Filters








13,670 Hits in 6.1 sec

Unsupervised Multi-Level Non-Negative Matrix Factorization Model: Binary Data Case

Qingquan Sun, Peng Wu, Yeqing Wu, Mengcheng Guo, Jiang Lu
2012 Journal of Information Security  
In our unsupervised multi-level model, we construct a three-level data structure for non-negative matrix factorization algorithm.  ...  In this paper, we propose an unsupervised multi-level non-negative matrix factorization model to extract the hidden data structure and seek the rank of base matrix.  ...  Unsupervised Multi-Level Non-Negative Matrix Factorization Model In our unsupervised multi-level NMF model, we introduce a hyper-prior level.  ... 
doi:10.4236/jis.2012.34031 fatcat:dbv25rw5obgr7ktdwzhprvxahm

Generating Cyber Threat Intelligence to Discover Potential Security Threats Using Classification and Topic Modeling [article]

Md Imran Hossen, Ashraful Islam, Farzana Anowar, Eshtiak Ahmed, Mohammad Masudur Rahman
2021 arXiv   pre-print
For this purpose, we leverage two topic modeling algorithms namely Latent Dirichlet Allocation (LDA) and Non-negative Matrix Factorization (NMF).  ...  To this end, we collect data from a real hacker forum and constructed two datasets: a binary dataset and a multi-class dataset.  ...  For topic modeling we utilize two popular algorithms: a) Latent Dirichlet Allocation (LDA) and b) Non-negative Matrix Factorization (NMF).  ... 
arXiv:2108.06862v2 fatcat:4wrmcilmmnenzoxftnkcdukvg4

Unsupervised and Online Update of Boosted Temporal Models: The UAL2Boost

Pedro Canotilho Ribeiro, Plinio Moreno, Jose Santos-Victor
2010 2010 Ninth International Conference on Machine Learning and Applications  
update that eliminates the necessity of labeled data in order to adapt the classifier and (iii) a multi-class adaptation method.  ...  models, thus improving the output of the models learned off-line on new video sequences, in a recursive and continuous way.  ...  In difference to the binary case that uses the two available classes, in multi-class problems we have c weights, one from each binary problem.  ... 
doi:10.1109/icmla.2010.143 dblp:conf/icmla/RibeiroMS10 fatcat:3qpurehbfvhxzja35j6fvnanu4

Machine Learning on Graphs: A Model and Comprehensive Taxonomy [article]

Ines Chami, Sami Abu-El-Haija, Bryan Perozzi, Christopher Ré, Kevin Murphy
2022 arXiv   pre-print
Here, we aim to bridge the gap between graph neural networks, network embedding and graph regularization models.  ...  There has been a surge of recent interest in learning representations for graph-structured data.  ...  The adjacency matrix is binary for unweighted graph, A ∈ {0, 1} |V |×|V | , and non-binary for weighted graphs W ∈ R |V |×|V | .  ... 
arXiv:2005.03675v3 fatcat:6eoicgprdvfbze732nsmpaumqe

A Novel Multi label Text Classification Model using Semi supervised learning

Shweta C Dharmadhikari
2012 International Journal of Data Mining & Knowledge Management Process  
Through this paper a classification model for ATC in multi-label domain is discussed.  ...  We are proposing a new multi label text classification model for assigning more relevant set of categories to every input text document.  ...  In 2006 Liu, Jin and Yan proposed Multi-label classification approach based on constrained non negative matrix factorization [8] .  ... 
doi:10.5121/ijdkp.2012.2402 fatcat:hhn3aa63zjdovnwgbvy25v236a

Solving Non-identifiable Latent Feature Models [article]

Ryota Suzuki, Shingo Takahashi, Murtuza Petladwala, Shigeru Kohmoto
2018 arXiv   pre-print
Latent feature models (LFM)s are widely employed for extracting latent structures of data.  ...  In this paper, a necessary and sufficient condition for non-identifiability is shown.  ...  For a context of non-negative matrix factorization (NMF), where Z = R N ×K + and W = R K×D + , Laurberg, et al.  ... 
arXiv:1809.03776v2 fatcat:x3hkvgprnva4feha37q45mqexq

MedLDA: A General Framework of Maximum Margin Supervised Topic Models [article]

Jun Zhu, Amr Ahmed, Eric P. Xing
2009 arXiv   pre-print
movie review and 20 Newsgroups data sets.  ...  The general principle of MedLDA can be applied to perform joint max-margin learning and maximum likelihood estimation for arbitrary topic models, directed or undirected, and supervised or unsupervised,  ...  Classification We perform binary and multi-class classification on the 20 Newsgroup data set.  ... 
arXiv:0912.5507v1 fatcat:xcv25naanrfwzl42iea3mm552q

Self-Adaptive Hierarchical Sentence Model [article]

Han Zhao, Zhengdong Lu, Pascal Poupart
2015 arXiv   pre-print
on 5 benchmark data sets.  ...  As an effort towards this goal we propose a self-adaptive hierarchical sentence model (AdaSent).  ...  This factorization of the wordembedding matrix also helps to reduce the effective number of parameters in our model when d D.  ... 
arXiv:1504.05070v2 fatcat:ozp6au5bzrbgrkeslo4pu4snly

