Filters








8,052 Hits in 5.9 sec

On the discriminative power of hyper-parameters in cross-validation and how to choose them

Vito Walter Anelli, Tommaso Di Noia, Eugenio Di Sciascio, Claudio Pomo, Azzurra Ragone
2019 Proceedings of the 13th ACM Conference on Recommender Systems - RecSys '19  
We investigated the discriminative power of nDCG, Precision, Recall, MRR, EFD, EPC, and, finally, we analyzed the role of parameters on model evaluation for Cross-Validation.  ...  We adopted a grid search strategy considering approximately 15 values for each parameter, and we then evaluated each combination of parameters in terms of accuracy and novelty.  ...  Even though the authors designed it to measure the robustness of metrics to changes in the cut-off (previously explored in Information Retrieval in [28] ), the overall procedure, along with the Discriminative  ... 
doi:10.1145/3298689.3347010 dblp:conf/recsys/AnelliNSPR19 fatcat:x5ix23yjdbfmzepmrnivmcanlq

MPEG-7 Visual Descriptors and Discriminant Analysis [chapter]

Jun Zhang, Lei Ye, Jianhua Ma
2010 The Handbook of MPEG Applications  
Since the core experiments applied for the MPEG-7 standards concentrate on single visual descriptor and retrieval ranking, the discriminant power of visual descriptors has not been sufficiently evaluated  ...  Particularly, the applications and technologies should be taken into account for evaluating the discriminant power of visual descriptors.  ...  The image distance can be computed using the recommended distance metrics. And, top k images in the ranked list will be returned as retrieval results.  ... 
doi:10.1002/9780470974582.ch10 fatcat:tiuxgwwzezgrpeuot2lmgev6bu

Evaluating diversified search results using per-intent graded relevance

Tetsuya Sakai, Ruihua Song
2011 Proceedings of the 34th international ACM SIGIR conference on Research and development in Information - SIGIR '11  
Our primary criterion is discriminative power, which represents the reliability of a metric in an experiment.  ...  We compare the properties of existing metrics given the premises that (1) queries may have multiple intents; (2) the likelihood of each intent given a query is available; and (3) graded relevance assessments  ...  Comparisons between l = 1000 and l = 10 show that, except for the top-heavy (n)ERR, using a small document cutoff (i.e. evaluating systems based on fewer data points) reduces discriminative power. 3.  ... 
doi:10.1145/2009916.2010055 dblp:conf/sigir/SakaiS11 fatcat:rh6vfuva3fcmjbpgrujhuvmnf4

User-curated image collections: Modeling and recommendation

Yuncheng Li, Tao Mei, Yang Cong, Jiebo Luo
2015 2015 IEEE International Conference on Big Data (Big Data)  
Most state-of-the-art image retrieval and recommendation systems predominantly focus on individual images.  ...  The evaluations on a large scale Pinterest data set have validated the effectiveness of our proposed methods for modeling and recommending image collections.  ...  maximal discriminant power.  ... 
doi:10.1109/bigdata.2015.7363803 dblp:conf/bigdataconf/LiMCL15 fatcat:ebw6hirf6jbz7iqdrzp5dhzysy

Towards automatic feature vector optimization for multimedia applications

Tobias Schreck, Dieter Fellner, Daniel Keim
2008 Proceedings of the 2008 ACM symposium on Applied computing - SAC '08  
A series of experiments using real and synthetic benchmark data is conducted, the results of which indicate the suitability of the method for unsupervised feature selection and optimization.  ...  We systematically evaluate a recently proposed method for unsupervised discrimination power analysis for feature selection and optimization in multimedia applications.  ...  Acknowledgments We thank Dietmar Saupe and Dejan Vranic for providing the 3D FV extractors and for valuable discussion. Many thanks to Benjamin Bustos for helpful comments.  ... 
doi:10.1145/1363686.1363964 dblp:conf/sac/SchreckFK08 fatcat:djstuclosncbfmvs2fx346ifdu

Effective music tagging through advanced statistical modeling

Jialie Shen, Wang Meng, Shuichang Yan, HweeHwa Pang, Xiansheng Hua
2010 Proceeding of the 33rd international ACM SIGIR conference on Research and development in information retrieval - SIGIR '10  
The structure preserves more discriminative information, leading to more accurate and robust tagging.  ...  Music information retrieval (MIR) holds great promise as a technology for managing large music archives. One of the key components of MIR that has been actively researched into is music tagging.  ...  ACKNOWLEDGEMENTS Shuichang Yan is partially supported by AcRF Tier-1 Grant of R-263-000-464-112, Singapore.  ... 
doi:10.1145/1835449.1835555 dblp:conf/sigir/ShenMYPH10 fatcat:iysleyr5svcinjvdiuva2wconq

Improving Retrieval Performance for Verbose Queries via Axiomatic Analysis of Term Discrimination Heuristic

Mozhdeh Ariannezhad, Ali Montazeralghaem, Hamed Zamani, Azadeh Shakery
2017 Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval - SIGIR '17  
A possible reason for this issue can be the extraneous terms in longer queries that makes it a challenge for the retrieval models to distinguish between the key and complementary concepts of the query.  ...  However, previous studies have shown that the performance of retrieval models varies for di erent query lengths, and it usually degrades when query length increases.  ...  Acknowledgements. is work was supported in part by the Center for Intelligent Information Retrieval and in part by a grant from the Institute for Research in Fundamental Sciences (No. CS1396-4-51).  ... 
doi:10.1145/3077136.3080761 dblp:conf/sigir/AriannezhadMZS17 fatcat:4iaffwtg7vconpe6jnjiqvgdz4

Effective Multi-Query Expansions: Collaborative Deep Networks for Robust Landmark Retrieval

