Filters








31,915 Hits in 4.0 sec

Detecting semantic anomalies [article]

Faruk Ahmed, Aaron Courville
2019 arXiv   pre-print
We make progress on these benchmarks by exploring a multi-task learning based approach, showing that auxiliary objectives for improved semantic awareness result in improved semantic anomaly detection,  ...  We critically appraise the recent interest in out-of-distribution (OOD) detection and question the practical relevance of existing benchmarks.  ...  Acknowledgements We thank Rachel Rolland for referencing and discussing the motivating examples of anomaly detection in nature studies.  ... 
arXiv:1908.04388v3 fatcat:r3neu3purzdk3k255d2ytot2by

Detecting Semantic Anomalies

Faruk Ahmed, Aaron Courville
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We make progress on these benchmarks by exploring a multi-task learning based approach, showing that auxiliary objectives for improved semantic awareness result in improved semantic anomaly detection,  ...  We critically appraise the recent interest in out-of-distribution (OOD) detection and question the practical relevance of existing benchmarks.  ...  naturally lead to better performance at detecting semantic anomalies.  ... 
doi:10.1609/aaai.v34i04.5712 fatcat:cuqcwa2fmfaqvemfnjjahj2fgi

Transfer-Based Semantic Anomaly Detection

Lucas Deecke, Lukas Ruff, Robert A. Vandermeulen, Hakan Bilen
2021 International Conference on Machine Learning  
Detecting semantic anomalies is challenging due to the countless ways in which they may appear in real-world data.  ...  In this paper, we show that a previously overlooked strategy for anomaly detection (AD) is to introduce an explicit inductive bias toward representations transferred over from some large and varied semantic  ...  Anomaly Detection with an Inductive Bias Due to the complex ways in which semantic anomalies may manifest themselves in images, we hypothesize that our best bet for robust semantic AD is to introduce a  ... 
dblp:conf/icml/DeeckeRVB21 fatcat:rmb62eliv5eejkchuzkyagtlhe

Semantic anomaly detection in online data sources

Orna Raz, Philip Koopman, Mary Shaw
2002 Proceedings of the 24th international conference on Software engineering - ICSE '02  
Anomaly detection e ! ¡ 5 F s ¢ § ' F T ¦ # C ! ' @ 5 £ 2 ! ' 6 ¤ 6 ¤ # § 1 ! ' £ C ¥ 5 0 § © 0 @ r ¡ 5 ! v C ! # C ! @ 5 9 s 9 A § 1 @ G H § c 4 ¤ 2 F c E 9 s ¥ 7 ¦ ¤ 6 ¤ F A 9 T £ 2 9 A | # C !  ...  Semantic quality $ @ ¤ F A 9 T @ ¤ n 4 5 ! D C ! W § 1 ¡ d # £ 0 0 S ! # u 9 A @ x 6 e ! D # 9 A £ 0 ¡ 5 F T ! # R @ ¤ 0 Q 4 § © 9 A G H 6 ¤ # § D ) P Q 4 4 d 0 6 8 0 @ 5 4 5 !  ... 
doi:10.1145/581339.581378 dblp:conf/icse/RazKS02 fatcat:il6j7hsq2rbjhkwirpqfz6xtri

Detecting Anomalies in Semantic Segmentation with Prototypes [article]

Dario Fontanel, Fabio Cermelli, Massimiliano Mancini, Barbara Caputo
2021 arXiv   pre-print
Traditional semantic segmentation methods can recognize at test time only the classes that are present in the training set.  ...  This is a significant limitation, especially for semantic segmentation algorithms mounted on intelligent autonomous systems, deployed in realistic settings.  ...  Conclusions In this paper we addressed the problem of anomaly detection in semantic segmentation, an important but yet scarcely researched topic.  ... 
arXiv:2106.00472v1 fatcat:iwrqogfymrakreg63drnd5yebe

Semantic anomaly detection in online data sources

Orna Raz, Philip Koopman, Mary Shaw
2002 Proceedings of the 24th international conference on Software engineering - ICSE '02  
ICSE'02 -2 -c Orna Raz, 2002 With anomaly detection ± Toby would be notified the data feed is behaving suspiciously ± Our results support semantic anomaly detection Stock quotes, weather forecasts,  ...  semantic anomalies ± Can do automatically, to a large extent µ infer invariants using multiple existing techniques µ use as proxies for missing specifications for anomaly detection where a model is a  ... 
doi:10.1145/581376.581378 fatcat:z4m5cuzit5ddhd22lenf4zjeea

Semantic Anomaly Detection in Medical Time Series [chapter]

Sven Festag, Cord Spreckelsen
2021 Studies in Health Technology and Informatics  
This corresponds to a precision and recall regarding the detection task of around 0.72.  ...  Denoising autoencoders based on recurrent neural networks with gated recurrent units were used for the semantic encoding of such time frames.  ...  Future studies could address the extension of the semantic anomaly detection approach for the recognition of anomalies with variable lengths.  ... 
doi:10.3233/shti210059 pmid:34042884 fatcat:msgt2leys5awhcrrmg5c4o73he

Synthesize then Compare: Detecting Failures and Anomalies for Semantic Segmentation [article]

