Generative kernels
نویسنده
چکیده
Kernel methods are a field of intensive research in machine learning. Lately, much attention has been dedicated to the problem of “kernel learning”, i.e., choosing the kernel that best suits a particular task. Many discriminative approaches avoid handling this problem directly, ignoring the process of data generation to represent them as vectors in a suitable Euclidean space. By contrast, generative approaches propose devising kernels directly by properly modeling the generation of data; this way prior knowledge about their structure may be introduced. In this framework, objects are modeled as outcomes of random processes, for example HMMs for strings, or multinomials for text documents. The aim of this report is providing a survey on generative kernels, capturing the essential theoretical aspects on which they settle, and highlighting the main geometrical ideas.
منابع مشابه
Generative Kernels for Exponential Families
In this paper, we propose a family of kernels for the data distributions belonging to the exponential family. We call these kernels generative kernels because they take into account the generative process of the data. Our proposed method considers the geometry of the data distribution to build a set of efficient closed-form kernels best suited for that distribution. We compare our generative ke...
متن کاملClassification of Biological Sequences with Kernel Methods
Classification of biological sequences with kernel methods Kernels and kernel methods Kernels for biological sequences Summary Outline 1 Kernels and kernel methods Kernels Kernel Methods 2 Kernels for biological sequences Motivations Feature space approach Using generative models Derive from a similarity measure Application: remote homology detection Jean-Philippe Vert Classification of biologi...
متن کاملEcient decoding with continuous rational kernels using the expectation semiring
Semi-Markov conditional randomelds are discriminativemodels that can be used for speech recognition. ey allow per-word (instead of per-frame) features. Since the segmentation into words is not known a priori, all possibilities must be considered. It is therefore important to consider the eciency of the feature extraction process. Features derived from generative models like hmms (log-likeliho...
متن کاملRenal Cancer Cell Classification Using Generative Embeddings and Information Theoretic Kernels
In this paper, we propose a hybrid generative/discriminative classification scheme and apply it to the detection of renal cell carcinoma (RCC) on tissue microarray (TMA) images. In particular we use probabilistic latent semantic analysis (pLSA) as a generative model to perform generative embedding onto the free energy score space (FESS). Subsequently, we use information theoretic kernels on the...
متن کاملCombining information theoretic kernels with generative embeddings for classification
Classical approaches to learn classifiers for structured objects (e.g., images, sequences) use generative models in a standard Bayesian framework. To exploit the state-of-the-art performance of discriminative learning, while also taking advantage of generative models of the data, generative embeddings have been recently proposed as a way of building hybrid discriminative/generative approaches. ...
متن کاملGenerative Kernels and Score-Spaces for Classication of Speech: Progress Report ii
January is is the second progress report for Project /// (Generative Kernels and Score Spaces for Classiication of Speech) within the Global Uncertainties Programme. is project combines the current generative models developed in the speech community with discriminative classiiers. An important aspect of the approach is that the generative models are used to deene a score-space that can be used ...
متن کامل