Exemplar-Supported Representation for Effective Class-Incremental Learning
نویسندگان
چکیده
منابع مشابه
Dynamic class imbalance learning for incremental LPSVM
Linear Proximal Support Vector Machines (LPSVMs), like decision trees, classic SVM, etc. are originally not equipped to handle drifting data streams that exhibit high and varying degrees of class imbalance. For online classification of data streams with imbalanced class distribution, we propose a dynamic class imbalance learning (DCIL) approach to incremental LPSVM (IncLPSVM) modeling. In doing...
متن کاملIncremental Class Dictionary Learning and Optimization
We have previously shown how the discovery of classes from objects can be automated, and how the resulting class organization can be eeciently optimized in the case where the optimum is a single inheritance class hierarchy. This paper extends our previous work by showing how an optimal class dictionary can be learned incrementally. The ability to expand a class organization incrementally as new...
متن کاملExemplar-based Sparse Representation for Posterior Features
Posterior features have been shown to yield very good performance in multiple contexts including speech recognition, spoken term detection, and template matching. These days, posterior features are usually estimated at the output of a neural network. More recently, sparse representation has also been shown to potentially provide additional advantages to improve discrimination and robustness. On...
متن کاملAn Effective Analysis Method for Computer-Supported Learning Systems Reusability
The development of computer-supported learning systems is a complex task that must take into account diverse issues and perspectives. Given the difficulties attached to this process, different authors have proposed the use of software engineering as a reference to optimise the learning material development process. Following this trend, and given the efficiency of the outcomes, it seems conveni...
متن کاملIncremental SampleBoost for Efficient Learning from Multi-Class Data Sets
Ensemble methods have been used for incremental learning. Yet, there are several issues that require attention, including elongated training time and smooth integration of new examples. In this article, we introduce an incremental SampleBoost method that learns efficiently from new data by employing a class-based down sampling strategy with an error parameter. Our novel weight initialization sc...
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 Access
سال: 2020
ISSN: 2169-3536
DOI: 10.1109/access.2020.2980386