Active Clustering with Model-Based Uncertainty Reduction
نویسندگان
چکیده
منابع مشابه
Uncertainty Reduction for Active Image Clustering via a Hybrid Global-Local Uncertainty Model
We propose a novel combined global/local model for active semi-supervised spectral clustering based on the principle of uncertainty reduction. We iteratively compute the derivative of the eigenvectors produced by spectral decomposition with respect to each item/image, and combine this with local label entropy provided by the current clustering results in order to estimate the uncertainty reduct...
متن کاملClustering-Based Control of Active Contour Model
To extract object regions from images, the methods using region-based active contour model (ACM) have been proposed. By controlling ACM with the statistical characteristics of the image properties, these methods effect robust region extraction. However, the existing methods require redundant processing and cannot adapt to complex scene images. To overcome these problems, we propose a new method...
متن کاملA Logic-based Model for uncertainty reduction: parallel processing
Geographic Information Systems (GIS) often use incomplete and uncertain information leading to inconsistency. Thus, a clear definition of information revision is required. In logic, most of the usual belief revision operations are characterized by a high computational complexity, which implies actual intractability even for rather small data sets. On the other hand, GIS deal with large amounts ...
متن کاملClustering-Based Active Learning
In contexts where obtaining labels for data points is expensive, active learning is a widely used methodology for deciding which data points to get a label for. In this project, we propose and evaluate an active learning algorithm in context of an automatic grading and feedback generation tool for an online embedded systems course currently under development at UC Berkeley. This tool learns a f...
متن کاملModel Based Approaches for Simultaneous Dimension Reduction and Clustering
High dimensional data are regularly generated from various sources. Thus subsets of data are usually concentrated in different subspaces. It is of interest to develop methods that achieve simultaneous dimension reduction and clustering. In this talk we will present a new solution based on a constrained version of mixture of factor analyzers. Our proposed technique imposes constraints on the und...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE Transactions on Pattern Analysis and Machine Intelligence
سال: 2017
ISSN: 0162-8828,2160-9292
DOI: 10.1109/tpami.2016.2539965