Unifying Explicit and Implicit Feedback for Top-N Recommendation
|
|
- Andrea King
- 5 years ago
- Views:
Transcription
1 017 IEEE nd International Conference on Big Data Analysis Unifying Exlicit and Imlicit Feedback for To-N Recommendation Siing Liu, Xiaohan Tu and Renfa Li College of Comuter Science and Electronic Engineering Hunan University Changsha 41008, China {liusiing, tutu16103, Abstract In the era of big data, the data are diverse and comlex. The issue that using multi-source data efficiently in recommender system is very essential. To solve this roblem, we rooses a ranking model that integrates exlicit feedback data with imlicit feedback data together. We use weighting factors to measure the imact of different user behaviors on recommendation quality. We solved the data fusion roblem and the To-N items recommendation roblem. We used matrix decomosition for collaborative filtering. Finally, a arallel otimization model based on distributed and arallel comuting is roosed, and the imlementation on Sark is rovided. Through comarison with several models, our model greatly enhanced the items recommendation quality and imroved the scalability and efficiency of ersonalized recommender systems. Keywords-comonent: Recommender systems; Imlicit and exlicit feedback; Collaborative Filtering; To-N recommendation. I. INTRODUCTION Most of recommender systems focus on analyzing exlicit feedback information [3, 5, 6], while ignoring imlicit feedback [4, 0]. That does not effectively analyze the otential value of data and limits the develoment of recommender system. The roblem of integrated with exlicit and imlicit feedback fully for items recommendation, which arises from the actual recommender engineering alication in industry. The data source include Weibo, WeChat, web sites and so on, which include exlicit and imlicit feedback data tyes. Therefore, the study of how to use a large number of imlicit feedback data combined with exlicit feedback information for ersonalized recommendation is the core value of big data. The inut data tyes of the recommender systems include exlicit feedback and imlicit feedback data. Exlicit user feedback refers to the exlicit tendency given by users, such as movie ratings, roduct ratings and so on. However, imlicit feedback includes ointing raise or commenting in Weibo or WeChat, or browsing the Netease news web site, or buying goods in Taobao, etc. Comared with exlicit feedback, imlicit feedback data has a broader alication landscae and a larger data size, but is more difficult to integrate effectively. We know that imlicit rating data has the roblem of information redundancy and are difficult to summarize. The relationshi between data is not very tight, including multile data sources, and data may be heterogeneous. Imlicit feedback does not have negative feedback, that is, even if the user does not act it does not mean that users do not like the item. Imlicit feedback also contains a lot of noise. For examle, the browse behavior may be a age that the user oen the age by default which leads to a higher score. These cannot really reflect the user's references. The value of exlicit feedback indicates the user's real reference, while the imlicit feedback reflects the user's reference, and the value reflects the confidence. If you want to use recommendations that include imlicit feedback, you need an aroriate methodology for your assessment. In addition, the imlicit feedback has a big data size, usually coming from the database and server logs, which requires the recommendation model is efficient and has good scalability. For multi-source data roblems, our aer combines the imlicit feedback and exlicit feedback data to make ersonalized recommendation service. By transforming the recommendation task into the maximization roblem of the robability of user choice behavior, we can directly model the imlicit feedback data, which avoid the introduction of noise. By otimizing the user and the roduct's otential characteristic factor of the roduct we model the choice between the user and the roduct. Based on the distributed matrix and artition strategy, we roosed a arallel Unifying Exlicit and Imlicit Feedback Model (UEIFM) to imrove the scalability and execution efficiency. II. RELATED WORK The Collaborative Filtering (CF) recommendation algorithm based on matrix decomosition has a good erformance. Matrix decomosition technology [3] has gradually relaced the collaborative filtering algorithm based on neighborhood (user or object) to become the research hotsot. This is mainly due to the fact that matrix decomosition has higher rediction recision and better scalability. The robability matrix decomosition [6] reconstructs the scoring matrix by maing the user (roduct) into the low-dimensional sace and estimating the otential feature vector of the user (roduct). The dimensionality reduction technique is used to deal with high dimensional sarse data to imrove the rediction of recision. Bayesian Personalized Ranking (BPR) model has been roosed in [7]. By maximizing the robability of the ositive cases ranked in front of negative cases, the recommendation task was translated into a roblem of learning to rank, and alied to two common recommendation algorithms, namely adative KNN and /17/$ IEEE 35
2 matrix decomosition. The model was evaluated by a rivate data set of a web-based transaction on an online store and a samle of Netflix dataset. However, since the latter did not include imlicit feedback, the researchers converted the exlicit rating to the imlicit rating, and used the AUC as the evaluation measure [8]. We insired by the idea of simulation imlicit feedback. If a combination of exlicit and imlicit user feedback is used, the erformance of the recommended system will be imroved [1]. So far, algorithms suitable for combining exlicit and imlicit user feedback in the recommendation include matrix decomosition, linear and logistic regression [1, 13, 14]. Collaborative filtering using neighborhood models has not yet been adated to a combination of exlicit and imlicit user feedback. The arallel otimization algorithm of matrix decomosition is classified into two categories. One is the otimization method itself, and the other is the alicable distributed comuting environment. Stochastic Gradient Descent (SGD) is articularly effective in solving the matrix decomosition otimization roblems []. The imlementation is simle and efficient. Moreover, the comutational comlexity of each iteration is less. However, arallelizing SGD is a big challenge, because each ste descent deends on the revious udate which will cause overwriting roblems. When the matrix is very large and sarse, the overwriting roblem is more serious. The SGD overwriting efficiency is roosed in a number of revious aers [15, 16, 17, 18, 19]. DSGD [16, 17] block the matrix, by adjusting the udated sequence, to avoid the arallel udate which brings the overwriting conflict. FPSGD [18] further otimized cache and load balancing by using efficient access for sharing memory on the basis of DSGD, but there is a size limit that requires shared memory to be