From e18a9a820895aeba3cd23757e9a790b74c538efe Mon Sep 17 00:00:00 2001 From: knmnyn <217067+knmnyn@users.noreply.github.com> Date: Sat, 19 Oct 2024 14:08:42 +0000 Subject: [PATCH] content: import publications from Bibtex --- .../publication/10-1002-asi-23834/cite.bib | 18 +++++++++ .../publication/10-1002-asi-23834/index.md | 31 ++++++++++++++ .../10-1007-978-3-319-48051-0-15/cite.bib | 16 ++++++++ .../10-1007-978-3-319-48051-0-15/index.md | 32 +++++++++++++++ .../10-1007-s-00799-014-0122-2/cite.bib | 19 +++++++++ .../10-1007-s-00799-014-0122-2/index.md | 40 +++++++++++++++++++ .../10-1145-2756406-2756946/cite.bib | 17 ++++++++ .../10-1145-2756406-2756946/index.md | 29 ++++++++++++++ .../10-1145-2808797-2808820/cite.bib | 17 ++++++++ .../10-1145-2808797-2808820/index.md | 30 ++++++++++++++ .../10-1145-2911451-2914698/cite.bib | 17 ++++++++ .../10-1145-2911451-2914698/index.md | 35 ++++++++++++++++ .../10-1145-3132847-3132946/cite.bib | 17 ++++++++ .../10-1145-3132847-3132946/index.md | 36 +++++++++++++++++ .../10-5555-2887007-2887012/cite.bib | 13 ++++++ .../10-5555-2887007-2887012/index.md | 25 ++++++++++++ .../10-5555-3171837-3171848/cite.bib | 13 ++++++ .../10-5555-3171837-3171848/index.md | 25 ++++++++++++ .../10-5555-3298023-3298065/cite.bib | 12 ++++++ .../10-5555-3298023-3298065/index.md | 25 ++++++++++++ 20 files changed, 467 insertions(+) create mode 100644 content/publication/10-1002-asi-23834/cite.bib create mode 100644 content/publication/10-1002-asi-23834/index.md create mode 100644 content/publication/10-1007-978-3-319-48051-0-15/cite.bib create mode 100644 content/publication/10-1007-978-3-319-48051-0-15/index.md create mode 100644 content/publication/10-1007-s-00799-014-0122-2/cite.bib create mode 100644 content/publication/10-1007-s-00799-014-0122-2/index.md create mode 100644 content/publication/10-1145-2756406-2756946/cite.bib create mode 100644 content/publication/10-1145-2756406-2756946/index.md create mode 100644 content/publication/10-1145-2808797-2808820/cite.bib create mode 100644 content/publication/10-1145-2808797-2808820/index.md create mode 100644 content/publication/10-1145-2911451-2914698/cite.bib create mode 100644 content/publication/10-1145-2911451-2914698/index.md create mode 100644 content/publication/10-1145-3132847-3132946/cite.bib create mode 100644 content/publication/10-1145-3132847-3132946/index.md create mode 100644 content/publication/10-5555-2887007-2887012/cite.bib create mode 100644 content/publication/10-5555-2887007-2887012/index.md create mode 100644 content/publication/10-5555-3171837-3171848/cite.bib create mode 100644 content/publication/10-5555-3171837-3171848/index.md create mode 100644 content/publication/10-5555-3298023-3298065/cite.bib create mode 100644 content/publication/10-5555-3298023-3298065/index.md diff --git a/content/publication/10-1002-asi-23834/cite.bib b/content/publication/10-1002-asi-23834/cite.bib new file mode 100644 index 0000000..31366e7 --- /dev/null +++ b/content/publication/10-1002-asi-23834/cite.bib @@ -0,0 +1,18 @@ +@article{10.1002/asi.23834, + abstract = {Big Science and cross-disciplinary collaborations have reshaped the intellectual structure of research areas. A number of works have tried to uncover this hidden intellectual structure by analyzing citation contexts. However, none of them analyzed by document logical structures such as sections. The two major goals of this study are to find characteristics of authors who are highly cited section-wise and to identify the differences in section-wise author networks. This study uses 29,158 of research articles culled from the ACL Anthology, which hosts articles on computational linguistics and natural language processing. We find that the distribution of citations across sections is skewed and that a different set of highly cited authors share distinct academic characteristics, according to their citation locations. Furthermore, the author networks based on citation context similarity reveal that the intellectual structure of a domain differs across different sections.}, + address = {USA}, + author = {An, Juyoung and Kim, Namhee and Kan, Min-Yen and Chandrasekaran, Muthu Kumar and Song, Min}, + doi = {10.1002/asi.23834}, + issn = {2330-1635}, + issue_date = {August 2017}, + journal = {J. Assoc. Inf. Sci. Technol.}, + month = {August}, + number = {8}, + numpages = {14}, + pages = {1975–1988}, + publisher = {John Wiley & Sons, Inc.