Paper Abstract and Keywords |
Presentation |
2017-11-10 13:00
[Poster Presentation]
Effect of maximum likelihood estimation after L1 regularization in learning of log-linear models Kazuya Takabatake, Shotaro Akaho (AIST) IBISML2017-86 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
$L_1$ regularization has two functions.
One function is the structure learning by parameter reduction, and another function is parameter learning that adjusts the remaining parameters to meet the observed data.
Ordinary $L_1$ regularization simultaneously performs these two functions.
Here, one possible option is that we firstly use $L_1$ regularization only for the structure learning and determines the remaining parameters, and then use maximum likelihood estimation for the parameter learning.
In this paper, we make performance comparisons between this option and the original $L_1$ regularization in the case that we use them in learning some log-linear models.
Comparisons by theoretical considerations and by numerical results are shown. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
L1 regularization / structure learning / parameter learning / maximum likelihood estimation / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 117, no. 293, IBISML2017-86, pp. 369-375, Nov. 2017. |
Paper # |
IBISML2017-86 |
Date of Issue |
2017-11-02 (IBISML) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
IBISML2017-86 |
Conference Information |
Committee |
IBISML |
Conference Date |
2017-11-08 - 2017-11-10 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
Univ. of Tokyo |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
Information-Based Induction Science Workshop (IBIS2017) |
Paper Information |
Registration To |
IBISML |
Conference Code |
2017-11-IBISML |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Effect of maximum likelihood estimation after L1 regularization in learning of log-linear models |
Sub Title (in English) |
|
Keyword(1) |
L1 regularization |
Keyword(2) |
structure learning |
Keyword(3) |
parameter learning |
Keyword(4) |
maximum likelihood estimation |
Keyword(5) |
|
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Kazuya Takabatake |
1st Author's Affiliation |
Advanced Industrial Science and Technoloty (AIST) |
2nd Author's Name |
Shotaro Akaho |
2nd Author's Affiliation |
Advanced Industrial Science and Technoloty (AIST) |
3rd Author's Name |
|
3rd Author's Affiliation |
() |
4th Author's Name |
|
4th Author's Affiliation |
() |
5th Author's Name |
|
5th Author's Affiliation |
() |
6th Author's Name |
|
6th Author's Affiliation |
() |
7th Author's Name |
|
7th Author's Affiliation |
() |
8th Author's Name |
|
8th Author's Affiliation |
() |
9th Author's Name |
|
9th Author's Affiliation |
() |
10th Author's Name |
|
10th Author's Affiliation |
() |
11th Author's Name |
|
11th Author's Affiliation |
() |
12th Author's Name |
|
12th Author's Affiliation |
() |
13th Author's Name |
|
13th Author's Affiliation |
() |
14th Author's Name |
|
14th Author's Affiliation |
() |
15th Author's Name |
|
15th Author's Affiliation |
() |
16th Author's Name |
|
16th Author's Affiliation |
() |
17th Author's Name |
|
17th Author's Affiliation |
() |
18th Author's Name |
|
18th Author's Affiliation |
() |
19th Author's Name |
|
19th Author's Affiliation |
() |
20th Author's Name |
|
20th Author's Affiliation |
() |
Speaker |
Author-1 |
Date Time |
2017-11-10 13:00:00 |
Presentation Time |
150 minutes |
Registration for |
IBISML |
Paper # |
IBISML2017-86 |
Volume (vol) |
vol.117 |
Number (no) |
no.293 |
Page |
pp.369-375 |
#Pages |
7 |
Date of Issue |
2017-11-02 (IBISML) |
|