sufficient to fully load the feature matrix. The arallel otimization algorithm based on the MaReduce framework have been imlemented [16, 19], but the inherent overwriting roblem of SGD in arallel distributed environment still cannot be well resolved. Another matrix decomosition algorithm (Alternative Least Squares), can ensure the final convergence of the algorithm by alternately fixing one of the otential feature matrices and otimizing another otential feature matrix. Aiming at the high comutational comlexity of each iteration of ALS, a cyclic coordinate descent (CCD) method was roosed [9], which only udated one variable at a time on the basis of fixing other variables. On this basis, CCD++ is roosed [10], which further reduces the comutational comlexity by changing the sequence of variable udating. The arallel otimization algorithm mentioned above in the dislay feedback recommendation has been well verified. For imlicit feedback, the arallelism is to aroach the otimization objective (least squares error sum) of the traditional matrix decomosition as faster as ossible. The arallelization otimization method roosed in this aer combines the dislay and imlicit feedback recommendation models, imlements the related algorithms based on Aache Sark framework [1], and imroves the rediction accuracy of the recommender system. III. THE PARALLEL UEIFM AND THE SPARK PLATFORM IMPLEMENTATION In the era of big data, data are diverse and naturally formed. The massive data are often structured, but sometimes are semi-structured or unstructured. To achieve the recision of the recommendation, data fusion is an indisensable ste, which make full use of the diversity of the multile data. Our aer integrates multi-source data and constructs a new ranking model based on the integration of exlicit feedback and imlicit feedback data. Because of the large scale of data which requires the recommendation model to be sufficiently efficient and scalable. Then a arallel Unifying Exlicit and Imlicit Feedback Model (UEIFM) is designed and imlemented in the distributed comuting framework. A. UEIFM According to the data from Weibo or WeChat for ointing raise or commenting, browsing the web sites, databases data and logs data, we introduce the data weight, which reflects the information quality of each data source. Suosing there are m exlicit data sources, the overall exlicit weight is s, the k -th data weight is k, the k -th exlicit score is E k, the number of imlicit data sources is n and the imlicit weight is t. The data weight of item v is q v. The imlicit score of item v is I v. We get one value R ij of the rating data matrix as follows: m R s E t q I ij k k v v k1 v1 n (1) Number of k score items of user i th k () q v Average number of score items available to the score grou Number of data items for the vth behavior of user i Average number of data items available to the behavior grou (3) Many values like R ij form a rating matrix. The resulting rating matrix is very large and sarse. Pave the way for the following arallelization, we choose the coordinate matrix on Sark, and use the rogramming model of Resilient Distributed Dataset (RDD). The essence of RDD is a customizable arallel data container. Different dataset formats corresond to different tyes of RDD. We create a distributed matrix named CoordinateRowMatrix with the MatrixEntry instance. Each term is as a tule of row and column values in the form (i: Long, j: Long, value: Double), where i is the row coordinate value, j is the column coordinate value, and value is the rating score. Then transform the tule into MatrixEntry (Long, Long, Double), which facilitates distributed arallel comuting. The otimization of our objective function is as follows: ) F c R x y x y (4) (X,Y) min ui ( ui u i ) ( u i ) x, y ui, u i 36
3 cui is the credibility, the weight of evaluation that user u values item i. Rui is the value of the user-item matrix R, reresenting the user u whose evaluation on the item i. is the regularization arameter, which controls the degree of regularization, and used to avoid over-fitting roblems. x u reresents the feature vector of user u, yi denotes the feature vector of item i. Due to the large amount of data, the stochastic gradient descent method is no longer suitable. Otherwise the ALS is designed to find two low dimensional matrix X ( m* k ) and matrix Y ( k *n ). To aroximate the aroximation R ( m* n), we use the arallel ALS method to get the value ste by ste with iterative aroximation. u F( X, Y) 0 X 1 ( u ) u ( ) (5) X Y C Y I Y C R u (6) The method to get Y i is similar to X u, Ru ( ) is the value of rating matrix we get from function (1). R X Y (7) * T ui u i The rogress of our algorithm is as follows: 1) Calculate the function (5). ) Calculate the function (6). 3) Make the derivative of Y i to be equal 0, then get the value of Y i as above. 4) Practice as the ste 1), ), 3). 5) The X u and Y i vectors are alternately determined and the other is calculated so that the value of the consumtion function is reduced. The rocess above is reeated until the value is convergent or stable. 6) R ui is calculated from function (7). B. Parallel Imlementation of UEIFM on Sark Due to our UEIFM is mainly the iterative udate rocess, in which the fixed x u to udate the y i, or the fixed y i to udate x u. The seudocode of inut is as follows: Inut: F as Number of features. ITERATIONS as the number of iterations. LAMBDA as Regularization coefficient. The exlicit and imlicit user feedback datasets. Outut: User matrix and item factors of features. 1) Calculate R as Grade CoordinateMatrix (Distributed Matrix) by RDD[MatrixEntry] data,because the original tye is tule (long, long, double), convert to MatrixEntry (long, long, double). ) Initialize U F as ms and F M as us matrix randomly. Due to the block size and artitioner,we can get the iteminblocks and userinblocks reectly. 3) Sarkcontext broadcast R and ms and um to all the Executor of s. 4) Iteratively udate items then users, Fixed M, one by one to udate each feature of each user, so that the artial derivative is equal to 0; and then fixed U, one by one to udate each item of each feature, so that the artial derivative is equal to 0. Then comute comute RMSE by user features and item features. 5) Then udate user or item rocess, Solve it with Cholesky solver for least square roblems with L regularization. 6) Finally, get the result of userfacotrs and itemfactors. IV. EXPERIMENT AND EVALUATION A. Exerimental Environment We use the HP-workstation comuter in our lab as exerimental environment. The comuter is configured to 55G memory, 1 cores CPU. We use the lightweight virtualization technology Docker [3], and then create containers over the Ubuntu host OS. Finally, the deloyment of 5-node Sark cluster are ready. We use HDFS (Hadoo Distributed File System) to store exerimental data, and virtual network adater docker0 to establish network connection. A cloud comuting latform for distributed arallel comuting has been built. Master 1 Docker Engine Ubuntu(Host OS) 3 Figure 1. The exerimental environment. 4 B. Exerimental Dataset We used MovieLens [1] as the exerimental datasets to simulate exlicit ratings and imlicit feedback data when comuting the ranking, which includes Movielens-1m, Movielens-10m, and Movielens-0m. Taking one of them as an examle, Movielens-1m contains ratings of 3883 movies and 6040 users. The user number is from 1 to 6040, and movie number is from 1 to 395, but only 706 movies are actually rated. Not all items are rated. The user number and movie number are not continuous, and a number reresents a single user. A single user rates 0 movies at least. The rating density is 4.47%. Details are shown as Table I. 37
