Learning to Advertise. Introduction Advertising on the Internet = $$$ –Especially search advertising and web page advertising Problem: –Selecting ads.

Slides:



Advertisements
Similar presentations
Query Classification Using Asymmetrical Learning Zheng Zhu Birkbeck College, University of London.
Advertisements

Relevance Feedback Limitations –Must yield result within at most 3-4 iterations –Users will likely terminate the process sooner –User may get irritated.
Spelling Correction for Search Engine Queries Bruno Martins, Mario J. Silva In Proceedings of EsTAL-04, España for Natural Language Processing Presenter:
Chapter 5: Introduction to Information Retrieval
CS6800 Advanced Theory of Computation
Genetic Algorithms Contents 1. Basic Concepts 2. Algorithm
Query Dependent Pseudo-Relevance Feedback based on Wikipedia SIGIR ‘09 Advisor: Dr. Koh Jia-Ling Speaker: Lin, Yi-Jhen Date: 2010/01/24 1.
Content Based Image Clustering and Image Retrieval Using Multiple Instance Learning Using Multiple Instance Learning Xin Chen Advisor: Chengcui Zhang Department.
Search Engines & Search Engine Optimization (SEO) Presentation by Saeed El-Darahali 7 th World Congress on the Management of e-Business.
Date:2011/06/08 吳昕澧 BOA: The Bayesian Optimization Algorithm.
Mechanics of Genetic Programming
TEMPLATE DESIGN © Genetic Algorithm and Poker Rule Induction Wendy Wenjie Xu Supervised by Professor David Aldous, UC.
1 Lecture 8: Genetic Algorithms Contents : Miming nature The steps of the algorithm –Coosing parents –Reproduction –Mutation Deeper in GA –Stochastic Universal.
Introduction to Genetic Algorithms Yonatan Shichel.
Artificial Intelligence Genetic Algorithms and Applications of Genetic Algorithms in Compilers Prasad A. Kulkarni.
Intro to AI Genetic Algorithm Ruth Bergman Fall 2004.
Search Engine Optimization (SEO)
Finding Advertising Keywords on Web Pages Scott Wen-tau YihJoshua Goodman Microsoft Research Vitor R. Carvalho Carnegie Mellon University.
1/16 Final project: Web Page Classification By: Xiaodong Wang Yanhua Wang Haitang Wang University of Cincinnati.
For REAL MEN REAL STYLE.  Search Engine Optimization  SEO is strategies, techniques and tactics to improve or promote a website in order to get a.
Genetic Algorithms Overview Genetic Algorithms: a gentle introduction –What are GAs –How do they work/ Why? –Critical issues Use in Data Mining –GAs.
Repository Method to suit different investment strategies Alma Lilia Garcia & Edward Tsang.
Introduction The large amount of traffic nowadays in Internet comes from social video streams. Internet Service Providers can significantly enhance local.
Title Extraction from Bodies of HTML Documents and its Application to Web Page Retrieval Microsoft Research Asia Yunhua Hu, Guomao Xin, Ruihua Song, Guoping.
Problems Premature Convergence Lack of genetic diversity Selection noise or variance Destructive effects of genetic operators Cloning Introns and Bloat.
An Approach of Artificial Intelligence Application for Laboratory Tests Evaluation Ş.l.univ.dr.ing. Corina SĂVULESCU University of Piteşti.
Cristian Urs and Ben Riveira. Introduction The article we chose focuses on improving the performance of Genetic Algorithms by: Use of predictive models.
Redeeming Relevance for Subject Search in Citation Indexes Shannon Bradshaw The University of Iowa
Search Engines & Search Engine Optimization (SEO).
UOS 1 Ontology Based Personalized Search Zhang Tao The University of Seoul.
When Experts Agree: Using Non-Affiliated Experts To Rank Popular Topics Meital Aizen.
Lecture 8: 24/5/1435 Genetic Algorithms Lecturer/ Kawther Abas 363CS – Artificial Intelligence.
Zorica Stanimirović Faculty of Mathematics, University of Belgrade
Boltzmann Machine (BM) (§6.4) Hopfield model + hidden nodes + simulated annealing BM Architecture –a set of visible nodes: nodes can be accessed from outside.
Xiaoying Gao Computer Science Victoria University of Wellington Intelligent Agents COMP 423.
