Jiongqian Liang, Deepak Ajwani, Patrick K. Nicholson, A. Sala, S. Parthasarathy
An increasing number of applications are modeled and analyzed in network form, where nodes represent entities of interest and edges represent interactions or relationships between entities. Commonly, such relationship analysis tools assume homogeneity in both node type and edge type. Recent research has sought to redress the assumption of homogeneity and focused on mining heterogeneous information networks (HINs) where both nodes and edges can be of different types. Building on such efforts, in this work we articulate a novel approach for mining relationships across entities in such networks while accounting for user preference (prioritization) over relationship type and interestingness metric. We formalize the problem as a top-$k$ lightest paths problem, contextualized in a real-world communication network, and seek to find the $k$ most interesting path instances matching the preferred relationship type. Our solution, PROphetic HEuristic Algorithm for Path Searching (PRO-HEAPS), leverages a combination of novel graph preprocessing techniques, well designed heuristics and the venerable A* search algorithm. We run our algorithm on real-world large-scale graphs and show that our algorithm significantly outperforms a wide variety of baseline approaches with speedups as large as 100X. We also conduct a case study and demonstrate valuable applications of our algorithm.
{"title":"What Links Alice and Bob?: Matching and Ranking Semantic Patterns in Heterogeneous Networks","authors":"Jiongqian Liang, Deepak Ajwani, Patrick K. Nicholson, A. Sala, S. Parthasarathy","doi":"10.1145/2872427.2883007","DOIUrl":"https://doi.org/10.1145/2872427.2883007","url":null,"abstract":"An increasing number of applications are modeled and analyzed in network form, where nodes represent entities of interest and edges represent interactions or relationships between entities. Commonly, such relationship analysis tools assume homogeneity in both node type and edge type. Recent research has sought to redress the assumption of homogeneity and focused on mining heterogeneous information networks (HINs) where both nodes and edges can be of different types. Building on such efforts, in this work we articulate a novel approach for mining relationships across entities in such networks while accounting for user preference (prioritization) over relationship type and interestingness metric. We formalize the problem as a top-$k$ lightest paths problem, contextualized in a real-world communication network, and seek to find the $k$ most interesting path instances matching the preferred relationship type. Our solution, PROphetic HEuristic Algorithm for Path Searching (PRO-HEAPS), leverages a combination of novel graph preprocessing techniques, well designed heuristics and the venerable A* search algorithm. We run our algorithm on real-world large-scale graphs and show that our algorithm significantly outperforms a wide variety of baseline approaches with speedups as large as 100X. We also conduct a case study and demonstrate valuable applications of our algorithm.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87047172","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Dominique Ritze, O. Lehmberg, Yaser Oulabi, Christian Bizer
Cross-domain knowledge bases such as DBpedia, YAGO, or the Google Knowledge Graph have gained increasing attention over the last years and are starting to be deployed within various use cases. However, the content of such knowledge bases is far from being complete, far from always being correct, and suffers from deprecation (i.e. population numbers become outdated after some time). Hence, there are efforts to leverage various types of Web data to complement, update and extend such knowledge bases. A source of Web data that potentially provides a very wide coverage are millions of relational HTML tables that are found on the Web. The existing work on using data from Web tables to augment cross-domain knowledge bases reports only aggregated performance numbers. The actual content of the Web tables and the topical areas of the knowledge bases that can be complemented using the tables remain unclear. In this paper, we match a large, publicly available Web table corpus to the DBpedia knowledge base. Based on the matching results, we profile the potential of Web tables for augmenting different parts of cross-domain knowledge bases and report detailed statistics about classes, properties, and instances for which missing values can be filled using Web table data as evidence. In order to estimate the potential quality of the new values, we empirically examine the Local Closed World Assumption and use it to determine the maximal number of correct facts that an ideal data fusion strategy could generate. Using this as ground truth, we compare three data fusion strategies and conclude that knowledge-based trust outperforms PageRank- and voting-based fusion.
