A lossy counting based approach for learning on streams of graphs on a budget

Giovanni Martino, Nicolò Navarin, Alessandro Sperduti

Research output: Chapter in Book/Report/Conference proceedingConference contribution

11 Citations (Scopus)

Abstract

In many problem settings, for example on graph domains, online learning algorithms on streams of data need to respect strict time constraints dictated by the throughput on which the data arrive. When only a limited amount of memory (budget) is available, a learning algorithm will eventually need to discard some of the information used to represent the current solution, thus negatively affecting its classification performance. More importantly, the overhead due to budget management may significantly increase the computational burden of the learning algorithm. In this paper we present a novel approach inspired by the Passive Aggressive and the Lossy Counting algorithms. Our algorithm uses a fast procedure for deleting the less influential features. Moreover, it is able to estimate the weighted frequency of each feature and use it for prediction.

Original languageEnglish
Title of host publicationIJCAI International Joint Conference on Artificial Intelligence
Pages1294-1301
Number of pages8
Publication statusPublished - 2013
Externally publishedYes
Event23rd International Joint Conference on Artificial Intelligence, IJCAI 2013 - Beijing, China
Duration: 3 Aug 20139 Aug 2013

Other

Other23rd International Joint Conference on Artificial Intelligence, IJCAI 2013
CountryChina
CityBeijing
Period3/8/139/8/13

Fingerprint

Learning algorithms
Throughput
Data storage equipment

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Martino, G., Navarin, N., & Sperduti, A. (2013). A lossy counting based approach for learning on streams of graphs on a budget. In IJCAI International Joint Conference on Artificial Intelligence (pp. 1294-1301)

A lossy counting based approach for learning on streams of graphs on a budget. / Martino, Giovanni; Navarin, Nicolò; Sperduti, Alessandro.

IJCAI International Joint Conference on Artificial Intelligence. 2013. p. 1294-1301.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Martino, G, Navarin, N & Sperduti, A 2013, A lossy counting based approach for learning on streams of graphs on a budget. in IJCAI International Joint Conference on Artificial Intelligence. pp. 1294-1301, 23rd International Joint Conference on Artificial Intelligence, IJCAI 2013, Beijing, China, 3/8/13.
Martino G, Navarin N, Sperduti A. A lossy counting based approach for learning on streams of graphs on a budget. In IJCAI International Joint Conference on Artificial Intelligence. 2013. p. 1294-1301
Martino, Giovanni ; Navarin, Nicolò ; Sperduti, Alessandro. / A lossy counting based approach for learning on streams of graphs on a budget. IJCAI International Joint Conference on Artificial Intelligence. 2013. pp. 1294-1301
@inproceedings{dd14bfeb904e4abda52a530691048d71,
title = "A lossy counting based approach for learning on streams of graphs on a budget",
abstract = "In many problem settings, for example on graph domains, online learning algorithms on streams of data need to respect strict time constraints dictated by the throughput on which the data arrive. When only a limited amount of memory (budget) is available, a learning algorithm will eventually need to discard some of the information used to represent the current solution, thus negatively affecting its classification performance. More importantly, the overhead due to budget management may significantly increase the computational burden of the learning algorithm. In this paper we present a novel approach inspired by the Passive Aggressive and the Lossy Counting algorithms. Our algorithm uses a fast procedure for deleting the less influential features. Moreover, it is able to estimate the weighted frequency of each feature and use it for prediction.",
author = "Giovanni Martino and Nicol{\`o} Navarin and Alessandro Sperduti",
year = "2013",
language = "English",
isbn = "9781577356332",
pages = "1294--1301",
booktitle = "IJCAI International Joint Conference on Artificial Intelligence",

}

TY - GEN

T1 - A lossy counting based approach for learning on streams of graphs on a budget

AU - Martino, Giovanni

AU - Navarin, Nicolò

AU - Sperduti, Alessandro

PY - 2013

Y1 - 2013

N2 - In many problem settings, for example on graph domains, online learning algorithms on streams of data need to respect strict time constraints dictated by the throughput on which the data arrive. When only a limited amount of memory (budget) is available, a learning algorithm will eventually need to discard some of the information used to represent the current solution, thus negatively affecting its classification performance. More importantly, the overhead due to budget management may significantly increase the computational burden of the learning algorithm. In this paper we present a novel approach inspired by the Passive Aggressive and the Lossy Counting algorithms. Our algorithm uses a fast procedure for deleting the less influential features. Moreover, it is able to estimate the weighted frequency of each feature and use it for prediction.

AB - In many problem settings, for example on graph domains, online learning algorithms on streams of data need to respect strict time constraints dictated by the throughput on which the data arrive. When only a limited amount of memory (budget) is available, a learning algorithm will eventually need to discard some of the information used to represent the current solution, thus negatively affecting its classification performance. More importantly, the overhead due to budget management may significantly increase the computational burden of the learning algorithm. In this paper we present a novel approach inspired by the Passive Aggressive and the Lossy Counting algorithms. Our algorithm uses a fast procedure for deleting the less influential features. Moreover, it is able to estimate the weighted frequency of each feature and use it for prediction.

UR - http://www.scopus.com/inward/record.url?scp=84896064017&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896064017&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781577356332

SP - 1294

EP - 1301

BT - IJCAI International Joint Conference on Artificial Intelligence

ER -