Yingda Xia, Yi Zhang, Fengze Liu, Wei Shen, Alan Yuille
2020 arXiv   pre-print
The ability to detect failures and anomalies are fundamental requirements for building reliable systems for computer vision applications, especially safety-critical applications of semantic segmentation  ...  In this paper, we systematically study failure and anomaly detection for semantic segmentation and propose a unified framework, consisting of two modules, to address these two related problems.  ...  We propose a unified framework to address failure detection and anomaly detection for semantic segmentation.  ... 
arXiv:2003.08440v2 fatcat:ypqsmpbomjbw3citwwiigdazfq

Unsupervised Discovery, Control, and Disentanglement of Semantic Attributes with Applications to Anomaly Detection [article]

William Paul, I-Jeng Wang, Fady Alajaji, Philippe Burlina
2021 arXiv   pre-print
For (c), we design several systems for anomaly detection exploiting representations learned in (a), and demonstrate their performance benefits when compared to state-of-the-art generative and discriminative  ...  detection methods that leverage representations learned in (a).  ...  Anomaly Detection From the above results in the experiment section with regard to anomaly detection, we remark that anomaly detection based on representation learning from these generative models provides  ... 
arXiv:2002.11169v4 fatcat:deey6zx65vcnnk7ye3t6guykja

HAWatcher: Semantics-Aware Anomaly Detection for Appified Smart Homes

Chenglong Fu, Qiang Zeng, Xiaojiang Du
2021 USENIX Security Symposium  
We propose Home Automation Watcher (HAWatcher), a semantics-aware anomaly detection system for appified smart homes.  ...  Prior works that utilize data mining techniques to detect anomalies suffer from high false alarm rates and missing many real anomalies.  ...  It is a promising direction to combine such semantic information with mining for anomaly detection.  ... 
dblp:conf/uss/00020D21 fatcat:d5xiur7l2ngdlbtttsqidr5amu

Towards combining ontology matchers via anomaly detection

Alexander C. Müller, Heiko Paulheim
2015 International Semantic Web Conference  
In this paper, we present the COMMAND matching system, an unsupervised method for combining base matchers, which uses anomaly detection to produce an alignment from the results delivered by several base  ...  The basic idea of our approach is that in a large set of potential mapping candidates, the scarce actual mappings should be visible as anomalies against the majority of non-mappings.  ...  Thus, it should be possible to discover them using anomaly detection methods.  ... 
dblp:conf/semweb/MullerP15 fatcat:bw56ydei7rgyrlul2g3j5pwi4y

A New Semantic Kernel Function for Online Anomaly Detection of Software

Saeed Parsa
2012 ETRI Journal  
In this letter, a new online anomaly detection approach for software systems is proposed.  ...  The novelty of the proposed approach is to apply a new semantic kernel function for a support vector machine (SVM) classifier to detect faultsuspicious execution paths at runtime in a reasonable amount  ...  To measure the quality of the anomaly reports generated by our semantic kernel, an important criterion is the distance between the exact location of a bug from the point where an anomaly has been detected  ... 
doi:10.4218/etrij.11.0211.0293 fatcat:w6udozxxgvhprmkt2csubtfvs4

A Video Anomaly Detection Framework based on Appearance-Motion Semantics Representation Consistency [article]

Xiangyu Huang, Caidan Zhao, Yilin Wang, Zhiqiang Wu
2022 arXiv   pre-print
Anomalies only occur in the moving foreground of surveillance videos, so the semantics expressed by video frame sequences and optical flow without background information in anomaly detection should be  ...  highly consistent and significant for anomaly detection.  ...  So during the anomaly detection phase, we use the semantics consistency of appearance-motion features and the frame prediction errors as final video anomaly detection cues.  ... 
arXiv:2204.04151v1 fatcat:g66yv3dh7zabjfqhpffxitqdiu

Data Cleaning for XML Electronic Dictionaries via Statistical Anomaly Detection

Michael Bloodgood, Benjamin Strauss
2016 2016 IEEE Tenth International Conference on Semantic Computing (ICSC)  
In this paper we describe methods for flagging statistical anomalies as likely errors in electronic dictionaries stored in XML format.  ...  For each system, we provide an intuitive analysis of the type of error that it is successful at detecting.  ...  Electronic dictionaries are a fundamentally important semantic computing resource.  ... 
doi:10.1109/icsc.2016.38 dblp:conf/semco/BloodgoodS16 fatcat:bejz54zxrbbc3oow2x4itac2yy

Learning Semantic Context from Normal Samples for Unsupervised Anomaly Detection

Xudong Yan, Huaidong Zhang, Xuemiao Xu, Xiaowei Hu, Pheng-Ann Heng
2021 AAAI Conference on Artificial Intelligence  
anomaly detection by learning the semantic context from the normal samples.  ...  The inference of abnormal regions on the input image requires an understanding of the surrounding semantic context.This work presents a Semantic Context based Anomaly Detection Network, SCADN, for unsupervised  ...  To this end, we design a novel deep network architecture via learning semantic context from normal samples, i.e., Semantic Context based Anomaly Detection Network (SCADN), which learns the semantic context  ... 
dblp:conf/aaai/YanZXHH21 fatcat:7okbjmft35cglmohwirjlos4sy
« Previous Showing results 1 — 15 out of 31,915 results