}, + title = {Exploring characteristics of highly cited authors according to citation location and content}, + url = {https://doi.org/10.1002/asi.23834}, + volume = {68}, + year = {2017} +} diff --git a/content/publication/10-1002-asi-23834/index.md b/content/publication/10-1002-asi-23834/index.md new file mode 100644 index 0000000..4140c72 --- /dev/null +++ b/content/publication/10-1002-asi-23834/index.md @@ -0,0 +1,31 @@ +--- +title: Exploring characteristics of highly cited authors according to citation location + and content +authors: +- Juyoung An +- Namhee Kim +- Min-Yen Kan +- Muthu Kumar Chandrasekaran +- Min Song +date: '2017-08-01' +publishDate: '2024-10-19T14:08:42.168581Z' +publication_types: +- article-journal +publication: '*J. Assoc. Inf. Sci. Technol.*' +doi: 10.1002/asi.23834 +abstract: Big Science and cross-disciplinary collaborations have reshaped the intellectual + structure of research areas. A number of works have tried to uncover this hidden + intellectual structure by analyzing citation contexts. However, none of them analyzed + by document logical structures such as sections. The two major goals of this study + are to find characteristics of authors who are highly cited section-wise and to + identify the differences in section-wise author networks. This study uses 29,158 + of research articles culled from the ACL Anthology, which hosts articles on computational + linguistics and natural language processing. We find that the distribution of citations + across sections is skewed and that a different set of highly cited authors share + distinct academic characteristics, according to their citation locations. Furthermore, + the author networks based on citation context similarity reveal that the intellectual + structure of a domain differs across different sections. +links: +- name: URL + url: https://doi.org/10.1002/asi.23834 +--- diff --git a/content/publication/10-1007-978-3-319-48051-0-15/cite.bib b/content/publication/10-1007-978-3-319-48051-0-15/cite.bib new file mode 100644 index 0000000..105662e --- /dev/null +++ b/content/publication/10-1007-978-3-319-48051-0-15/cite.bib @@ -0,0 +1,16 @@ +@inproceedings{10.1007/978-3-319-48051-0_15, + abstract = {Among several traditional and novel mobile app recommender techniques that utilize a diverse set of app-related features (such as an app’s Twitter followers, various version instances, etc.), which app-related features are the most important indicators for app recommendation? In this paper, we develop a hybrid app recommender framework that integrates a variety of app-related features and recommendation techniques, and then identify the most important indicators for the app recommendation task. Our results reveal an interesting correlation with data from third-party app analytics companies; and suggest that, in the context of mobile app recommendation, more focus could be placed in user and trend analysis via social networks.}, + address = {Berlin, Heidelberg}, + author = {Lin, Jovian and Sugiyama, Kazunari and Kan, Min-Yen and Chua, Tat-Seng}, + booktitle = {Information Retrieval Technology: 12th Asia Information Retrieval Societies Conference, AIRS 2016, Beijing, China, November 30 – December 2, 2016, Proceedings}, + doi = {10.1007/978-3-319-48051-0_15}, + isbn = {978-3-319-48050-3}, + keywords = {Recommender systems, Mobile apps, Gradient tree boosting}, + location = {Beijing, China}, + numpages = {15}, + pages = {197–211}, + publisher = {Springer-Verlag}, + title = {Scrutinizing Mobile App Recommendation: Identifying Important App-Related Indicators}, + url = {https://doi.org/10.1007/978-3-319-48051-0_15}, + year = {2016} +} diff --git a/content/publication/10-1007-978-3-319-48051-0-15/index.md b/content/publication/10-1007-978-3-319-48051-0-15/index.md new file mode 100644 index 0000000..5f988a2 --- /dev/null +++ b/content/publication/10-1007-978-3-319-48051-0-15/index.md @@ -0,0 +1,32 @@ +--- +title: 'Scrutinizing Mobile App Recommendation: Identifying Important App-Related + Indicators' +authors: +- Jovian Lin +- Kazunari Sugiyama +- Min-Yen Kan +- Tat-Seng Chua +date: '2016-01-01' +publishDate: '2024-10-19T14:08:42.190112Z' +publication_types: +- paper-conference +publication: '*Information Retrieval Technology: 12th Asia Information Retrieval Societies + Conference, AIRS 2016, Beijing, China, November 30 – December 2, 2016, Proceedings*' +doi: 10.1007/978-3-319-48051-0_15 +abstract: Among several traditional and novel mobile app recommender techniques that + utilize a diverse set of app-related features (such as an app’s Twitter followers, + various version instances, etc.), which app-related features are the most important + indicators for app recommendation? In this paper, we develop a hybrid app recommender + framework that integrates a variety of app-related features and recommendation techniques, + and then identify the most important indicators for the app recommendation task. + Our results reveal an interesting correlation with data from third-party app analytics + companies; and