Yang Wang, Xuemin Lin, Lin Wu, Wenjie Zhang
2017 IEEE Transactions on Image Processing  
focus on exploiting geometries of landmarks for similarity matches between candidate photos and a query photo.  ...  Given a query photo issued by a user (q-user), the landmark retrieval is to return a set of photos with their landmarks similar to those of the query, while the existing studies on the landmark retrieval  ...  Settings Evaluation Metric We choose precision-recall as the metrics to evaluate both our model and state-of-the-art methods over top-100 retrieval ranked list w.r.t each query landmark.  ... 
doi:10.1109/tip.2017.2655449 pmid:28103558 fatcat:k6wsv2mjszes7kgegyrg4urtsa

Cross-Domain Image Retrieval with a Dual Attribute-Aware Ranking Network

Junshi Huang, Rogerio Feris, Qiang Chen, Shuicheng Yan
2015 2015 IEEE International Conference on Computer Vision (ICCV)  
and about 90,000 exact offline counterpart images of those online ones.  ...  More specifically, DARN consists of two sub-networks, one for each domain, whose retrieval feature representations are driven by semantic attribute learning.  ...  We conduct extensive experiments proving the effectiveness and robustness of the framework and each one of its components for the clothing retrieval problem.  ... 
doi:10.1109/iccv.2015.127 dblp:conf/iccv/HuangFCY15 fatcat:oxj6mbtg6fduxhpdyfqjkuzsle

Trajectory fingerprint: one-shot human trajectory identification using Siamese network

Zipei Fan, Xuan Song, Quanjun Chen, Renhe Jiang, Ryosuke Shibasaki, Kota Tsubouchi
2020 CCF Transactions on Pervasive Computing and Interaction  
Sliced encoder neural network is designed to encoder the spatiotemporal characteristics and Siamese network is applied to extract the discriminative features for identification.  ...  Moreover, we could hardly define a clear user set for user identification because the set of users are dynamic and changing everyday (there are always new users and inactive users everyday in the real-world  ...  local information better by shortening the path of information traverse in the network, and thus perform better on these metrics focusing on top-rankings.  ... 
doi:10.1007/s42486-020-00034-2 fatcat:7dutlyy6rnb6hnxopei2xawe6a

Exploring Artist Gender Bias in Music Recommendation [article]

Dougal Shakespeare, Lorenzo Porcaro, Emilia Gómez, Carlos Castillo
2020 arXiv   pre-print
To assess group biases introduced by CF, we deploy a recently proposed metric of bias disparity on two listening event datasets: the LFM-1b dataset, and the earlier constructed Celma's dataset.  ...  Our work traces the causes of disparity to variations in input gender distributions and user-item preferences, highlighting the effect such configurations can have on user's gender bias after recommendation  ...  Precision (p@n) captures the proportion of relevant items in top-N recommendations, such that relevance is a binary function that represents the relevance of item for a user .  ... 
arXiv:2009.01715v2 fatcat:pt5q3msxm5eurnfw5xd7zmexay

Deep Metric Learning: A Survey

Kaya, Bilge
2019 Symmetry  
Moreover, sampling strategy, appropriate distance metric, and the structure of the network are the challenging factors for researchers to improve the performance of the network model.  ...  This article aims to reveal the importance of deep metric learning and the problems dealt with in this field in the light of recent studies.  ...  Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/sym11091066 fatcat:ltmug3vjt5aa3dx6gwc537kqca

A multi-label classification method using a hierarchical and transparent representation for paper-reviewer recommendation [article]

Dong Zhang, Shu Zhao, Zhen Duan, Jie Chen, Yangping Zhang, Jie Tang
2019 arXiv   pre-print
Paper-reviewer recommendation task is of significant academic importance for conference chairs and journal editors.  ...  How to effectively and accurately recommend reviewers for the submitted papers is a meaningful and still tough task.  ...  and #61673020), the Provincial Natural Science Foundation of Anhui Province (#1708085QF156), and the Recruitment Project of Anhui University for Academic and Technology Leader.  ... 
arXiv:1912.08976v1 fatcat:c2sb7yupjrgz3alzfudcflup2y

Evaluation with informational and navigational intents

Tetsuya Sakai
2012 Proceedings of the 21st international conference on World Wide Web - WWW '12  
bootstrap test that only considers one run pair at a time, and therefore recommend the former approach for significance testing when a set of runs is available for evaluation.  ...  Our main experimental findings are: (a) In terms of discriminative power which reflects statistical reliability, the proposed metrics, DIN -nDCG and P+Q , are comparable to intent recall and D -nDCG, and  ...  Moreover, note that, by means of discriminative power, we are measuring the robustness of metrics to variations in the choice of topics and therefore the reliability of experiments: we are not discussing  ... 
doi:10.1145/2187836.2187904 dblp:conf/www/Sakai12 fatcat:ciaz5oyuwncanni4yori74z4pi

Web Search Evaluation with Informational and Navigational Intents

Tetsuya Sakai
2013 Journal of Information Processing  
Our main experimental findings are: (a) In terms of discriminative power which reflects statistical reliability, the proposed metrics, DIN -nDCG and P+Q , are comparable to intent recall and D -nDCG, and  ...  bootstrap test that only considers one run pair at a time, and therefore recommend the former approach for significance testing when a set of runs is available for evaluation.  ...  Moreover, note that, by means of discriminative power, we are measuring the robustness of metrics to variations in the choice of topics and therefore the reliability of experiments: we are not discussing  ... 
doi:10.2197/ipsjjip.21.145 fatcat:lciuokv25jc45d2gnttcrorgze
« Previous Showing results 1 — 15 out of 8,052 results