1,059 Hits in 4.1 sec

Performance-Agnostic Fusion of Probabilistic Classifier Outputs [article]

Jordan F. Masakuna, Simukai W. Utete, Steve Kroon
2020 arXiv   pre-print
We propose a method for combining probabilistic outputs of classifiers to make a single consensus class prediction when no further information about the individual classifiers is available, beyond that  ...  Performance of the model is demonstrated on different benchmark datasets.  ...  Such information about classifier performance is invaluable for high-performance classifier fusion, since it allows one to account explicitly for the quality of each classifier.  ... 
arXiv:2009.00565v1 fatcat:hmg6rjz27jbdtjrsd7rd3c5psu

Multichannel Semantic Segmentation with Unsupervised Domain Adaptation [article]

Kohei Watanabe, Kuniaki Saito, Yoshitaka Ushiku, Tatsuya Harada
2018 arXiv   pre-print
One is a fusion-based approach that uses depth images as inputs. The other is a multitask learning approach that uses depth images as outputs.  ...  However, models trained simply on synthetic images tend to demonstrate poor performance on real images.  ...  RGB-HHA Feature Generator Classifier Segmentation output In this figure, one classifier (two classifiers in the score fusion model) exists but actually two classifiers (four classifiers in the score  ... 
arXiv:1812.04351v1 fatcat:zx3poueipfcmnbqqxpuvs2yqky

Multichannel Semantic Segmentation with Unsupervised Domain Adaptation [chapter]

Kohei Watanabe, Kuniaki Saito, Yoshitaka Ushiku, Tatsuya Harada
2019 Lecture Notes in Computer Science  
One is a fusion-based approach that uses depth images as inputs. The other is a multitask learning approach that uses depth images as outputs.  ...  However, models trained simply on synthetic images tend to demonstrate poor performance on real images.  ...  We use the notation p 1 (y 1 |x), p 2 (y 1 |x) to denote the (K × |x|)-dimensional probabilistic outputs for input x obtained by C 1 and C 2 respectively.  ... 
doi:10.1007/978-3-030-11021-5_37 fatcat:5d35nzuiyjd35epnqnyfcsyuhi

Deep Learning-Based Methods for Sentiment Analysis on Nepali COVID-19-Related Tweets

C. Sitaula, A. Basnet, A. Mainali, T. B. Shahi, Thippa Reddy G
2021 Computational Intelligence and Neuroscience  
For this, we, first, propose to use three different feature extraction methods—fastText-based (ft), domain-specific (ds), and domain-agnostic (da)—for the representation of tweets.  ...  Moreover, the proposed CNN models impart robust and stable performance on the proposed features.  ...  Decision Fusion. We perform a fusion of decisions obtained from three different pretrained CNNs using ensemble CNN model for the end results (equation ( 17 )).  ... 
doi:10.1155/2021/2158184 pmid:34737773 pmcid:PMC8561567 fatcat:i2nwtwhlgjatze2kk5vwp4ogfy

Fusing Saliency Maps with Region Proposals for Unsupervised Object Localization [article]

Hakan Karaoguz, Patric Jensfelt
2018 arXiv   pre-print
The resulting saliency map is matched with region proposals from a class agnostic region proposal network to roughly localize the candidate object regions.  ...  In this paper we address the problem of unsupervised localization of objects in single images.  ...  [23] uses color and depth cues independently to find the object candidates for both cues and perform late-fusion to refine the results of separate cues. Abbeloos et al.  ... 
arXiv:1804.03905v1 fatcat:yxeig2urevf55eckailw3jzinq

Staircase Regression in OA RVM, Data Selection and Gender Dependency in AVEC 2016