suggest that, in the context of mobile app recommendation, more focus + could be placed in user and trend analysis via social networks. +tags: +- Recommender systems +- Mobile apps +- Gradient tree boosting +links: +- name: URL + url: https://doi.org/10.1007/978-3-319-48051-0_15 +--- diff --git a/content/publication/10-1007-s-00799-014-0122-2/cite.bib b/content/publication/10-1007-s-00799-014-0122-2/cite.bib new file mode 100644 index 0000000..5bf64fc --- /dev/null +++ b/content/publication/10-1007-s-00799-014-0122-2/cite.bib @@ -0,0 +1,19 @@ +@article{10.1007/s00799-014-0122-2, + abstract = {To help generate relevant suggestions for researchers, recommendation systems have started to leverage the latent interests in the publication profiles of the researchers themselves. While using such a publication citation network has been shown to enhance performance, the network is often sparse, making recommendation difficult. To alleviate this sparsity, in our former work, we identified "potential citation papers" through the use of collaborative filtering. Also, as different logical sections of a paper have different significance, as a secondary contribution, we investigated which sections of papers can be leveraged to represent papers effectively. While this initial approach works well for researchers vested in a single discipline, it generates poor predictions for scientists who work on several different topics in the discipline (hereafter, "intra-disciplinary"). We thus extend our previous work in this paper by proposing an adaptive neighbor selection method to overcome this problem in our imputation-based collaborative filtering framework. On a publicly-available scholarly paper recommendation dataset, we show that recommendation accuracy significantly outperforms state-of-the-art recommendation baselines as measured by nDCG and MRR, when using our adaptive neighbor selection method. While recommendation performance is enhanced for all researchers, improvements are more marked for intra-disciplinary researchers, showing that our method does address the targeted audience.}, + address = {Berlin, Heidelberg}, + author = {Sugiyama, Kazunari and Kan, Min-Yen}, + doi = {10.1007/s00799-014-0122-2}, + issn = {1432-5012}, + issue_date = {June 2015}, + journal = {Int. J. Digit. Libr.}, + keywords = {Recommendation, Information retrieval, Digital library, Collaborative filtering, Citation analysis}, + month = {June}, + number = {2}, + numpages = {19}, + pages = {91–109}, + publisher = {Springer-Verlag}, + title = {A comprehensive evaluation of scholarly paper recommendation using potential citation papers}, + url = {https://doi.org/10.1007/s00799-014-0122-2}, + volume = {16}, + year = {2015} +} diff --git a/content/publication/10-1007-s-00799-014-0122-2/index.md b/content/publication/10-1007-s-00799-014-0122-2/index.md new file mode 100644 index 0000000..65b3345 --- /dev/null +++ b/content/publication/10-1007-s-00799-014-0122-2/index.md @@ -0,0 +1,40 @@ +--- +title: A comprehensive evaluation of scholarly paper recommendation using potential + citation papers +authors: +- Kazunari Sugiyama +- Min-Yen Kan +date: '2015-06-01' +publishDate: '2024-10-19T14:08:42.212532Z' +publication_types: +- article-journal +publication: '*Int. J. Digit. Libr.*' +doi: 10.1007/s00799-014-0122-2 +abstract: To help generate relevant suggestions for researchers, recommendation systems + have started to leverage the latent interests in the publication profiles of the + researchers themselves. While using such a publication citation network has been + shown to enhance performance, the network is often sparse, making recommendation + difficult. To alleviate this sparsity, in our former work, we identified \"potential + citation papers\" through the use of collaborative filtering. Also, as different + logical sections of a paper have different significance, as a secondary contribution, + we investigated which sections of papers can be leveraged to represent papers effectively. + While this initial approach works well for researchers vested in a single discipline, + it generates poor predictions for scientists who work on several different topics + in the discipline (hereafter, \"intra-disciplinary\"). We thus extend our previous + work in this paper by proposing an adaptive neighbor selection method to overcome + this problem in our imputation-based collaborative filtering framework. On a publicly-available + scholarly paper recommendation dataset, we show that recommendation accuracy significantly + outperforms state-of-the-art recommendation baselines as measured by nDCG and MRR, + when using our adaptive neighbor selection method. While recommendation performance + is enhanced for all researchers, improvements