Topic Modeling: A Comprehensive Review

Pooja Kherwa, Poonam Bansal
2018 EAI Endorsed Transactions on Scalable Information Systems  
Quantitative evaluation of topic modeling techniques is also presented in detail for better understanding the concept of topic modeling.  ...  After analysing approximately 300 research articles on topic modeling, a comprehensive survey on topic modelling has been presented in this paper.  ...  S. (1999) Learning the parts of objects by non-negative matrix factorization.  ... 
doi:10.4108/eai.13-7-2018.159623 fatcat:lu6al57vp5aahbytyejhqrlzry

Beyond Tracking: Modelling Activity and Understanding Behaviour

Tao Xiang, Shaogang Gong
2006 International Journal of Computer Vision  
networks including a Multi-Observation Hidden Markov Model (MOHMM), a Parallel Hidden Markov Model (PaHMM) and a Coupled Hidden Markov Model (CHMM).  ...  Dynamic Probabilistic Networks (DPNs) are formulated for modelling the temporal and causal correlations among discrete events for robust and holistic scene-level behaviour interpretation.  ...  Acknowledgements We shall thank Huw Farmer and Mark Ealing at BAA for providing us with the aircraft cargo activity data under the DTI/EPSRC MI LINK project ICONS. Notes  ... 
doi:10.1007/s11263-006-4329-6 fatcat:jfg4mig2ureoxb5kbcocfvr5xm

Model recommendation: Generating object detectors from few samples

Yu-Xiong Wang, Martial Hebert
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
In this paper, we explore an approach to generating detectors that is radically different from the conventional way of learning a detector from a large corpus of annotated positive and negative data samples  ...  Given a new target task, we evaluate a subset of the models on few samples from the new task and we use the matrix of modelstasks ratings to predict the performance of all the models in the library on  ...  Formally, given the rating matrix R ∈ R n×m ≥0 with non-negative elements, NMF seeks to decompose R into a non-negative n × d basis matrix U (model factor) and a non-negative d × m coefficient matrix V  ... 
doi:10.1109/cvpr.2015.7298770 dblp:conf/cvpr/WangH15 fatcat:auejhxzbibb7zb5o3fmjfay2de

A Survey on Machine Learning in COVID-19 Diagnosis

Xing Guo, Yu-Dong Zhang, Siyuan Lu, Zhihai Lu
2022 CMES - Computer Modeling in Engineering & Sciences  
For example, the highly unbalanced dataset, the difficulty of collecting labeled data, and the poor quality of the data.  ...  Then, we review seven methods in detail: transfer learning, ensemble learning, unsupervised learning and semi-supervised learning, convolutional neural networks, graph neural networks, explainable deep  ...  Gray level co-occurrence matrix, local binary gray level co-occurrence matrix, gray level run length matrix, as well as segmentation-based fractal texture analysis and synthetic minority over-sampling  ... 
doi:10.32604/cmes.2022.017679 fatcat:hre5zxtekvaevleu335faqilwu

Multi-Object Classification and Unsupervised Scene Understanding Using Deep Learning Features and Latent Tree Probabilistic Models [article]

Tejaswi Nimmagadda, Anima Anandkumar
2015 arXiv   pre-print
Thus, we present a unified framework for multi-object classification and unsupervised scene understanding.  ...  COCO, consisting of non-iconic images.  ...  In the case of structures involving latent variables h, we use negative marginal log-likelihood loss (2) for training.  ... 
arXiv:1505.00308v1 fatcat:sbkgstfcdrcltiu55zuninv4e4

TzK: Flow-Based Conditional Generative Model [article]

Micha Livne, David Fleet
2019 arXiv   pre-print
This allows one to train generative models from multiple, heterogeneous datasets, while retaining strong prior models over subsets of the data (e.g., from a single dataset, class label, or attribute).  ...  We formulate a new class of conditional generative models based on probability flows.  ...  In the next experiment we train a much richer t-flow from the entire multi-data training set of 1,892,916 images, again unsupervised.  ... 
arXiv:1902.01893v4 fatcat:b4bs76fs2nfnhgpduwsrpob5fe

Unsupervised Cyberbullying Detection via Time-Informed Gaussian Mixture Model [article]

Lu Cheng, Kai Shu, Siqi Wu, Yasin N. Silva, Deborah L. Hall, Huan Liu
2020 arXiv   pre-print
Our core contribution is an unsupervised cyberbullying detection model that not only experimentally outperforms the state-of-the-art unsupervised models, but also achieves competitive performance compared  ...  multi-task learning network that simultaneously fits the comment inter-arrival times and estimates the bullying likelihood based on a Gaussian Mixture Model.  ...  (b) Predicted as non-bullying session. Figure 6 : Case study using the Instagram dataset. • Multi-modal features.  ... 
arXiv:2008.02642v1 fatcat:hx54aoyc4bh4jeidljlexyrjl4
« Previous Showing results 1 — 15 out of 13,670 results