4 TABLE I. EXPERIMENTAL MOVIELENS DATASETS datasets users items(actual) ratings density 1m 10m 0m (3706) % (10677) % (6744) % The total number of Movielens datasets s ratings is After calculating, the distribution of ratings are as shown in Table II. We can see that rating score 4 is the most, and second is rating score 3.The lower rating score is much less than the higher rating score. TABLE II. RATING DISTRIBUTION rating numbers Ratio % % % % % % % % % % C. Exeriment Evaluation After the fusion of exlicit feedback and imlicit feedback data, we train the recommendation model, then get the To-N recommendation. To-N recommended recision is generally measured by the recision rate, recall rate [8] and so on. So in our aer, we mainly calculated the rec@n (recision at N), recall@n (Recall at N), and a few additional auxiliary evaluation indicators, which are AUC (Area Under the ROC curve), MAP (Mean Average Precision), which calculate the average recision at all N value, And the NDCG (Normalized Discounted Cumulative Gain). Finally, we calculated the F 1 score. According to our model UEIFM (the integration of the user's exlicit feedback and imlicit feedback data), we make recommendation lists as Ru ( ) for the user, and use the test set to get the test item lists as Tu. ( ) The recision of the recommendation results is defined as: Precision uu Ru ( ) Tu ( ) Ru ( ) uu The Recall for recommendation results are defined as: Recall uu Ru ( ) Tu ( ) uu Tu ( ) (8) (9) The average AUC is usually defined as in [7]: 1 AUC AUC( u) U u (10) U The discounted CG accumulated at a articular rank osition is defined as: reli DCG i1 log ( i 1) (11) And then, NDCG is comuted as: NDCG The F 1 score defined as: DCG (1) IDCG recision recall F 1 recision recall (13) D. Exeriment Results We comare our model UEIFM with ItemKNN [], UserKNN, WRMF [4], and BPRMF [7]. We use MyMediaLite [], which has imlemented some of items recommendation model and evaluated by 5-fold cross validation. It means that the exerimental datasets was divided into training data set and test data set by 0.8 to 0.. We followed to imlement the evaluation metrics. The results are shown in Table III and Table IV. Our UEIFM model got the recision at to-5 and to-10 are and resectively. The recall value of our model are and In the comarison of recision, both rec@5 and rec@10 are better than WRMF AND BPRMF. In the comarison of recall@5 and reacall@10, there is a more significant reduction than WRMF and BPRMF. We are concerned with the recommendation recision, not the recommended relevance, because we get a lot of user-related recommendations, but we only care about the revious To-N recommendations such as 1, 5, or 10. Therefore, we got high recision and low recall. TABLE III. PRECISION AND RECALL COMPARISON Model rec@5 rec@10 recall@5 recall@10 ItemKNN UserKNN WRMF BPRMF UEIFM By using the advantages of using exlicit and imlicit feedback data, the AUC value of our model is , which is better than that of WRMF. Our model s NDCG value is , which is better than WRMF and BPRMF, due to suerimosing the results of calculations on To-5 and To-10. AUC and NDCG showed that the UEIFM is 38
5 ossible to recommend the user's favorite item with a greater robability. But the MAP is too low as , which calculate the all recision. The F 1 score takes into account the recision and recall of the model. F 1 score that the recall and recision are equally imortant, but our model condsider that the recision is more imortant, so the F 1 score is less than other models. TABLE IV. AUC, MAP,NDCG AND F1 SCORE COMPARISON Model AUC MAP NDCG F1 Score ItemKNN UserKNN WRMF BPRMF UEIFM V. CONCLUSION AND FUTURE WORK Based on the integration of exlicit feedback and imlicit feedback data, a new model UEIFM is constructed and imlemented by the distributed comuting framework of Sark on cloud comuting technology, and then we get the To-N recommendation by ranking. We comared our model with ItemKNN [], UserKNN, WRMF [4] and BPRMF [7]. Whether in recision or recall, our model is relatively better. The value of AUC is better than WRMF. Due to the distributed comuting framework and cloud comuting environment, our model is more sufficiently scalable and efficiency. Because of the combination of multile exerimental datasets, the enhancement erformance in the exlicit and imlicit feedback data is reduced. In future work, we need to consider more about the imrovement of recision erformance, and increase the versatility of our model. ACKNOWLEDGMENT This aer was suorted by the National Natural Science Foundation of China (No: and 61089), the Science and Technology Plan of Hunan Province (No. 015GK3015) and The National High-Tech Research and Develoment Plan of China under Grant (No.01AA01A301-01). REFERENCES [1] MovieLens: htt://groulens.org/datasets/movielens/ [] MyMediaLite: htt://mymedialite.net/ [3] Koren, Yehuda, Robert Bell, and Chris Volinsky, "Matrix factorization techniques for recommender systems," Comuter 4.8 (009): [4] Hu, Yifan, Yehuda Koren, and Chris Volinsky, "Collaborative filtering for imlicit feedback datasets," 008 Eighth IEEE International Conference on Data Mining. Ieee, 008. [5] Zhou, Y., Wilkinson, D., Schreiber, R., & Pan, R, "Large-scale arallel collaborative filtering for the netflix rize," International Conference on Algorithmic Alications in Management. Sringer Berlin Heidelberg, 008. [6] Salakhutdinov, Ruslan, and Andriy Mnih, "Probabilistic matrix factorization," NIPS. Vol [7] Rendle, S., Freudenthaler, C., Gantner, Z., & Schmidt-Thieme, L, "BPR: Bayesian ersonalized ranking from imlicit feedback," Proceedings of the twenty-fifth conference on uncertainty in artificial intelligence. AUAI Press, 009. [8] Herlocker, J. L., Konstan, J. A., Terveen, L. G., & Riedl, J. T, "Evaluating collaborative filtering recommender systems," ACM Transactions on Information Systems (TOIS).1 (004): [9] Pilászy, István, Dávid Zibriczky, and Domonkos Tikk, "Fast alsbased matrix factorization for exlicit and imlicit feedback datasets," Proceedings of the fourth ACM conference on Recommender systems. ACM, 010. [10] Yu, H. F., Hsieh, C. J., Si, S., & Dhillon, I, "Scalable coordinate descent aroaches to arallel matrix factorization for recommender systems," 01 IEEE 1th International Conference on Data Mining. IEEE, 01. [11] Yu, H. F., Hsieh, C. J., Si, S., & Dhillon, I. S, "Parallel matrix factorization for recommender systems." Knowledge and Information Systems 41.3 (014): [1] Koren, Yehuda, "Factor in the neighbors: Scalable and