are more marked for intra-disciplinary + researchers, showing that our method does address the targeted audience. +tags: +- Recommendation +- Information retrieval +- Digital library +- Collaborative filtering +- Citation analysis +links: +- name: URL + url: https://doi.org/10.1007/s00799-014-0122-2 +--- diff --git a/content/publication/10-1145-2756406-2756946/cite.bib b/content/publication/10-1145-2756406-2756946/cite.bib new file mode 100644 index 0000000..e66f765 --- /dev/null +++ b/content/publication/10-1145-2756406-2756946/cite.bib @@ -0,0 +1,17 @@ +@inproceedings{10.1145/2756406.2756946, + abstract = {We address the tasks of recovering bibliographic and document structure metadata from scholarly documents. We leverage higher order semi-Markov conditional random fields to model long-distance label sequences, improving upon the performance of the linear-chain conditional random field model. We introduce the notion of extensible features, which allows the expensive inference process to be simplified through memoization, resulting in lower computational complexity. Our method significantly betters the state-of-the-art on three related scholarly document extraction tasks.}, + address = {New York, NY, USA}, + author = {Cuong, Nguyen Viet and Chandrasekaran, Muthu Kumar and Kan, Min-Yen and Lee, Wee Sun}, + booktitle = {Proceedings of the 15th ACM/IEEE-CS Joint Conference on Digital Libraries}, + doi = {10.1145/2756406.2756946}, + isbn = {9781450335942}, + keywords = {metadata extraction, logical structure discovery, conditional random fields}, + location = {Knoxville, Tennessee, USA}, + numpages = {4}, + pages = {61–64}, + publisher = {Association for Computing Machinery}, + series = {JCDL '15}, + title = {Scholarly Document Information Extraction using Extensible Features for Efficient Higher Order Semi-CRFs}, + url = {https://doi.org/10.1145/2756406.2756946}, + year = {2015} +} diff --git a/content/publication/10-1145-2756406-2756946/index.md b/content/publication/10-1145-2756406-2756946/index.md new file mode 100644 index 0000000..be26b7e --- /dev/null +++ b/content/publication/10-1145-2756406-2756946/index.md @@ -0,0 +1,29 @@ +--- +title: Scholarly Document Information Extraction using Extensible Features for Efficient + Higher Order Semi-CRFs +authors: +- Nguyen Viet Cuong +- Muthu Kumar Chandrasekaran +- Min-Yen Kan +- Wee Sun Lee +date: '2015-01-01' +publishDate: '2024-10-19T14:08:42.182824Z' +publication_types: +- paper-conference +publication: '*Proceedings of the 15th ACM/IEEE-CS Joint Conference on Digital Libraries*' +doi: 10.1145/2756406.2756946 +abstract: We address the tasks of recovering bibliographic and document structure + metadata from scholarly documents. We leverage higher order semi-Markov conditional + random fields to model long-distance label sequences, improving upon the performance + of the linear-chain conditional random field model. We introduce the notion of extensible + features, which allows the expensive inference process to be simplified through + memoization, resulting in lower computational complexity. Our method significantly + betters the state-of-the-art on three related scholarly document extraction tasks. +tags: +- metadata extraction +- logical structure discovery +- conditional random fields +links: +- name: URL + url: https://doi.org/10.1145/2756406.2756946 +--- diff --git a/content/publication/10-1145-2808797-2808820/cite.bib b/content/publication/10-1145-2808797-2808820/cite.bib new file mode 100644 index 0000000..d5a631b --- /dev/null +++ b/content/publication/10-1145-2808797-2808820/cite.bib @@ -0,0 +1,17 @@ +@inproceedings{10.1145/2808797.2808820, + abstract = {We study how users of multiple online social networks (OSNs) employ and share information by studying a common user pool that use six OSNs -- Flickr, Google+, Instagram, Tumblr, Twitter, and YouTube. We analyze the temporal and topical signature of users' sharing behaviour, showing how they exhibit distinct behaviorial patterns on different networks. We also examine cross-sharing (i.e., the act of user broadcasting their activity to multiple OSNs near-simultaneously), a previously-unstudied behaviour and demonstrate how certain OSNs play the roles of originating source and destination sinks.