{"title":"Profiling the Potential of Web Tables for Augmenting Cross-domain Knowledge Bases","authors":"Dominique Ritze, O. Lehmberg, Yaser Oulabi, Christian Bizer","doi":"10.1145/2872427.2883017","DOIUrl":"https://doi.org/10.1145/2872427.2883017","url":null,"abstract":"Cross-domain knowledge bases such as DBpedia, YAGO, or the Google Knowledge Graph have gained increasing attention over the last years and are starting to be deployed within various use cases. However, the content of such knowledge bases is far from being complete, far from always being correct, and suffers from deprecation (i.e. population numbers become outdated after some time). Hence, there are efforts to leverage various types of Web data to complement, update and extend such knowledge bases. A source of Web data that potentially provides a very wide coverage are millions of relational HTML tables that are found on the Web. The existing work on using data from Web tables to augment cross-domain knowledge bases reports only aggregated performance numbers. The actual content of the Web tables and the topical areas of the knowledge bases that can be complemented using the tables remain unclear. In this paper, we match a large, publicly available Web table corpus to the DBpedia knowledge base. Based on the matching results, we profile the potential of Web tables for augmenting different parts of cross-domain knowledge bases and report detailed statistics about classes, properties, and instances for which missing values can be filled using Web table data as evidence. In order to estimate the potential quality of the new values, we empirically examine the Local Closed World Assumption and use it to determine the maximal number of correct facts that an ideal data fusion strategy could generate. Using this as ground truth, we compare three data fusion strategies and conclude that knowledge-based trust outperforms PageRank- and voting-based fusion.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85361829","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A. Halevy, Natasha Noy, Sunita Sarawagi, Steven Euijong Whang, Xiao Yu
Recently, search engines have invested significant effort to answering entity--attribute queries from structured data, but have focused mostly on queries for frequent attributes. In parallel, several research efforts have demonstrated that there is a long tail of attributes, often thousands per class of entities, that are of interest to users. Researchers are beginning to leverage these new collections of attributes to expand the ontologies that power search engines and to recognize entity--attribute queries. Because of the sheer number of potential attributes, such tasks require us to impose some structure on this long and heavy tail of attributes. This paper introduces the problem of organizing the attributes by expressing the compositional structure of their names as a rule-based grammar. These rules offer a compact and rich semantic interpretation of multi-word attributes, while generalizing from the observed attributes to new unseen ones. The paper describes an unsupervised learning method to generate such a grammar automatically from a large set of attribute names. Experiments show that our method can discover a precise grammar over 100,000 attributes of {sc Countries} while providing a 40-fold compaction over the attribute names. Furthermore, our grammar enables us to increase the precision of attributes from 47% to more than 90% with only a minimal curation effort. Thus, our approach provides an efficient and scalable way to expand ontologies with attributes of user interest.
{"title":"Discovering Structure in the Universe of Attribute Names","authors":"A. Halevy, Natasha Noy, Sunita Sarawagi, Steven Euijong Whang, Xiao Yu","doi":"10.1145/2872427.2882975","DOIUrl":"https://doi.org/10.1145/2872427.2882975","url":null,"abstract":"Recently, search engines have invested significant effort to answering entity--attribute queries from structured data, but have focused mostly on queries for frequent attributes. In parallel, several research efforts have demonstrated that there is a long tail of attributes, often thousands per class of entities, that are of interest to users. Researchers are beginning to leverage these new collections of attributes to expand the ontologies that power search engines and to recognize entity--attribute queries. Because of the sheer number of potential attributes, such tasks require us to impose some structure on this long and heavy tail of attributes. This paper introduces the problem of organizing the attributes by expressing the compositional structure of their names as a rule-based grammar. These rules offer a compact and rich semantic interpretation of multi-word attributes, while generalizing from the observed attributes to new unseen ones. The paper describes an unsupervised learning method to generate such a grammar automatically from a large set of attribute names. Experiments show that our method can discover a precise grammar over 100,000 attributes of {sc Countries} while providing a 40-fold compaction over the attribute names. Furthermore, our grammar enables us to increase the precision of attributes from 47% to more than 90% with only a minimal curation effort. Thus, our approach provides an efficient and scalable way to expand ontologies with attributes of user interest.