accurate collaborative filtering." ACM Transactions on Knowledge Discovery from Data (TKDD) 4.1 (010): 1. [13] Parra, D., Karatzoglou, A., Amatriain, X., & Yavuz, I, "Imlicit feedback recommendation via imlicit-to-exlicit ordinal logistic regression maing." Proceedings of the CARS-011 (011). [14] Nathan N. Liu, Evan W. Xiang, Min Zhao, and Qiang Yang, "Unifying exlicit and imlicit feedback for collaborative filtering," Proceedings of the 19th ACM international conference on Information and knowledge management. ACM, 010. [15] Niu, F., Recht, B., Ré, C., & Wright, S. J. H, "A lock-free aroach to arallelizing stochastic gradient descent. arxiv rerint," ArXiv 1106 (011). [16] Gemulla, R., Nijkam, E., Haas, P. J., & Sismanis, Y, "Large-scale matrix factorization with distributed stochastic gradient descent," Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining. ACM, 011. [17] Recht, Benjamin, and Christoher Ré, "Parallel stochastic gradient algorithms for large-scale matrix comletion," Mathematical Programming Comutation 5. (013): [18] Zhuang, Y., Chin, W. S., Juan, Y. C., & Lin, C. J, "A fast arallel SGD for matrix factorization in shared memory systems," Proceedings of the 7th ACM conference on Recommender systems. ACM, 013. [19] Schelter, S., Boden, C., Schenck, M., Alexandrov, A., & Markl, V, "Distributed matrix factorization with mareduce using a series of broadcast-joins," Proceedings of the 7th ACM conference on Recommender systems. ACM, 013. [0] Koren, Yehuda, "Factorization meets the neighborhood: a multifaceted collaborative filtering model," Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining. ACM, 008. [1] Zaharia, M., Chowdhury, M., Franklin, M. J., Shenker, S., & Stoica, I, "Sark: cluster comuting with working sets," HotCloud10 (010): [] Linden, Greg, Brent Smith, and Jeremy York. "Amazon.com recommendations: Item-to-item collaborative filtering," IEEE Internet comuting 7.1 (003): [3] Docker: htts:// 39
PREDICTING LINKS IN LARGE COAUTHORSHIP NETWORKS
PREDICTING LINKS IN LARGE COAUTHORSHIP NETWORKS Kevin Miller, Vivian Lin, and Rui Zhang Grou ID: 5 1. INTRODUCTION The roblem we are trying to solve is redicting future links or recovering missing links
More informationEfficient Parallel Hierarchical Clustering
Efficient Parallel Hierarchical Clustering Manoranjan Dash 1,SimonaPetrutiu, and Peter Scheuermann 1 Deartment of Information Systems, School of Comuter Engineering, Nanyang Technological University, Singaore
More informationA GPU Heterogeneous Cluster Scheduling Model for Preventing Temperature Heat Island
A GPU Heterogeneous Cluster Scheduling Model for Preventing Temerature Heat Island Yun-Peng CAO 1,2,a and Hai-Feng WANG 1,2 1 School of Information Science and Engineering, Linyi University, Linyi Shandong,
More informationAn Efficient Coding Method for Coding Region-of-Interest Locations in AVS2
An Efficient Coding Method for Coding Region-of-Interest Locations in AVS2 Mingliang Chen 1, Weiyao Lin 1*, Xiaozhen Zheng 2 1 Deartment of Electronic Engineering, Shanghai Jiao Tong University, China
More informationAn empirical analysis of loopy belief propagation in three topologies: grids, small-world networks and random graphs
An emirical analysis of looy belief roagation in three toologies: grids, small-world networks and random grahs R. Santana, A. Mendiburu and J. A. Lozano Intelligent Systems Grou Deartment of Comuter Science
More informationInteractive Image Segmentation
Interactive Image Segmentation Fahim Mannan (260 266 294) Abstract This reort resents the roject work done based on Boykov and Jolly s interactive grah cuts based N-D image segmentation algorithm([1]).
More informationFace Recognition Using Legendre Moments
Face Recognition Using Legendre Moments Dr.S.Annadurai 1 A.Saradha Professor & Head of CSE & IT Research scholar in CSE Government College of Technology, Government College of Technology, Coimbatore, Tamilnadu,
More informationEfficient Processing of Top-k Dominating Queries on Multi-Dimensional Data
Efficient Processing of To-k Dominating Queries on Multi-Dimensional Data Man Lung Yiu Deartment of Comuter Science Aalborg University DK-922 Aalborg, Denmark mly@cs.aau.dk Nikos Mamoulis Deartment of
More informationFigure 8.1: Home age taken from the examle health education site (htt:// Setember 14, 2001). 201
200 Chater 8 Alying the Web Interface Profiles: Examle Web Site Assessment 8.1 Introduction This chater describes the use of the rofiles develoed in Chater 6 to assess and imrove the quality of an examle
More informationIMS Network Deployment Cost Optimization Based on Flow-Based Traffic Model
IMS Network Deloyment Cost Otimization Based on Flow-Based Traffic Model Jie Xiao, Changcheng Huang and James Yan Deartment of Systems and Comuter Engineering, Carleton University, Ottawa, Canada {jiexiao,
More informationThe Research on Curling Track Empty Value Fill Algorithm Based on Similar Forecast
Research Journal of Alied Sciences, Engineering and Technology 6(8): 1472-1478, 2013 ISSN: 2040-7459; e-issn: 2040-7467 Maxwell Scientific Organization, 2013 Submitted: October 31, 2012 Acceted: January
More informationStereo Disparity Estimation in Moment Space
Stereo Disarity Estimation in oment Sace Angeline Pang Faculty of Information Technology, ultimedia University, 63 Cyberjaya, alaysia. angeline.ang@mmu.edu.my R. ukundan Deartment of Comuter Science, University
More informationA CLASS OF STRUCTURED LDPC CODES WITH LARGE GIRTH
A CLASS OF STRUCTURED LDPC CODES WITH LARGE GIRTH Jin Lu, José M. F. Moura, and Urs Niesen Deartment of Electrical and Comuter Engineering Carnegie Mellon University, Pittsburgh, PA 15213 jinlu, moura@ece.cmu.edu
More informationA label distance maximum-based classifier for multi-label learning
Bio-Medical Materials and Engineering 26 (2015) S1969 S1976 DOI 10.3233/BME-151500 IOS ress S1969 A label distance maximum-based classifier for multi-label learning Xiaoli Liu a,b, Hang Bao a, Dazhe Zhao
More informationLearning Motion Patterns in Crowded Scenes Using Motion Flow Field
Learning Motion Patterns in Crowded Scenes Using Motion Flow Field Min Hu, Saad Ali and Mubarak Shah Comuter Vision Lab, University of Central Florida {mhu,sali,shah}@eecs.ucf.edu Abstract Learning tyical
More informationExtracting Optimal Paths from Roadmaps for Motion Planning
Extracting Otimal Paths from Roadmas for Motion Planning Jinsuck Kim Roger A. Pearce Nancy M. Amato Deartment of Comuter Science Texas A&M University College Station, TX 843 jinsuckk,ra231,amato @cs.tamu.edu
More informationAPPLICATION OF PARTICLE FILTERS TO MAP-MATCHING ALGORITHM
APPLICATION OF PARTICLE FILTERS TO MAP-MATCHING ALGORITHM Pavel Davidson 1, Jussi Collin 2, and Jarmo Taala 3 Deartment of Comuter Systems, Tamere University of Technology, Finland e-mail: avel.davidson@tut.fi
More informationPivot Selection for Dimension Reduction Using Annealing by Increasing Resampling *
ivot Selection for Dimension Reduction Using Annealing by Increasing Resamling * Yasunobu Imamura 1, Naoya Higuchi 1, Tetsuji Kuboyama 2, Kouichi Hirata 1 and Takeshi Shinohara 1 1 Kyushu Institute of