}, + address = {New York, NY, USA}, + author = {Lim, Bang Hui and Lu, Dongyuan and Chen, Tao and Kan, Min-Yen}, + booktitle = {Proceedings of the 2015 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining 2015}, + doi = {10.1145/2808797.2808820}, + isbn = {9781450338547}, + keywords = {Online Social Networks, cross-sharing, user behaviour}, + location = {Paris, France}, + numpages = {8}, + pages = {113–120}, + publisher = {Association for Computing Machinery}, + series = {ASONAM '15}, + title = {#mytweet via Instagram: Exploring User Behaviour across Multiple Social Networks}, + url = {https://doi.org/10.1145/2808797.2808820}, + year = {2015} +} diff --git a/content/publication/10-1145-2808797-2808820/index.md b/content/publication/10-1145-2808797-2808820/index.md new file mode 100644 index 0000000..3e29d83 --- /dev/null +++ b/content/publication/10-1145-2808797-2808820/index.md @@ -0,0 +1,30 @@ +--- +title: '#mytweet via Instagram: Exploring User Behaviour across Multiple Social Networks' +authors: +- Bang Hui Lim +- Dongyuan Lu +- Tao Chen +- Min-Yen Kan +date: '2015-01-01' +publishDate: '2024-10-19T14:08:42.205353Z' +publication_types: +- paper-conference +publication: '*Proceedings of the 2015 IEEE/ACM International Conference on Advances + in Social Networks Analysis and Mining 2015*' +doi: 10.1145/2808797.2808820 +abstract: We study how users of multiple online social networks (OSNs) employ and + share information by studying a common user pool that use six OSNs -- Flickr, Google+, + Instagram, Tumblr, Twitter, and YouTube. We analyze the temporal and topical signature + of users' sharing behaviour, showing how they exhibit distinct behaviorial patterns + on different networks. We also examine cross-sharing (i.e., the act of user broadcasting + their activity to multiple OSNs near-simultaneously), a previously-unstudied behaviour + and demonstrate how certain OSNs play the roles of originating source and destination + sinks. +tags: +- Online Social Networks +- cross-sharing +- user behaviour +links: +- name: URL + url: https://doi.org/10.1145/2808797.2808820 +--- diff --git a/content/publication/10-1145-2911451-2914698/cite.bib b/content/publication/10-1145-2911451-2914698/cite.bib new file mode 100644 index 0000000..22f6bf7 --- /dev/null +++ b/content/publication/10-1145-2911451-2914698/cite.bib @@ -0,0 +1,17 @@ +@inproceedings{10.1145/2911451.2914698, + abstract = {Many organizations possess social media accounts on different social networks, but these profiles are not always linked. End applications, users, as well as the organization themselves, can benefit when the profiles are appropriately identified and linked. Most existing works on social network entity linking focus on linking individuals, and do not model features specific for organizational linking. We address this gap not only to link official social media accounts but also to discover and solve the identification and linking of associated affiliate accounts -- such as geographical divisions and brands -- which are important to distinguish.We instantiate our method for classifying profiles on social network services for Twitter and Facebook, which major organizations use. We classify profiles as to whether they belong to an organization or its affiliates. Our best classifier achieves an accuracy of 0.976 on average in both datasets, significantly improving baselines that exploit the features used in state-of-the-art comparable user linkage strategies.}, + address = {New York, NY, USA}, + author = {Cheng, Jerome and Sugiyama, Kazunari and Kan, Min-Yen}, + booktitle = {Proceedings of the 39th International ACM SIGIR Conference on Research and Development in Information Retrieval}, + doi = {10.1145/2911451.2914698}, + isbn = {9781450340694}, + keywords = {organization entity profiling, organizational social profiles, record linkage, social networks}, + location = {Pisa, Italy}, + numpages = {4}, + pages = {901–904}, + publisher = {Association for Computing Machinery}, + series = {SIGIR '16}, + title = {Linking Organizational Social Network Profiles}, + url = {https://doi.org/10.1145/2911451.2914698}, + year = {2016} +} diff --git a/content/publication/10-1145-2911451-2914698/index.md b/content/publication/10-1145-2911451-2914698/index.md new file mode 100644 index 0000000..cf85c89 --- /dev/null +++ b/content/publication/10-1145-2911451-2914698/index.md @@ -0,0 +1,35 @@ +--- +title: Linking Organizational Social Network Profiles +authors: +- Jerome Cheng +- Kazunari Sugiyama +- Min-Yen Kan +date: '2016-01-01' +publishDate: '2024-10-19T14:08:42.197769Z' +publication_types: +- paper-conference +publication: '*Proceedings of the 39th International ACM SIGIR Conference on Research + and Development in Information Retrieval*' +doi: 10.1145/2911451.2914698 +abstract: Many organizations possess social media accounts on different social networks, + but these profiles are not always linked. End applications, users, as well as the + organization themselves, can benefit when the profiles are appropriately identified + and linked. Most existing works on social network entity linking focus on linking + individuals, and do not model features specific for