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82329793","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Twitter (and similar microblogging services) has become a central nexus for discussion of the topics of the day. Twitter data contains rich content and structured information on users' topics of interest and behavior patterns. Correctly analyzing and modeling Twitter data enables the prediction of the user behavior and preference in a variety of practical applications, such as tweet recommendation and followee recommendation. Although a number of models have been developed on Twitter data in prior work, most of these only model the tweets from users, while neglecting their valuable retweet information in the data. Models would enhance their predictive power by incorporating users' retweet content as well as their retweet behavior. In this paper, we propose two novel Bayesian nonparametric models, URM and UCM, on retweet data. Both of them are able to integrate the analysis of tweet text and users' retweet behavior in the same probabilistic framework. Moreover, they both jointly model users' interest in tweet and retweet. As nonparametric models, URM and UCM can automatically determine the parameters of the models based on input data, avoiding arbitrary parameter settings. Extensive experiments on real-world Twitter data show that both URM and UCM are superior to all the baselines, while UCM further outperforms URM, confirming the appropriateness of our models in retweet modeling.
{"title":"Modeling a Retweet Network via an Adaptive Bayesian Approach","authors":"Bin Bi, Junghoo Cho","doi":"10.1145/2872427.2882985","DOIUrl":"https://doi.org/10.1145/2872427.2882985","url":null,"abstract":"Twitter (and similar microblogging services) has become a central nexus for discussion of the topics of the day. Twitter data contains rich content and structured information on users' topics of interest and behavior patterns. Correctly analyzing and modeling Twitter data enables the prediction of the user behavior and preference in a variety of practical applications, such as tweet recommendation and followee recommendation. Although a number of models have been developed on Twitter data in prior work, most of these only model the tweets from users, while neglecting their valuable retweet information in the data. Models would enhance their predictive power by incorporating users' retweet content as well as their retweet behavior. In this paper, we propose two novel Bayesian nonparametric models, URM and UCM, on retweet data. Both of them are able to integrate the analysis of tweet text and users' retweet behavior in the same probabilistic framework. Moreover, they both jointly model users' interest in tweet and retweet. As nonparametric models, URM and UCM can automatically determine the parameters of the models based on input data, avoiding arbitrary parameter settings. Extensive experiments on real-world Twitter data show that both URM and UCM are superior to all the baselines, while UCM further outperforms URM, confirming the appropriateness of our models in retweet modeling.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87179120","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Chikashi Nobata, Joel R. Tetreault, A. Thomas, Yashar Mehdad, Yi Chang
Detection of abusive language in user generated online content has become an issue of increasing importance in recent years. Most current commercial methods make use of blacklists and regular expressions, however these measures fall short when contending with more subtle, less ham-fisted examples of hate speech. In this work, we develop a machine learning based method to detect hate speech on online user comments from two domains which outperforms a state-of-the-art deep learning approach. We also develop a corpus of user comments annotated for abusive language, the first of its kind. Finally, we use our detection tool to analyze abusive language over time and in different settings to further enhance our knowledge of this behavior.
