Relative Novelty Detection
نویسندگان
چکیده
Novelty detection is an important tool for unsupervised data analysis. It relies on finding regions of low density within which events are then flagged as novel. By design this is dependent on the underlying measure of the space. In this paper we derive a formulation which is able to address this problem by allowing for a reference measure to be given in the form of a sample from an alternative distribution. We show that this optimization problem can be solved efficiently and that it works well in practice.
منابع مشابه
Striatal and midbrain connectivity with the hippocampus selectively boosts memory for contextual novelty
The role of contextual expectation in processing familiar and novel stimuli was investigated in a series of experiments combining eye tracking, functional magnetic resonance imaging, and behavioral methods. An experimental paradigm emphasizing either familiarity or novelty detection at retrieval was used. The detection of unexpected familiar and novel stimuli, which were characterized by lower ...
متن کاملNovelty Detection on a Mobile Robot Using Habituation
In this paper a novelty filter is introduced which allows a robot operating in an unstructured environment to produce a self-organised model of its surroundings and to detect deviations from the learned model. The environment is perceived using the robot’s 16 sonar sensors. The algorithm produces a novelty measure for each sensor scan relative to the model it has learned. This means that it hig...
متن کاملUsing P300 to Evaluate the Effect of Object Color Knowledge in Novelty Detection
A B S T R A C T Introduction: In an oddball experiment, the context in which novel stimuli are presented affects characteristics of novelty P3, i.e. as long as there is a difficult task in which the difference between standard and target stimuli is small, recurrent presentation of a highly discrepant stimulus can lead to P300 highly similar to novelty P3. Effect of stimulus properties on P300 h...
متن کاملProbabilistic Anomaly Detection in Dynamic Systems
In This paper describes probabilistic methods for novelty detection when using pattern recognition methods for fault monitoring of dynamic systems. The problem of novelty detection is particularly acute when prior knowledge and training data only allow one to construct an incomplete classification model. Allowance must be made in model design so that the classifier will be robust to data genera...
متن کاملHuman hippocampal processing of environmental novelty during spatial navigation
The detection and processing of novel information encountered as we explore our environment is crucial for learning and adaptive behavior. The human hippocampus has been strongly implicated in laboratory tests of novelty detection and episodic memory, but has been less well studied during more ethological tasks such as spatial navigation, typically used in animals. We examined fMRI BOLD activit...
متن کامل