More informationA New and Efficient Algorithm-Based Fault Tolerance Scheme for A Million Way Parallelism
A New and Efficient Algorithm-Based Fault Tolerance Scheme for A Million Way Parallelism Erlin Yao, Mingyu Chen, Rui Wang, Wenli Zhang, Guangming Tan Key Laboratory of Comuter System and Architecture Institute
More informationOMNI: An Efficient Overlay Multicast. Infrastructure for Real-time Applications
OMNI: An Efficient Overlay Multicast Infrastructure for Real-time Alications Suman Banerjee, Christoher Kommareddy, Koushik Kar, Bobby Bhattacharjee, Samir Khuller Abstract We consider an overlay architecture
More informationImproved Image Super-Resolution by Support Vector Regression
Proceedings of International Joint Conference on Neural Networks, San Jose, California, USA, July 3 August 5, 0 Imroved Image Suer-Resolution by Suort Vector Regression Le An and Bir Bhanu Abstract Suort
More informationSensitivity Analysis for an Optimal Routing Policy in an Ad Hoc Wireless Network
1 Sensitivity Analysis for an Otimal Routing Policy in an Ad Hoc Wireless Network Tara Javidi and Demosthenis Teneketzis Deartment of Electrical Engineering and Comuter Science University of Michigan Ann
More informationImproved heuristics for the single machine scheduling problem with linear early and quadratic tardy penalties
Imroved heuristics for the single machine scheduling roblem with linear early and quadratic tardy enalties Jorge M. S. Valente* LIAAD INESC Porto LA, Faculdade de Economia, Universidade do Porto Postal
More informationAutonomic Physical Database Design - From Indexing to Multidimensional Clustering
Autonomic Physical Database Design - From Indexing to Multidimensional Clustering Stehan Baumann, Kai-Uwe Sattler Databases and Information Systems Grou Technische Universität Ilmenau, Ilmenau, Germany
More informationAUTOMATIC GENERATION OF HIGH THROUGHPUT ENERGY EFFICIENT STREAMING ARCHITECTURES FOR ARBITRARY FIXED PERMUTATIONS. Ren Chen and Viktor K.
inuts er clock cycle Streaming ermutation oututs er clock cycle AUTOMATIC GENERATION OF HIGH THROUGHPUT ENERGY EFFICIENT STREAMING ARCHITECTURES FOR ARBITRARY FIXED PERMUTATIONS Ren Chen and Viktor K.
More informationShuigeng Zhou. May 18, 2016 School of Computer Science Fudan University
Query Processing Shuigeng Zhou May 18, 2016 School of Comuter Science Fudan University Overview Outline Measures of Query Cost Selection Oeration Sorting Join Oeration Other Oerations Evaluation of Exressions
More informationLecture 18. Today, we will discuss developing algorithms for a basic model for parallel computing the Parallel Random Access Machine (PRAM) model.
U.C. Berkeley CS273: Parallel and Distributed Theory Lecture 18 Professor Satish Rao Lecturer: Satish Rao Last revised Scribe so far: Satish Rao (following revious lecture notes quite closely. Lecture
More informationPatterned Wafer Segmentation
atterned Wafer Segmentation ierrick Bourgeat ab, Fabrice Meriaudeau b, Kenneth W. Tobin a, atrick Gorria b a Oak Ridge National Laboratory,.O.Box 2008, Oak Ridge, TN 37831-6011, USA b Le2i Laboratory Univ.of
More informationA Model-Adaptable MOSFET Parameter Extraction System
A Model-Adatable MOSFET Parameter Extraction System Masaki Kondo Hidetoshi Onodera Keikichi Tamaru Deartment of Electronics Faculty of Engineering, Kyoto University Kyoto 66-1, JAPAN Tel: +81-7-73-313
More informationPrivacy Preserving Moving KNN Queries
Privacy Preserving Moving KNN Queries arxiv:4.76v [cs.db] 4 Ar Tanzima Hashem Lars Kulik Rui Zhang National ICT Australia, Deartment of Comuter Science and Software Engineering University of Melbourne,
More information10. Parallel Methods for Data Sorting
10. Parallel Methods for Data Sorting 10. Parallel Methods for Data Sorting... 1 10.1. Parallelizing Princiles... 10.. Scaling Parallel Comutations... 10.3. Bubble Sort...3 10.3.1. Sequential Algorithm...3
More informationTop-N Recommendations from Implicit Feedback Leveraging Linked Open Data
Top-N Recommendations from Implicit Feedback Leveraging Linked Open Data Vito Claudio Ostuni, Tommaso Di Noia, Roberto Mirizzi, Eugenio Di Sciascio Polytechnic University of Bari, Italy {ostuni,mirizzi}@deemail.poliba.it,
More informationA Morphological LiDAR Points Cloud Filtering Method based on GPGPU
A Morhological LiDAR Points Cloud Filtering Method based on GPGPU Shuo Li 1, Hui Wang 1, Qiuhe Ma 1 and Xuan Zha 2 1 Zhengzhou Institute of Surveying & Maing, No.66, Longhai Middle Road, Zhengzhou, China
More informationRecommender Systems Based on Doubly Structural Network
Proceedings of the 8th nternational Conference on nnovation & Management 975 Recommender Systems Based on Doubly Structural Network Na Chang, Takao Terano Deartment of Comutational ntelligence and Systems
More informationA Novel Iris Segmentation Method for Hand-Held Capture Device
A Novel Iris Segmentation Method for Hand-Held Cature Device XiaoFu He and PengFei Shi Institute of Image Processing and Pattern Recognition, Shanghai Jiao Tong University, Shanghai 200030, China {xfhe,
More informationA Parallel Algorithm for Constructing Obstacle-Avoiding Rectilinear Steiner Minimal Trees on Multi-Core Systems
A Parallel Algorithm for Constructing Obstacle-Avoiding Rectilinear Steiner Minimal Trees on Multi-Core Systems Cheng-Yuan Chang and I-Lun Tseng Deartment of Comuter Science and Engineering Yuan Ze University,
More informationSwift Template Matching Based on Equivalent Histogram
Swift emlate Matching ased on Equivalent istogram Wangsheng Yu, Xiaohua ian, Zhiqiang ou * elecommunications Engineering Institute Air Force Engineering University Xi an, PR China *corresonding author:
More informationRandomized algorithms: Two examples and Yao s Minimax Principle
Randomized algorithms: Two examles and Yao s Minimax Princile Maximum Satisfiability Consider the roblem Maximum Satisfiability (MAX-SAT). Bring your knowledge u-to-date on the Satisfiability roblem. Maximum
More informationParametric Optimization in WEDM of WC-Co Composite by Neuro-Genetic Technique
Parametric Otimization in WEDM of WC-Co Comosite by Neuro-Genetic Technique P. Saha*, P. Saha, and S. K. Pal Abstract The resent work does a multi-objective otimization in wire electro-discharge machining
More informationComplexity Issues on Designing Tridiagonal Solvers on 2-Dimensional Mesh Interconnection Networks
Journal of Comuting and Information Technology - CIT 8, 2000, 1, 1 12 1 Comlexity Issues on Designing Tridiagonal Solvers on 2-Dimensional Mesh Interconnection Networks Eunice E. Santos Deartment of Electrical
More informationAUTOMATIC EXTRACTION OF BUILDING OUTLINE FROM HIGH RESOLUTION AERIAL IMAGERY
AUTOMATIC EXTRACTION OF BUILDING OUTLINE FROM HIGH RESOLUTION AERIAL IMAGERY Yandong Wang EagleView Technology Cor. 5 Methodist Hill Dr., Rochester, NY 1463, the United States yandong.wang@ictometry.com
More informationDistributed Estimation from Relative Measurements in Sensor Networks
Distributed Estimation from Relative Measurements in Sensor Networks #Prabir Barooah and João P. Hesanha Abstract We consider the roblem of estimating vectorvalued variables from noisy relative measurements.