{"title":"Abusive Language Detection in Online User Content","authors":"Chikashi Nobata, Joel R. Tetreault, A. Thomas, Yashar Mehdad, Yi Chang","doi":"10.1145/2872427.2883062","DOIUrl":"https://doi.org/10.1145/2872427.2883062","url":null,"abstract":"Detection of abusive language in user generated online content has become an issue of increasing importance in recent years. Most current commercial methods make use of blacklists and regular expressions, however these measures fall short when contending with more subtle, less ham-fisted examples of hate speech. In this work, we develop a machine learning based method to detect hate speech on online user comments from two domains which outperforms a state-of-the-art deep learning approach. We also develop a corpus of user comments annotated for abusive language, the first of its kind. Finally, we use our detection tool to analyze abusive language over time and in different settings to further enhance our knowledge of this behavior.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82193485","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Christopher J. Riederer, Yunsung Kim, A. Chaintreau, Nitish Korula, Silvio Lattanzi
Linking accounts of the same user across datasets -- even when personally identifying information is removed or unavailable -- is an important open problem studied in many contexts. Beyond many practical applications, (such as cross domain analysis, recommendation, and link prediction), understanding this problem more generally informs us on the privacy implications of data disclosure. Previous work has typically addressed this question using either different portions of the same dataset or observing the same behavior across thematically similar domains. In contrast, the general cross-domain case where users have different profiles independently generated from a common but unknown pattern raises new challenges, including difficulties in validation, and remains under-explored. In this paper, we address the reconciliation problem for location-based datasets and introduce a robust method for this general setting. Location datasets are a particularly fruitful domain to study: such records are frequently produced by users in an increasing number of applications and are highly sensitive, especially when linked to other datasets. Our main contribution is a generic and self-tunable algorithm that leverages any pair of sporadic location-based datasets to determine the most likely matching between the users it contains. While making very general assumptions on the patterns of mobile users, we show that the maximum weight matching we compute is provably correct. Although true cross-domain datasets are a rarity, our experimental evaluation uses two entirely new data collections, including one we crawled, on an unprecedented scale. The method we design outperforms naive rules and prior heuristics. As it combines both sparse and dense properties of location-based data and accounts for probabilistic dynamics of observation, it can be shown to be robust even when data gets sparse.
{"title":"Linking Users Across Domains with Location Data: Theory and Validation","authors":"Christopher J. Riederer, Yunsung Kim, A. Chaintreau, Nitish Korula, Silvio Lattanzi","doi":"10.1145/2872427.2883002","DOIUrl":"https://doi.org/10.1145/2872427.2883002","url":null,"abstract":"Linking accounts of the same user across datasets -- even when personally identifying information is removed or unavailable -- is an important open problem studied in many contexts. Beyond many practical applications, (such as cross domain analysis, recommendation, and link prediction), understanding this problem more generally informs us on the privacy implications of data disclosure. Previous work has typically addressed this question using either different portions of the same dataset or observing the same behavior across thematically similar domains. In contrast, the general cross-domain case where users have different profiles independently generated from a common but unknown pattern raises new challenges, including difficulties in validation, and remains under-explored. In this paper, we address the reconciliation problem for location-based datasets and introduce a robust method for this general setting. Location datasets are a particularly fruitful domain to study: such records are frequently produced by users in an increasing number of applications and are highly sensitive, especially when linked to other datasets. Our main contribution is a generic and self-tunable algorithm that leverages any pair of sporadic location-based datasets to determine the most likely matching between the users it contains. While making very general assumptions on the patterns of mobile users, we show that the maximum weight matching we compute is provably correct. Although true cross-domain datasets are a rarity, our experimental evaluation uses two entirely new data collections, including one we crawled, on an unprecedented scale. The method we design outperforms naive rules and prior heuristics. As it combines both sparse and dense properties of location-based data and accounts for probabilistic dynamics of observation, it can be shown to be robust even when data gets sparse.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74833354","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Rodérick Fanou, Gareth Tyson, Pierre François, A. Sathiaseelan
It is well known that Africa's mobile and fixed Internet infrastructure is progressing at a rapid pace. A flurry of recent research has quantified this, highlighting the expansion of its underlying connectivity network. However, improving the infrastructure is not useful without appropriately provisioned services to utilise it. This paper measures the availability of web content infrastructure in Africa. Whereas others have explored web infrastructure in developed regions, we shed light on practices in developing regions. To achieve this, we apply a comprehensive measurement methodology to collect data from a variety of sources. We focus on a large content delivery network to reveal that Africa's content infrastructure is, indeed, expanding. However, we find much web content is still served from the US and Europe. We discover that many of the problems faced are actually caused by significant inter-AS delays in Africa, which contribute to local ISPs not sharing their cache capacity. We discover that a related problem is the poor DNS configuration used by some ISPs, which confounds the attempts of providers to optimise their delivery. We then explore a number of other websites to show that large web infrastructure deployments are a rarity in Africa and that even regional websites host their services abroad. We conclude by making suggestions for improvements.
