Filters








20 Hits in 15.3 sec

TAGLETS: A System for Automatic Semi-Supervised Learning with Auxiliary Data [article]

Wasu Piriyakulkij and Cristina Menghini and Ross Briden and Nihal V. Nayak and Jeffrey Zhu and Elaheh Raisi and Stephen H. Bach
2021 arXiv   pre-print
The key components of TAGLETS are: (1) auxiliary data organized according to a knowledge graph, (2) modules encapsulating different methods for exploiting auxiliary and unlabeled data, and (3) a distillation  ...  We compare TAGLETS with state-of-the-art transfer learning and semi-supervised learning methods on four image classification tasks.  ...  Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon.  ... 
arXiv:2111.04798v2 fatcat:3x4jaa4h3nfobhd3jhlfbyotdi

Harnessing the Power of Ego Network Layers for Link Prediction in Online Social Networks [article]

Mustafa Toprak, Chiara Boldrini, Andrea Passarella, Marco Conti
2021 arXiv   pre-print
Being able to recommend links between users in online social networks is important for users to connect with like-minded individuals as well as for the platforms themselves and third parties leveraging  ...  Finally, we show that social-awareness can be used in place of using a classifier (which may be costly or impractical) for targeting a specific category of users.  ...  SEAL [5] is a graph neural network method that focuses on the local enclosing subgraphs of the target links (node pairs).  ... 
arXiv:2109.09190v1 fatcat:m2ock43efzd7vppwjkmqhkvz6y

New Frontiers for Scan Statistics: Network, Trajectory, and Text Data [chapter]

Renato M. Assunção, Roberto C. S. N. P. Souza, Marcos O. Prates
2020 Handbook of Scan Statistics  
Graph embedding is particularly popular because it is effective in a number of downstream machine learning tasks (e.g., link prediction, node classification, anomaly detection).  ...  moving average and link prediction."  ... 
doi:10.1007/978-1-4614-8414-1_47-1 fatcat:i7mruwigejc4ti2kufrvbd5qwm

Visualization of Heterogeneous Data

Mike Cammarano, Xin Dong, Bryan Chan, Jeff Klingner, Justin Talbot, Alon Halevey, Pat Hanrahan
2007 IEEE Transactions on Visualization and Computer Graphics  
We formulate this as a schema matching problem, finding appropriate paths in the data model for each required visualization attribute in a visualization template.  ...  Both the Resource Description Framework (RDF), used in the semantic web, and Maya Viz u-forms represent data as a graph of objects connected by labeled edges.  ...  ACKNOWLEDGEMENTS This material is based on work supported by SRI International/DARPA contract #55-000679 TO-5, Battelle Memorial Institute/US DOE under contract #15597-3, and Boeing under contract #SPO  ... 
doi:10.1109/tvcg.2007.70617 pmid:17968065 fatcat:zv5jgudnwvdm5br4qzdhef5o7q

Anomaly detection

Varun Chandola, Arindam Banerjee, Vipin Kumar
2009 ACM Computing Surveys  
In graph data, data instances are represented as vertices in a graph and are connected to other vertices with edges.  ...  To prune the search space for nearest neighbors, several techniques partition the attribute space into a hypergrid consisting of hypercubes of fixed sizes.  ... 
doi:10.1145/1541880.1541882 fatcat:bdimpqcz5bbyhjqatvpw45cmja

Social Networks and the Semantic Web

Peter Mika
2015 Proceedings of the 24th International Conference on World Wide Web - WWW '15 Companion  
For example, a clique in a graph is maximal complete subgraph of three or more nodes.  ...  For example, a k-plex is a maximal subgraph in which each node is adjacent to no fewer than g s − k nodes in the subgraph, where g s is the number of nodes in the subgraph.  ...  chances of a direct hit were forecast at 17%, with strike probability rising to 29% by August 28.  ... 
doi:10.1145/2740908.2742138 dblp:conf/www/Mika15 fatcat:ptyntdmnjrc4tcayyytykwlaba

Outlier Analysis [chapter]

Charu C. Aggarwal
2015 Data Mining  
Leman and Saket provided detailed feedback on several sections and chapters of this book. xvii  ...  In particular, I would like to acknowledge Leman Akoglu, Chih-Jen Lin, Saket Sathe, Jiliang Tang, and Suhang Wang.  ...  Therefore, if one uses a regression model to predict one of the features from the other features, and the average squared error is too large, then such a feature should be pruned.  ... 
doi:10.1007/978-3-319-14142-8_8 fatcat:6rqpe4bv3ba77hlyfpej2qotyy