More informationA DEA-bases Approach for Multi-objective Design of Attribute Acceptance Sampling Plans
Available online at htt://ijdea.srbiau.ac.ir Int. J. Data Enveloment Analysis (ISSN 2345-458X) Vol.5, No.2, Year 2017 Article ID IJDEA-00422, 12 ages Research Article International Journal of Data Enveloment
More informationEquality-Based Translation Validator for LLVM
Equality-Based Translation Validator for LLVM Michael Ste, Ross Tate, and Sorin Lerner University of California, San Diego {mste,rtate,lerner@cs.ucsd.edu Abstract. We udated our Peggy tool, reviously resented
More informationJ. Parallel Distrib. Comput.
J. Parallel Distrib. Comut. 71 (2011) 288 301 Contents lists available at ScienceDirect J. Parallel Distrib. Comut. journal homeage: www.elsevier.com/locate/jdc Quality of security adatation in arallel
More informationA Fast Image Restoration Method Based on an Improved Criminisi Algorithm
A Fast Image Restoration Method Based on an Imroved Algorithm Yue Chi1, Ning He2*, Qi Zhang1 Beijing Key Laboratory of Information Services Engineering, Beijing Union University, Beijing 100101, China.
More informationEfficient Sequence Generator Mining and its Application in Classification
Efficient Sequence Generator Mining and its Alication in Classification Chuancong Gao, Jianyong Wang 2, Yukai He 3 and Lizhu Zhou 4 Tsinghua University, Beijing 0084, China {gaocc07, heyk05 3 }@mails.tsinghua.edu.cn,
More informationWeighted Page Rank Algorithm based on In-Out Weight of Webpages
Indian Journal of Science and Technology, Vol 8(34), DOI: 10.17485/ijst/2015/v8i34/86120, December 2015 ISSN (Print) : 0974-6846 ISSN (Online) : 0974-5645 eighted Page Rank Algorithm based on In-Out eight
More informationUsing Rational Numbers and Parallel Computing to Efficiently Avoid Round-off Errors on Map Simplification
Using Rational Numbers and Parallel Comuting to Efficiently Avoid Round-off Errors on Ma Simlification Maurício G. Grui 1, Salles V. G. de Magalhães 1,2, Marcus V. A. Andrade 1, W. Randolh Franklin 2,
More informationWeb Personalization and Recommendation: A Review
Web Personalization and Recommendation: A Review 1 Nayab Zya, 2 Mohammad Suaib 3 M S Husain 1, 2, 3 Integral University, Lucknow Uttar Pradesh, India Abstract: During the ast few years the World Wide Web
More informationSpace-efficient Region Filling in Raster Graphics
"The Visual Comuter: An International Journal of Comuter Grahics" (submitted July 13, 1992; revised December 7, 1992; acceted in Aril 16, 1993) Sace-efficient Region Filling in Raster Grahics Dominik Henrich
More informationWavelet Based Statistical Adapted Local Binary Patterns for Recognizing Avatar Faces
Wavelet Based Statistical Adated Local Binary atterns for Recognizing Avatar Faces Abdallah A. Mohamed 1, 2 and Roman V. Yamolskiy 1 1 Comuter Engineering and Comuter Science, University of Louisville,
More informationA Study of Protocols for Low-Latency Video Transport over the Internet
A Study of Protocols for Low-Latency Video Transort over the Internet Ciro A. Noronha, Ph.D. Cobalt Digital Santa Clara, CA ciro.noronha@cobaltdigital.com Juliana W. Noronha University of California, Davis
More informationLeak Detection Modeling and Simulation for Oil Pipeline with Artificial Intelligence Method
ITB J. Eng. Sci. Vol. 39 B, No. 1, 007, 1-19 1 Leak Detection Modeling and Simulation for Oil Pieline with Artificial Intelligence Method Pudjo Sukarno 1, Kuntjoro Adji Sidarto, Amoranto Trisnobudi 3,
More informationBayesian Oil Spill Segmentation of SAR Images via Graph Cuts 1
Bayesian Oil Sill Segmentation of SAR Images via Grah Cuts 1 Sónia Pelizzari and José M. Bioucas-Dias Instituto de Telecomunicações, I.S.T., TULisbon,Lisboa, Portugal sonia@lx.it.t, bioucas@lx.it.t Abstract.
More informationFace Recognition Based on Wavelet Transform and Adaptive Local Binary Pattern
Face Recognition Based on Wavelet Transform and Adative Local Binary Pattern Abdallah Mohamed 1,2, and Roman Yamolskiy 1 1 Comuter Engineering and Comuter Science, University of Louisville, Louisville,
More informationA Symmetric FHE Scheme Based on Linear Algebra
A Symmetric FHE Scheme Based on Linear Algebra Iti Sharma University College of Engineering, Comuter Science Deartment. itisharma.uce@gmail.com Abstract FHE is considered to be Holy Grail of cloud comuting.