organizational linking. We address + this gap not only to link official social media accounts but also to discover and + solve the identification and linking of associated affiliate accounts -- such as + geographical divisions and brands -- which are important to distinguish.We instantiate + our method for classifying profiles on social network services for Twitter and Facebook, + which major organizations use. We classify profiles as to whether they belong to + an organization or its affiliates. Our best classifier achieves an accuracy of 0.976 + on average in both datasets, significantly improving baselines that exploit the + features used in state-of-the-art comparable user linkage strategies. +tags: +- organization entity profiling +- organizational social profiles +- record linkage +- social networks +links: +- name: URL + url: https://doi.org/10.1145/2911451.2914698 +--- diff --git a/content/publication/10-1145-3132847-3132946/cite.bib b/content/publication/10-1145-3132847-3132946/cite.bib new file mode 100644 index 0000000..dc9e025 --- /dev/null +++ b/content/publication/10-1145-3132847-3132946/cite.bib @@ -0,0 +1,17 @@ +@inproceedings{10.1145/3132847.3132946, + abstract = {We introduce a general, interest-aware topic model (IATM), in which known higher-level interests on topics expressed by each user can be modeled. We then specialize the IATM for use in consumer health forum thread recommendation by equating each user's self-reported medical conditions as interests and topics as symptoms of treatments for recommendation. The IATM additionally models the implicit interests embodied by users' textual descriptions in their profiles. To further enhance the personalized nature of the recommendations, we introduce jointly normalized collaborative topic regression (JNCTR) which captures how users interact with the various symptoms belonging to the same clinical condition. In our experiments on two real-world consumer health forums, our proposed model significantly outperforms competitive state-of-the-art baselines by over 10% in recall. Importantly, we show that our IATM+JNCTR pipeline also imbues the recommendation process with added transparency, allowing a recommendation system to justify its recommendation with respect to each user's interest in certain health conditions.}, + address = {New York, NY, USA}, + author = {Halder, Kishaloy and Kan, Min-Yen and Sugiyama, Kazunari}, + booktitle = {Proceedings of the 2017 ACM on Conference on Information and Knowledge Management}, + doi = {10.1145/3132847.3132946}, + isbn = {9781450349185}, + keywords = {topic models, recommender systems, graphical model, collaborative filtering}, + location = {Singapore, Singapore}, + numpages = {10}, + pages = {1589–1598}, + publisher = {Association for Computing Machinery}, + series = {CIKM '17}, + title = {Health Forum Thread Recommendation Using an Interest Aware Topic Model}, + url = {https://doi.org/10.1145/3132847.3132946}, + year = {2017} +} diff --git a/content/publication/10-1145-3132847-3132946/index.md b/content/publication/10-1145-3132847-3132946/index.md new file mode 100644 index 0000000..07ac3fe --- /dev/null +++ b/content/publication/10-1145-3132847-3132946/index.md @@ -0,0 +1,36 @@ +--- +title: Health Forum Thread Recommendation Using an Interest Aware Topic Model +authors: +- Kishaloy Halder +- Min-Yen Kan +- Kazunari Sugiyama +date: '2017-01-01' +publishDate: '2024-10-19T14:08:42.138840Z' +publication_types: +- paper-conference +publication: '*Proceedings of the 2017 ACM on Conference on Information and Knowledge + Management*' +doi: 10.1145/3132847.3132946 +abstract: We introduce a general, interest-aware topic model (IATM), in which known + higher-level interests on topics expressed by each user can be modeled. We then + specialize the IATM for use in consumer health forum thread recommendation by equating + each user's self-reported medical conditions as interests and topics as symptoms + of treatments for recommendation. The IATM additionally models the implicit interests + embodied by users' textual descriptions in their profiles. To further enhance the + personalized nature of the recommendations, we introduce jointly normalized collaborative + topic regression (JNCTR) which captures how users interact with the various symptoms + belonging to the same clinical condition. In our experiments on two real-world consumer + health forums, our proposed model significantly outperforms competitive state-of-the-art + baselines by over 10% in recall. Importantly, we show that our IATM+JNCTR pipeline + also imbues the recommendation process with added transparency, allowing a recommendation + system to justify its recommendation with respect to each user's interest in