Spatiotemporal enabled Content-based Image Retrieval

Mariana Belgiu, Martin Sudmanns, Tiede Dirk, Andrea Baraldi, Stefan Lang
2016 International Conference on GIScience Short Paper Proceedings  
In this paper, we propose a probabilistic method for estimation of the coverage of a sensor network based on 3D vector representation of the environment.  ...  For efficient deployment of sensors in a WSN the coverage estimation is a critical issue. Probabilistic methods are among the most accurate models proposed for sensor coverage estimation.  ...  Future Work Future research will extend our machine learning approach on Spark in the following four directions: (1) scanning a wider parameterization space and further optimizing search methods for the  ... 
doi:10.21433/b311729295dw fatcat:fulw4pw3kfh5nmfzcsy3pkisvm

Threat Detection using Information Flow Analysis on Kernel Audit Logs

Sadegh Momeni Milajerdi
2020
The results indicate that the proposed methods are capable of efficiently searching these audit logs and pinpoint threats in real-time with high precision and low false alarm rate.  ...  For cyber threat-hunting, we develop an inexact graph pattern matching approach to align a query graph extracted from cyber threat intelligence to a provenance graph constructed out of kernel audit logs  ...  We present several criteria for pruning these searches in order to produce a compact graph.  ... 
doi:10.25417/uic.13474962 fatcat:66suvjwjzvda5hsvbpfogiheyi

Building blocks for semantic data organization on the desktop

Niko Popitsch
2011 unpublished
Descriptions are accessible via a uniform Linked Data interface and can be linked with other descriptions and resources.  ...  These metadata are predominantly stored in embedded file headers, using a magnitude of mainly proprietary formats.  ...  Embedded links. The embedded link model [Dav98] is the most common way how links are modeled on the desktop and in the Web.  ... 
doi:10.25365/thesis.16182 fatcat:ovjhdn5srrbojcd72t7stgipu4

Comparing graphs

Nils Morten Kriege, Technische Universität Dortmund, Technische Universität Dortmund
2015
We present algorithms with improved bounds on running time for the subclass of outerplanar graphs.  ...  We study maximum common subgraph problems, which are based on classical graph-theoretical concepts for graph comparison and are NP-hard in the general case.  ...  In any case only an extremely small fraction of these compounds is of interest for a specific disease and the search for a potential new drug is therefore often compared to "searching a needle in a haystack  ... 
doi:10.17877/de290r-16358 fatcat:btrj2zdwlngxncyr7gpywat4be

Planning in design synthesis : abstraction-based LOOS (ABLOOS)

Robert F. Coyne, Carnegie Mellon University.Engineering Design Research Center.
2018
In these cases the structure or subgraph (or alternative subgraphs), for a GOB will raise issues such as: When does a GOB have a subgraph? Can it have a partial subgraph?  ...  The generalized notation for a problem reduction search is an AND/OR graph, whereas for a state space search it is an OR graph.  ... 
doi:10.1184/r1/6469145.v1 fatcat:hempkuax4bfevjfhbuzo6anq2y

Exploiting metadata for context creation and ranking on the desktop [article]

Stefania Costache, University, My
2019
They also demonstrated by their web search experiments that the time that a user spends on one document is the most relevant in order to predict a certain satisfaction, as in our case it will be able to  ...  The list of relevant resources is expanded by searching in the RDF graph for resources linked via a property which indicates the similarity (see [MPC + 10] for more details).  ...  For the desktop ranking using time links we intend to explore content based heuristics to provide us with further additional links between similar desktop documents.  ... 
doi:10.15488/7375 fatcat:zlfza73345dcbfw6hvln5oonna

Evolutionary reverse engineering of gene networks

Daniel Marbach
2009
Here, we advocate a new approach for embedding prior knowledge in a reverse engineering method.  ...  The biomimetic approach circumvents this problem by embedding prior knowledge at a more fundamental level, namely by using a search method that bears close similarity with the evolutionary process (the  ... 
doi:10.5075/epfl-thesis-4503 fatcat:hr73bykfprhilka3itsnynn5wu

Social Networks and the Semantic Web

P. Mika
IEEE/WIC/ACM International Conference on Web Intelligence (WI'04)  
For example, a clique in a graph is maximal complete subgraph of three or more nodes.  ...  For example, a k-plex is a maximal subgraph in which each node is adjacent to no fewer than g s − k nodes in the subgraph, where g s is the number of nodes in the subgraph.  ...  chances of a direct hit were forecast at 17%, with strike probability rising to 29% by August 28.  ... 
doi:10.1109/wi.2004.10039 fatcat:ymtxnlnpvnakpgiecoebdn64ie
« Previous Showing results 1 — 15 out of 20 results