{"title":"Pushing the Frontier: Exploring the African Web Ecosystem","authors":"Rodérick Fanou, Gareth Tyson, Pierre François, A. Sathiaseelan","doi":"10.1145/2872427.2882997","DOIUrl":"https://doi.org/10.1145/2872427.2882997","url":null,"abstract":"It is well known that Africa's mobile and fixed Internet infrastructure is progressing at a rapid pace. A flurry of recent research has quantified this, highlighting the expansion of its underlying connectivity network. However, improving the infrastructure is not useful without appropriately provisioned services to utilise it. This paper measures the availability of web content infrastructure in Africa. Whereas others have explored web infrastructure in developed regions, we shed light on practices in developing regions. To achieve this, we apply a comprehensive measurement methodology to collect data from a variety of sources. We focus on a large content delivery network to reveal that Africa's content infrastructure is, indeed, expanding. However, we find much web content is still served from the US and Europe. We discover that many of the problems faced are actually caused by significant inter-AS delays in Africa, which contribute to local ISPs not sharing their cache capacity. We discover that a related problem is the poor DNS configuration used by some ISPs, which confounds the attempts of providers to optimise their delivery. We then explore a number of other websites to show that large web infrastructure deployments are a rarity in Africa and that even regional websites host their services abroad. We conclude by making suggestions for improvements.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79601139","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Online tracking poses a serious privacy challenge that has drawn significant attention in both academia and industry. Existing approaches for preventing user tracking, based on curated blocklists, suffer from limited coverage and coarse-grained resolution for classification, rely on exceptions that impact sites' functionality and appearance, and require significant manual maintenance. In this paper we propose a novel approach, based on the concepts leveraged from $k$-Anonymity, in which users collectively identify unsafe data elements, which have the potential to identify uniquely an individual user, and remove them from requests. We deployed our system to 200,000 German users running the Cliqz Browser or the Cliqz Firefox extension to evaluate its efficiency and feasibility. Results indicate that our approach achieves better privacy protection than blocklists, as provided by Disconnect, while keeping the site breakage to a minimum, even lower than the community-optimized AdBlock Plus. We also provide evidence of the prevalence and reach of trackers to over 21 million pages of 350,000 unique sites, the largest scale empirical evaluation to date. 95% of the pages visited contain 3rd party requests to potential trackers and 78% attempt to transfer unsafe data. Tracker organizations are also ranked, showing that a single organization can reach up to 42% of all page visits in Germany.
{"title":"Tracking the Trackers","authors":"Zhonghao Yu, S. Macbeth, Konark Modi, J. M. Pujol","doi":"10.1145/2872427.2883028","DOIUrl":"https://doi.org/10.1145/2872427.2883028","url":null,"abstract":"Online tracking poses a serious privacy challenge that has drawn significant attention in both academia and industry. Existing approaches for preventing user tracking, based on curated blocklists, suffer from limited coverage and coarse-grained resolution for classification, rely on exceptions that impact sites' functionality and appearance, and require significant manual maintenance. In this paper we propose a novel approach, based on the concepts leveraged from $k$-Anonymity, in which users collectively identify unsafe data elements, which have the potential to identify uniquely an individual user, and remove them from requests. We deployed our system to 200,000 German users running the Cliqz Browser or the Cliqz Firefox extension to evaluate its efficiency and feasibility. Results indicate that our approach achieves better privacy protection than blocklists, as provided by Disconnect, while keeping the site breakage to a minimum, even lower than the community-optimized AdBlock Plus. We also provide evidence of the prevalence and reach of trackers to over 21 million pages of 350,000 unique sites, the largest scale empirical evaluation to date. 95% of the pages visited contain 3rd party requests to potential trackers and 78% attempt to transfer unsafe data. Tracker organizations are also ranked, showing that a single organization can reach up to 42% of all page visits in Germany.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76119833","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Web search via voice is becoming increasingly popular, taking advantage of recent advances in automatic speech recognition. Speech recognition systems are trained using audio transcripts, which can be generated by a paid annotator listening to some audio and manually transcribing it. This paper considers an alternative source of training data for speech recognition, called implicit transcription. This is based on Web search clicks and reformulations, which can be interpreted as validating or correcting the recognition done during a real Web search. This can give a large amount of free training data that matches the exact characteristics of real incoming voice searches and the implicit transcriptions can better reflect the needs of real users because they come from the user who generated the audio. On an overall basis we demonstrate that the new training data has value in improving speech recognition. We further show that the in-context feedback from real users can allow the speech recognizer to exploit contextual signals, and reduce the recognition error rate further by up to 23%.
