norm to tensors
11/16
Approach 3: Mixture of low-rank tensors
Each mixture componentZk is regularized to be
lowrank only in modek.
minimizeZ1,...,ZK
1
2
Y
Kk=1
Zk
2
F
+K
k=1
kZk(k),
Pro: Each Zk takes care of each modeCon: Sum is not low-rank
-
8/8/2019 On the extension of trace norm to tensors
12/16
Numerical experiment
True tensor: Size 50x50x20, rank 7x8x9. No noise (=0).
Random train/test split.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
104
10
2
100
102
Fraction of observed elements
Genera
lizationerror
As a Matrix (mode 1)
As a Matrix (mode 2)As a Matrix (mode 3)
Constraint
Mixture
Tucker (large)
Tucker (exact)Optimization tolerance
Tucker=EM
algo
(nonconvex)
-
8/8/2019 On the extension of trace norm to tensors
13/16
Computation time
Convex formula6on is also fast
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
10
20
30
40
50
Fraction of observed elements
Computationtime(s)
As a Matrix
ConstraintMixture
Tucker (large)
Tucker (exact)
-
8/8/2019 On the extension of trace norm to tensors
14/16
Phase transition behaviour
0 20 40 60 80
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
[10 12 13]
[10 12 9][16 6 8]
[17 13 15]
[20 20 5]
[20 3 2]
[30 12 10]
[32 3 4]
[40 20 6]
[40 3 2]
[40 9 7]
[4 42 3]
[5 4 3]
[7 8 15]
[7 8 9][8 5 6]
Sum of true ranks
Fraction
required
forperfectrecon
struction
Sum of true ranks= min(r1,r2r3)+ min(r2,r3r1)+ min(r3,r1r2)
-
8/8/2019 On the extension of trace norm to tensors
15/16
Summary
Lowrank tensor comple6on can be computed in a convex
op6miza6on problem using the trace norm of the unfoldings.
No need to specify the rank beforehand.
Convex formula6on is more accurate and faster thanconven6onal EMbased Tucker decomposi6on.
Curious phase transi6on found compressivesensing
type analysis is an ongoing work.
Technical report: Arxiv:1010.0789 (including op6miza6on)
Code:
h]p://www.ibis.t.utokyo.ac.jp/RyotaTomioka/So_wares/Tensor
-
8/8/2019 On the extension of trace norm to tensors
16/16
Acknowledgment
This work was supported in part by MEXT KAKENHI
22700138, 22700289, and NTT Communica6on
Science Laboratories.