More informationGrouping of Patches in Progressive Radiosity
Grouing of Patches in Progressive Radiosity Arjan J.F. Kok * Abstract The radiosity method can be imroved by (adatively) grouing small neighboring atches into grous. Comutations normally done for searate
More informationA NN Image Classification Method Driven by the Mixed Fitness Function
Comuter and Information Science November, 2009 A NN Image Classification Method Driven by the Mixed Fitness Function Shan Gai, Peng Liu, Jiafeng Liu & Xianglong Tang School of Comuter Science and Technology,
More informationIntroduction to Parallel Algorithms
CS 1762 Fall, 2011 1 Introduction to Parallel Algorithms Introduction to Parallel Algorithms ECE 1762 Algorithms and Data Structures Fall Semester, 2011 1 Preliminaries Since the early 1990s, there has
More informationA BICRITERION STEINER TREE PROBLEM ON GRAPH. Mirko VUJO[EVI], Milan STANOJEVI] 1. INTRODUCTION
Yugoslav Journal of Oerations Research (00), umber, 5- A BICRITERIO STEIER TREE PROBLEM O GRAPH Mirko VUJO[EVI], Milan STAOJEVI] Laboratory for Oerational Research, Faculty of Organizational Sciences University
More informationEnsemble Learning Based on Parametric Triangular Norms
Send Orders for Rerints to rerints@benthamscience.ae The Oen Automation and Control Systems Journal, 2014, 6, 997-1003 997 Ensemble Learning Based on Parametric Triangular Norms Oen Access Pengtao Jia
More informationUse of Multivariate Statistical Analysis in the Modelling of Chromatographic Processes
Use of Multivariate Statistical Analysis in the Modelling of Chromatograhic Processes Simon Edwards-Parton 1, Nigel itchener-hooker 1, Nina hornhill 2, Daniel Bracewell 1, John Lidell 3 Abstract his aer
More informationVisualization, Estimation and User-Modeling for Interactive Browsing of Image Libraries
Visualization, Estimation and User-Modeling for Interactive Browsing of Image Libraries Qi Tian, Baback Moghaddam 2 and Thomas S. Huang Beckman Institute, University of Illinois, Urbana-Chamaign, IL 680,
More informationRecord Route IP Traceback: Combating DoS Attacks and the Variants
Record Route IP Traceback: Combating DoS Attacks and the Variants Abdullah Yasin Nur, Mehmet Engin Tozal University of Louisiana at Lafayette, Lafayette, LA, US ayasinnur@louisiana.edu, metozal@louisiana.edu
More informationAn Efficient Video Program Delivery algorithm in Tree Networks*
3rd International Symosium on Parallel Architectures, Algorithms and Programming An Efficient Video Program Delivery algorithm in Tree Networks* Fenghang Yin 1 Hong Shen 1,2,** 1 Deartment of Comuter Science,
More informationAn Efficient VLSI Architecture for Adaptive Rank Order Filter for Image Noise Removal
International Journal of Information and Electronics Engineering, Vol. 1, No. 1, July 011 An Efficient VLSI Architecture for Adative Rank Order Filter for Image Noise Removal M. C Hanumantharaju, M. Ravishankar,
More informationMultidimensional Service Weight Sequence Mining based on Cloud Service Utilization in Jyaguchi
Proceedings of the International MultiConference of Engineers and Comuter Scientists 2013 Vol I, Multidimensional Service Weight Sequence Mining based on Cloud Service Utilization in Jyaguchi Shree Krishna
More informationWeight Co-occurrence based Integrated Color and Intensity Matrix for CBIR
Weight Co-occurrence based Integrated Color and Intensity Matrix for CBIR Megha Agarwal, 2R.P. Maheshwari Indian Institute of Technology Roorkee 247667, Uttarakhand, India meghagarwal29@gmail.com, 2rmaheshwari@gmail.com
More informationSEARCH ENGINE MANAGEMENT
e-issn 2455 1392 Volume 2 Issue 5, May 2016. 254 259 Scientific Journal Imact Factor : 3.468 htt://www.ijcter.com SEARCH ENGINE MANAGEMENT Abhinav Sinha Kalinga Institute of Industrial Technology, Bhubaneswar,
More informationExperiments on Patent Retrieval at NTCIR-4 Workshop
Working Notes of NTCIR-4, Tokyo, 2-4 June 2004 Exeriments on Patent Retrieval at NTCIR-4 Worksho Hironori Takeuchi Λ Naohiko Uramoto Λy Koichi Takeda Λ Λ Tokyo Research Laboratory, IBM Research y National
More informationAuto-Tuning Distributed-Memory 3-Dimensional Fast Fourier Transforms on the Cray XT4
Auto-Tuning Distributed-Memory 3-Dimensional Fast Fourier Transforms on the Cray XT4 M. Gajbe a A. Canning, b L-W. Wang, b J. Shalf, b H. Wasserman, b and R. Vuduc, a a Georgia Institute of Technology,
More informationarxiv: v1 [cs.mm] 18 Jan 2016
Lossless Intra Coding in with 3-ta Filters Saeed R. Alvar a, Fatih Kamisli a a Deartment of Electrical and Electronics Engineering, Middle East Technical University, Turkey arxiv:1601.04473v1 [cs.mm] 18
More informationAN ANALYTICAL MODEL DESCRIBING THE RELATIONSHIPS BETWEEN LOGIC ARCHITECTURE AND FPGA DENSITY
AN ANALYTICAL MODEL DESCRIBING THE RELATIONSHIPS BETWEEN LOGIC ARCHITECTURE AND FPGA DENSITY Andrew Lam 1, Steven J.E. Wilton 1, Phili Leong 2, Wayne Luk 3 1 Elec. and Com. Engineering 2 Comuter Science
More informationModel-Based Annotation of Online Handwritten Datasets
Model-Based Annotation of Online Handwritten Datasets Anand Kumar, A. Balasubramanian, Anoo Namboodiri and C.V. Jawahar Center for Visual Information Technology, International Institute of Information
More informationS16-02, URL:
Self Introduction A/Prof ay Seng Chuan el: Email: scitaysc@nus.edu.sg Office: S-0, Dean s s Office at Level URL: htt://www.hysics.nus.edu.sg/~hytaysc I was a rogrammer from to. I have been working in NUS
More informationSensitivity of multi-product two-stage economic lotsizing models and their dependency on change-over and product cost ratio s
Sensitivity two stage EOQ model 1 Sensitivity of multi-roduct two-stage economic lotsizing models and their deendency on change-over and roduct cost ratio s Frank Van den broecke, El-Houssaine Aghezzaf,
More informationCENTRAL AND PARALLEL PROJECTIONS OF REGULAR SURFACES: GEOMETRIC CONSTRUCTIONS USING 3D MODELING SOFTWARE
CENTRAL AND PARALLEL PROJECTIONS OF REGULAR SURFACES: GEOMETRIC CONSTRUCTIONS USING 3D MODELING SOFTWARE Petra Surynková Charles University in Prague, Faculty of Mathematics and Physics, Sokolovská 83,
More informationSimultaneous Tracking of Multiple Objects Using Fast Level Set-Like Algorithm
Simultaneous Tracking of Multile Objects Using Fast Level Set-Like Algorithm Martin Maška, Pavel Matula, and Michal Kozubek Centre for Biomedical Image Analysis, Faculty of Informatics Masaryk University,