{"title":"Did You Say U2 or YouTube?: Inferring Implicit Transcripts from Voice Search Logs","authors":"Milad Shokouhi, Umut Ozertem, Nick Craswell","doi":"10.1145/2872427.2882994","DOIUrl":"https://doi.org/10.1145/2872427.2882994","url":null,"abstract":"Web search via voice is becoming increasingly popular, taking advantage of recent advances in automatic speech recognition. Speech recognition systems are trained using audio transcripts, which can be generated by a paid annotator listening to some audio and manually transcribing it. This paper considers an alternative source of training data for speech recognition, called implicit transcription. This is based on Web search clicks and reformulations, which can be interpreted as validating or correcting the recognition done during a real Web search. This can give a large amount of free training data that matches the exact characteristics of real incoming voice searches and the implicit transcriptions can better reflect the needs of real users because they come from the user who generated the audio. On an overall basis we demonstrate that the new training data has value in improving speech recognition. We further show that the in-context feedback from real users can allow the speech recognizer to exploit contextual signals, and reduce the recognition error rate further by up to 23%.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91168250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
C. Hsieh, Longqi Yang, Honghao Wei, Mor Naaman, D. Estrin
We propose a new user-centric recommendation model, called Immersive Recommendation, that incorporates cross-platform and diverse personal digital traces into recommendations. Our context-aware topic modeling algorithm systematically profiles users' interests based on their traces from different contexts, and our hybrid recommendation algorithm makes high-quality recommendations by fusing users' personal profiles, item profiles, and existing ratings. Specifically, in this work we target personalized news and local event recommendations for their utility and societal importance. We evaluated the model with a large-scale offline evaluation leveraging users' public Twitter traces. In addition, we conducted a direct evaluation of the model's recommendations in a 33-participant study using Twitter, Facebook and email traces. In the both cases, the proposed model showed significant improvement over the state-of-the-art algorithms, suggesting the value of using this new user-centric recommendation model to improve recommendation quality, including in cold-start situations.
{"title":"Immersive Recommendation: News and Event Recommendations Using Personal Digital Traces","authors":"C. Hsieh, Longqi Yang, Honghao Wei, Mor Naaman, D. Estrin","doi":"10.1145/2872427.2883006","DOIUrl":"https://doi.org/10.1145/2872427.2883006","url":null,"abstract":"We propose a new user-centric recommendation model, called Immersive Recommendation, that incorporates cross-platform and diverse personal digital traces into recommendations. Our context-aware topic modeling algorithm systematically profiles users' interests based on their traces from different contexts, and our hybrid recommendation algorithm makes high-quality recommendations by fusing users' personal profiles, item profiles, and existing ratings. Specifically, in this work we target personalized news and local event recommendations for their utility and societal importance. We evaluated the model with a large-scale offline evaluation leveraging users' public Twitter traces. In addition, we conducted a direct evaluation of the model's recommendations in a 33-participant study using Twitter, Facebook and email traces. In the both cases, the proposed model showed significant improvement over the state-of-the-art algorithms, suggesting the value of using this new user-centric recommendation model to improve recommendation quality, including in cold-start situations.","PeriodicalId":20455,"journal":{"name":"Proceedings of the 25th International Conference on World Wide Web","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89816563","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：