certain + health conditions. +tags: +- topic models +- recommender systems +- graphical model +- collaborative filtering +links: +- name: URL + url: https://doi.org/10.1145/3132847.3132946 +--- diff --git a/content/publication/10-5555-2887007-2887012/cite.bib b/content/publication/10-5555-2887007-2887012/cite.bib new file mode 100644 index 0000000..122b2c3 --- /dev/null +++ b/content/publication/10-5555-2887007-2887012/cite.bib @@ -0,0 +1,13 @@ +@inproceedings{10.5555/2887007.2887012, + abstract = {Image tweets are becoming a prevalent form of social media, but little is known about their content - textual and visual - and the relationship between the two mediums. Our analysis of image tweets shows that while visual elements certainly play a large role in image-text relationships, other factors such as emotional elements, also factor into the relationship. We develop Visual-Emotional LDA (VELDA), a novel topic model to capture the image-text correlation from multiple perspectives (namely, visual and emotional).Experiments on real-world image tweets in both English and Chinese and other user generated content, show that VELDA significantly outperforms existing methods on cross-modality image retrieval. Even in other domains where emotion does not factor in image choice directly, our VELDA model demonstrates good generalization ability, achieving higher fidelity modeling of such multimedia documents.}, + author = {Chen, Tao and SalahEldeen, Hany M. and He, Xiangnan and Kan, Min-Yen and Lu, Dongyuan}, + booktitle = {Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence}, + isbn = {0262511290}, + location = {Austin, Texas}, + numpages = {7}, + pages = {30–36}, + publisher = {AAAI Press}, + series = {AAAI'15}, + title = {VELDA: relating an image tweet's text and images}, + year = {2015} +} diff --git a/content/publication/10-5555-2887007-2887012/index.md b/content/publication/10-5555-2887007-2887012/index.md new file mode 100644 index 0000000..accaa75 --- /dev/null +++ b/content/publication/10-5555-2887007-2887012/index.md @@ -0,0 +1,25 @@ +--- +title: "VELDA: relating an image tweet's text and images" +authors: +- Tao Chen +- Hany M. SalahEldeen +- Xiangnan He +- Min-Yen Kan +- Dongyuan Lu +date: '2015-01-01' +publishDate: '2024-10-19T14:08:42.220393Z' +publication_types: +- paper-conference +publication: '*Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence*' +abstract: Image tweets are becoming a prevalent form of social media, but little is + known about their content - textual and visual - and the relationship between the + two mediums. Our analysis of image tweets shows that while visual elements certainly + play a large role in image-text relationships, other factors such as emotional elements, + also factor into the relationship. We develop Visual-Emotional LDA (VELDA), a novel + topic model to capture the image-text correlation from multiple perspectives (namely, + visual and emotional).Experiments on real-world image tweets in both English and + Chinese and other user generated content, show that VELDA significantly outperforms + existing methods on cross-modality image retrieval. Even in other domains where + emotion does not factor in image choice directly, our VELDA model demonstrates good + generalization ability, achieving higher fidelity modeling of such multimedia documents. +--- diff --git a/content/publication/10-5555-3171837-3171848/cite.bib b/content/publication/10-5555-3171837-3171848/cite.bib new file mode 100644 index 0000000..9a0b82f --- /dev/null +++ b/content/publication/10-5555-3171837-3171848/cite.bib @@ -0,0 +1,13 @@ +@inproceedings{10.5555/3171837.3171848, + abstract = {Capturing the semantic interaction of pairs of words across arguments and proper argument representation are both crucial issues in implicit discourse relation recognition. The current state-of-the-art represents arguments as distributional vectors that are computed via bi-directional Long Short-Term Memory networks (BiLSTMs), known to have significant model complexity.In contrast, we demonstrate that word-weighted averaging can encode argument representation which can be incorporated with word pair information efficiently. By saving an order of magnitude in parameters and eschewing the recurrent structure, our proposed model achieves equivalent performance, but trains seven times faster.