What is Genetic Programming? Genetic programming is a model of programming which uses the ideas (and some of the terminology) of biological evolution to.
XP New Perspectives on The Internet, Sixth Edition— Comprehensive Tutorial 3 1 Searching the Web Using Search Engines and Directories Effectively Tutorial.
An Introduction to Genetic Algorithms Lecture 2 November, 2010 Ivan Garibay
Chapter 6: Information Retrieval and Web Search
Evolving Virtual Creatures & Evolving 3D Morphology and Behavior by Competition Papers by Karl Sims Presented by Sarah Waziruddin.
Impedance Coupling in Content-targeted Advertising Berthier Ribeiro-Neto, UFMG, Brazil Marco Cristo, UFMG, Brazil Paulo Golgher, Akwan, Brazil Edleno Moura.
Neural and Evolutionary Computing - Lecture 9 1 Evolutionary Neural Networks Design  Motivation  Evolutionary training  Evolutionary design of the architecture.
Introduction to Digital Libraries hussein suleman uct cs honours 2003.
GENETIC ALGORITHMS.  Genetic algorithms are a form of local search that use methods based on evolution to make small changes to a popula- tion of chromosomes.
Parallel and Distributed Searching. Lecture Objectives Review Boolean Searching Indicate how Searches may be carried out in parallel Overview Distributed.
Contextual Ranking of Keywords Using Click Data Utku Irmak, Vadim von Brzeski, Reiner Kraft Yahoo! Inc ICDE 09’ Datamining session Summarized.
Tuning Before Feedback: Combining Ranking Discovery and Blind Feedback for Robust Retrieval* Weiguo Fan, Ming Luo, Li Wang, Wensi Xi, and Edward A. Fox.
How Useful are Your Comments? Analyzing and Predicting YouTube Comments and Comment Ratings Stefan Siersdorfer, Sergiu Chelaru, Wolfgang Nejdl, Jose San.
LOGO 1 Corroborate and Learn Facts from the Web Advisor : Dr. Koh Jia-Ling Speaker : Tu Yi-Lang Date : Shubin Zhao, Jonathan Betz (KDD '07 )
Chapter 9 Genetic Algorithms.  Based upon biological evolution  Generate successor hypothesis based upon repeated mutations  Acts as a randomized parallel.
Genetic Algorithms Genetic algorithms provide an approach to learning that is based loosely on simulated evolution. Hypotheses are often described by bit.
Genetic Algorithms What is a GA Terms and definitions Basic algorithm.
Harvesting Social Knowledge from Folksonomies Harris Wu, Mohammad Zubair, Kurt Maly, Harvesting social knowledge from folksonomies, Proceedings of the.
2101INT – Principles of Intelligent Systems Lecture 11.
EE749 I ntroduction to Artificial I ntelligence Genetic Algorithms The Simple GA.
Effective Automatic Image Annotation Via A Coherent Language Model and Active Learning Rong Jin, Joyce Y. Chai Michigan State University Luo Si Carnegie.
Automated discovery in math Machine learning techniques (GP, ILP, etc.) have been successfully applied in science Machine learning techniques (GP, ILP,
Post-Ranking query suggestion by diversifying search Chao Wang.
Learning in a Pairwise Term-Term Proximity Framework for Information Retrieval Ronan Cummins, Colm O’Riordan Digital Enterprise Research Institute SIGIR.
An Introduction to Genetic Algorithms Lecture 2 November, 2010 Ivan Garibay
Agenda  INTRODUCTION  GENETIC ALGORITHMS  GENETIC ALGORITHMS FOR EXPLORING QUERY SPACE  SYSTEM ARCHITECTURE  THE EFFECT OF DIFFERENT MUTATION RATES.
Instance Discovery and Schema Matching With Applications to Biological Deep Web Data Integration Tantan Liu, Fan Wang, Gagan Agrawal {liut, wangfa,
Estimation of Distribution Algorithm and Genetic Programming Structure Complexity Lab,Seoul National University KIM KANGIL.
 Presented By: Abdul Aziz Ghazi  Roll No:  Presented to: Sir Harris.
LEARNING IN A PAIRWISE TERM-TERM PROXIMITY FRAMEWORK FOR INFORMATION RETRIEVAL Ronan Cummins, Colm O’Riordan (SIGIR’09) Speaker : Yi-Ling Tai Date : 2010/03/15.
Genetic Algorithms.
Evolutionary Algorithms Jim Whitehead
Example: Applying EC to the TSP Problem
Boltzmann Machine (BM) (§6.4)
Learning to Rank with Ties
Presentation transcript:

Learning to Advertise

Introduction Advertising on the Internet = $$$ –Especially search advertising and web page advertising Problem: –Selecting ads relevant to the users, and profitable for the publishers and advertisers In this paper: –A new framework for associating ads with web pages based on Genetic Programming –Finding a ranking function to increase overall precision and minimize the number of misplacements

Search advertising Paid placement strategy –Advertiser pays a placement fee to get a prominent position in ad lists The most popular = Keyword-targeted advertising –Query terms matched against ads keywords (provided by advertisers) Has led to Content-targeted advertising –Use of the content of a web page to select ads to display

Ads Ad is composed of –Title + textual description + hyperlink –Set of keywords chosen by advertiser Describe the topics that should appear on the web page Pay for each keyword Campaign = group of ads related to one product –Only one ad per campaign should be placed in a web page in order to ensure fair use of the page advertising space

Vector Space Model Document Query

Problem to address Ranking ads according to their relevance to the page  ranking function –Indicates how relevant is an ad to a web page: rank(a i,p) Tools: –Statistics about the structural parts of the ads Term frequency in an ad or in a group of ads, number of ads, … –Keywords provided by the advertisers

Genetic Programming Machine learning technique to find solutions optimized for certain problem characteristics Based on biological inheritance and evolution 3 key components: –Individuals –Fitness function: indicator of individuals performance –Genetic operations, to create more diverse and better individuals Reproduction: breed new individuals identical to their parents Crossover: breed an individual sharing attributes with 2 parents Mutation: simulates the deviations occurring during reproduction

Individuals Individuals are presented using a tree structure – easy parsing, easy implementing and interpretation: Nodes in the tree represent functions, leaves represent terminals (statistics about structural parts of the ads and the information provided by the advertisers like keywords )

List of Terminals Functions: P Stands for different structural parts of the ad (keywords, title, etc.) G indicated whether the ads are grouped.

Evaluation The evaluation function should take into consideration the number of relevant ads and the order in which they appear, that is it should be a combination of precision and recall. An example is given by:

Genetic Operators The genetic operations in the model are those commonly used in GP, that is – mutation, crossover and reproduction. Mutation – implemented in such a way that a randomly selected sub-tree is replaced by a new sub-tree, also created randomly. Crossover – taking two trees and exchanging randomly selected sub nodes of these trees forming two new children. Reproduction – Cloning a selected individual to the next generation.

Crossover

Best Individual The last step in the GP is determining the best individual to applied to the test set. The natural choice id the individual with best performance in the training set, but due to over fitting during the learning, the best individuals evolve over Ng generations are applied to a second document collection, which we call a validation set. Avg. performance of individual in the training and validation sets. Corresponding standard deviation.

Experiments Test Collection: A set of 100 pages extracted from a Brazilian newspaper. Cover different subjects. 50 pages were used as training, 30 pages for validation, and 20 pages for test. Relevant ads: Same pooling method used to evaluate the TREC web based collection. The size of the population was fixed at 750 individuals. Crossover, Mutation and Reproduction at rates 85%, 10% and 5%, respectively. 30 generations.

Results Comparing results against the AAK_H method which consists the Cosine similarity function to match a page to the ad.

Conclusions Genetic Programming => discover a good ranking estimation Very effective in placing ads in web pages By using a real ad collection and web pages from a Brazilian newspaper, the GP model obtained a gain of 61.7% over the baseline method.