More informationEarthenware Reconstruction Based on the Shape Similarity among Potsherds
Original Paer Forma, 16, 77 90, 2001 Earthenware Reconstruction Based on the Shae Similarity among Potsherds Masayoshi KANOH 1, Shohei KATO 2 and Hidenori ITOH 1 1 Nagoya Institute of Technology, Gokiso-cho,
More informationContinuous Visible k Nearest Neighbor Query on Moving Objects
Continuous Visible k Nearest Neighbor Query on Moving Objects Yaniu Wang a, Rui Zhang b, Chuanfei Xu a, Jianzhong Qi b, Yu Gu a, Ge Yu a, a Deartment of Comuter Software and Theory, Northeastern University,
More informationSPARSE SIGNAL REPRESENTATION FOR COMPLEX-VALUED IMAGING Sadegh Samadi 1, M üjdat Çetin 2, Mohammad Ali Masnadi-Shirazi 1
SPARSE SIGNAL REPRESENTATION FOR COMPLEX-VALUED IMAGING Sadegh Samadi 1, M üjdat Çetin, Mohammad Ali Masnadi-Shirazi 1 1. Shiraz University, Shiraz, Iran,. Sabanci University, Istanbul, Turkey ssamadi@shirazu.ac.ir,
More informationMatlab Virtual Reality Simulations for optimizations and rapid prototyping of flexible lines systems
Matlab Virtual Reality Simulations for otimizations and raid rototying of flexible lines systems VAMVU PETRE, BARBU CAMELIA, POP MARIA Deartment of Automation, Comuters, Electrical Engineering and Energetics
More informationFactoring Variations in Natural Images with Deep Gaussian Mixture Models
Factoring Variations in Natural Images with Dee Gaussian Mixture Models Aäron van den Oord, Benjamin Schrauwen Electronics and Information Systems deartment (ELIS), Ghent University {aaron.vandenoord,
More informationImproved Gaussian Mixture Probability Hypothesis Density for Tracking Closely Spaced Targets
INTL JOURNAL OF ELECTRONICS AND TELECOMMUNICATIONS, 7, VOL., NO., PP. 7-5 Manuscrit received October 7, 5; revised June, 7. DOI:.55/eletel-7- Imroved Gaussian Mixture Probability Hyothesis Density for
More informationComplexity analysis of matrix product on multicore architectures
Comlexity analysis of matrix roduct on multicore architectures Mathias Jacquelin, Loris Marchal and Yves Robert École Normale Suérieure de Lyon, France {Mathias.Jacquelin Loris.Marchal Yves.Robert}@ens-lyon.fr
More informationTOPP Probing of Network Links with Large Independent Latencies
TOPP Probing of Network Links with Large Indeendent Latencies M. Hosseinour, M. J. Tunnicliffe Faculty of Comuting, Information ystems and Mathematics, Kingston University, Kingston-on-Thames, urrey, KT1
More informationImplementation of Evolvable Fuzzy Hardware for Packet Scheduling Through Online Context Switching
Imlementation of Evolvable Fuzzy Hardware for Packet Scheduling Through Online Context Switching Ju Hui Li, eng Hiot Lim and Qi Cao School of EEE, Block S Nanyang Technological University Singaore 639798
More informationApplying the fuzzy preference relation to the software selection
Proceedings of the 007 WSEAS International Conference on Comuter Engineering and Alications, Gold Coast, Australia, January 17-19, 007 83 Alying the fuzzy reference relation to the software selection TIEN-CHIN
More informationReducing the Communication Costs of Graph Analysis by Read-only Replicas and Prioritized Execution
6 IEEE 8th International Conference on High Performance Comuting and Communications; IEEE th International Conference on Smart City; IEEE nd International Conference on Data Science and Systems Reducing
More informationLecture 2: Fixed-Radius Near Neighbors and Geometric Basics
structure arises in many alications of geometry. The dual structure, called a Delaunay triangulation also has many interesting roerties. Figure 3: Voronoi diagram and Delaunay triangulation. Search: Geometric
More informationChapter 8: Adaptive Networks
Chater : Adative Networks Introduction (.1) Architecture (.2) Backroagation for Feedforward Networks (.3) Jyh-Shing Roger Jang et al., Neuro-Fuzzy and Soft Comuting: A Comutational Aroach to Learning and
More informationParallel Construction of Multidimensional Binary Search Trees. Ibraheem Al-furaih, Srinivas Aluru, Sanjay Goil Sanjay Ranka
Parallel Construction of Multidimensional Binary Search Trees Ibraheem Al-furaih, Srinivas Aluru, Sanjay Goil Sanjay Ranka School of CIS and School of CISE Northeast Parallel Architectures Center Syracuse
More informationUsing Permuted States and Validated Simulation to Analyze Conflict Rates in Optimistic Replication
Using Permuted States and Validated Simulation to Analyze Conflict Rates in Otimistic Relication An-I A. Wang Comuter Science Deartment Florida State University Geoff H. Kuenning Comuter Science Deartment
More informationarxiv: v1 [cs.dc] 13 Nov 2018
Task Grah Transformations for Latency Tolerance arxiv:1811.05077v1 [cs.dc] 13 Nov 2018 Victor Eijkhout November 14, 2018 Abstract The Integrative Model for Parallelism (IMP) derives a task grah from a
More informationSemi-Supervised Learning Based Object Detection in Aerial Imagery
Semi-Suervised Learning Based Obect Detection in Aerial Imagery Jian Yao Zhongfei (Mark) Zhang Deartment of Comuter Science, State University of ew York at Binghamton, Y 13905, USA yao@binghamton.edu Zhongfei@cs.binghamton.edu
More informationLearning Robust Locality Preserving Projection via p-order Minimization
Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence Learning Robust Locality Preserving Projection via -Order Minimization Hua Wang, Feiing Nie, Heng Huang Deartment of Electrical
More informationSkip List Based Authenticated Data Structure in DAS Paradigm
009 Eighth International Conference on Grid and Cooerative Comuting Ski List Based Authenticated Data Structure in DAS Paradigm Jieing Wang,, Xiaoyong Du,. Key Laboratory of Data Engineering and Knowledge
More informationTo appear in IEEE TKDE Title: Efficient Skyline and Top-k Retrieval in Subspaces Keywords: Skyline, Top-k, Subspace, B-tree
To aear in IEEE TKDE Title: Efficient Skyline and To-k Retrieval in Subsaces Keywords: Skyline, To-k, Subsace, B-tree Contact Author: Yufei Tao (taoyf@cse.cuhk.edu.hk) Deartment of Comuter Science and
More informationOptimization of Collective Communication Operations in MPICH
To be ublished in the International Journal of High Performance Comuting Alications, 5. c Sage Publications. Otimization of Collective Communication Oerations in MPICH Rajeev Thakur Rolf Rabenseifner William
More information