}, + author = {Lei, Wenqiang and Wang, Xuancong and Liu, Meichun and Ilievski, Ilija and He, Xiangnan and Kan, Min-Yen}, + booktitle = {Proceedings of the 26th International Joint Conference on Artificial Intelligence}, + isbn = {9780999241103}, + location = {Melbourne, Australia}, + numpages = {7}, + pages = {4026–4032}, + publisher = {AAAI Press}, + series = {IJCAI'17}, + title = {SWIM: a simple word interaction model for implicit discourse relation recognition}, + year = {2017} +} diff --git a/content/publication/10-5555-3171837-3171848/index.md b/content/publication/10-5555-3171837-3171848/index.md new file mode 100644 index 0000000..feb3579 --- /dev/null +++ b/content/publication/10-5555-3171837-3171848/index.md @@ -0,0 +1,25 @@ +--- +title: 'SWIM: a simple word interaction model for implicit discourse relation recognition' +authors: +- Wenqiang Lei +- Xuancong Wang +- Meichun Liu +- Ilija Ilievski +- Xiangnan He +- Min-Yen Kan +date: '2017-01-01' +publishDate: '2024-10-19T14:08:42.156820Z' +publication_types: +- paper-conference +publication: '*Proceedings of the 26th International Joint Conference on Artificial + Intelligence*' +abstract: Capturing the semantic interaction of pairs of words across arguments and + proper argument representation are both crucial issues in implicit discourse relation + recognition. The current state-of-the-art represents arguments as distributional + vectors that are computed via bi-directional Long Short-Term Memory networks (BiLSTMs), + known to have significant model complexity.In contrast, we demonstrate that word-weighted + averaging can encode argument representation which can be incorporated with word + pair information efficiently. By saving an order of magnitude in parameters and + eschewing the recurrent structure, our proposed model achieves equivalent performance, + but trains seven times faster. +--- diff --git a/content/publication/10-5555-3298023-3298065/cite.bib b/content/publication/10-5555-3298023-3298065/cite.bib new file mode 100644 index 0000000..26ece44 --- /dev/null +++ b/content/publication/10-5555-3298023-3298065/cite.bib @@ -0,0 +1,12 @@ +@inproceedings{10.5555/3298023.3298065, + abstract = {We tackle the prediction of instructor intervention in student posts from discussion forums in Massive Open Online Courses (MOOCs). Our key finding is that using automatically obtained discourse relations improves the prediction of when instructors intervene in student discussions, when compared with a state-of-the-art, feature-rich baseline. Our supervised classifier makes use of an automatic discourse parser which outputs Penn Discourse Treebank (PDTB) tags that represent in-post discourse features. We show PDTB relation-based features increase the robustness of the classifier and complement baseline features in recalling more diverse instructor intervention patterns. In comprehensive experiments over 14 MOOC offerings from several disciplines, the PDTB discourse features improve performance on average. The resultant models are less dependent on domain-specific vocabulary, allowing them to better generalize to new courses.}, + author = {Chandrasekaran, Muthu Kumar and Epp, Carrie Demmans and Kan, Min-Yen and Litman, Diane}, + booktitle = {Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence}, + location = {San Francisco, California, USA}, + numpages = {7}, + pages = {3415–3421}, + publisher = {AAAI Press}, + series = {AAAI'17}, + title = {Using discourse signals for robust instructor intervention prediction}, + year = {2017} +} diff --git a/content/publication/10-5555-3298023-3298065/index.md b/content/publication/10-5555-3298023-3298065/index.md new file mode 100644 index 0000000..e0d52ce --- /dev/null +++ b/content/publication/10-5555-3298023-3298065/index.md @@ -0,0 +1,25 @@ +--- +title: Using discourse signals for robust instructor intervention prediction +authors: +- Muthu Kumar Chandrasekaran +- Carrie Demmans Epp +- Min-Yen Kan +- Diane Litman +date: '2017-01-01' +publishDate: '2024-10-19T14:08:42.175983Z' +publication_types: +- paper-conference +publication: '*Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence*' +abstract: We tackle the prediction of instructor intervention in student posts from + discussion forums in Massive Open Online Courses (MOOCs). Our key finding is that + using automatically obtained discourse relations improves the prediction of when + instructors intervene in student discussions, when compared with a state-of-the-art, + feature-rich baseline. Our supervised classifier makes use of an automatic discourse + parser which outputs Penn Discourse Treebank (PDTB) tags that represent in-post + discourse features. We show PDTB relation-based features increase the robustness + of the classifier and complement baseline features in recalling more diverse instructor + intervention patterns. In comprehensive experiments over 14 MOOC offerings from + several disciplines, the PDTB discourse features improve performance on average. + The resultant models are less dependent on domain-specific vocabulary, allowing + them to better generalize to new courses. +---