Zhaocheng Huang, Brian Stasak, Ting Dang, Kalani Wataraka Gamage, Phu Le, Vidhyasaharan Sethu, Julien Epps
2016 Proceedings of the 6th International Workshop on Audio/Visual Emotion Challenge - AVEC '16  
For emotion prediction, we investigate emotionally salient data selection based on emotion change, an output-associative regression approach based on the probabilistic outputs of relevance vector machine  ...  classifiers operating on low-high class pairs (OA RVM-SR), and gender-dependent systems.  ...  1 probabilistic output 1 RVM classifier 2 probabilistic output 2 RVM classifier N-1 probabilistic output N-1 Prediction score • • • • • • • • • • • • Table 1 : 1 F1 scores  ... 
doi:10.1145/2988257.2988265 dblp:conf/mm/HuangSDGLSE16 fatcat:boq2umnkknhtfbaqqtytsetuqi

Model-Agnostic Multi-Agent Perception Framework [article]

Weizhe Chen, Runsheng Xu, Hao Xiang, Lantao Liu, Jiaqi Ma
2022 arXiv   pre-print
of their neighbors.  ...  In this work, we propose a model-agnostic multi-agent framework to reduce the negative effect caused by model discrepancies and maintain confidentiality.  ...  For a probabilistic classifier, the probability associated with the predicted class label should reflect its correctness likelihood.  ... 
arXiv:2203.13168v1 fatcat:ohfxpl3tybggdfy2t2sxm6slhi

Real-Time Volumetric-Semantic Exploration and Mapping: An Uncertainty-Aware Approach [article]

Rui Pimentel de Figueiredo, Jonas le Fevre Sejersen, Jakob Grimm Hansen, Martim Brandão, Erdal Kayacan
2021 arXiv   pre-print
Second, we extend the reward function to incorporate not only geometric but also semantic probabilistic information, provided by a DCNN for semantic segmentation that operates in real-time.  ...  A set of simulations in realistic scenarios demonstrate the efficacy and efficiency of the proposed framework when compared with the state-of-the-art.  ...  In this case although planning and mapping runs slower due to the computations involved in estimating the quality (probabilistic) of both geometric and semantic measurements and probabilistic fusion in  ... 
arXiv:2109.01474v1 fatcat:gumsu2hrffa4rf67fegjz3a5ty

ConceptDistil: Model-Agnostic Distillation of Concept Explanations [article]

João Bento Sousa, Ricardo Moreira, Vladimir Balayan, Pedro Saleiro, Pedro Bizarro
2022 arXiv   pre-print
In this work, we propose ConceptDistil, a method to bring concept explanations to any black-box classifier using knowledge distillation.  ...  ConceptDistil is decomposed into two components:(1) a concept model that predicts which domain concepts are present in a given instance, and (2) a distillation model that tries to mimic the predictions of  ...  Information Fusion, 81:84-90, 2022.  ... 
arXiv:2205.03601v1 fatcat:xxh4wdo35nbethkd4vcbvf3fsy

Monitoring of Perception Systems: Deterministic, Probabilistic, and Learning-based Fault Detection and Identification [article]

Pasquale Antonante, Heath Nilsen, Luca Carlone
2022 arXiv   pre-print
We then provide a set of deterministic, probabilistic, and learning-based algorithms that use diagnostic graphs to perform fault detection and identification.  ...  Moreover, we investigate fundamental limits and provide deterministic and probabilistic guarantees on the fault detection and identification results.  ...  For example, imagine a diagnostic test that compares the output of two object classifiers: if one of them produces a wrong label, it is easy to detect there is a failure; however, if both classifiers are  ... 
arXiv:2205.10906v1 fatcat:qrfjohpx4rglhdh5xdcsbavnwy

Privileged Information for Modeling Affect In The Wild [article]

Konstantinos Makantasis, David Melhart, Antonios Liapis, Georgios N. Yannakakis
2021 arXiv   pre-print
drop in their modeling performance.  ...  The proposed privileged information framework is tested in a game arousal corpus that contains physiological signals in the form of heart rate and electrodermal activity, game telemetry, and pixels of  ...  As a baseline performance, we also report the accuracy of a dummy classifier, denoted as majority class, which always outputs the most frequent class in the training set.  ... 
arXiv:2107.10552v1 fatcat:u334gcq45na6bk7zf64eag57py

Enhancing action recognition through simultaneous semantic mapping from body-worn motion sensors

Michael Hardegger, Long-Van Nguyen-Dinh, Alberto Calatroni, Gerhard Tröster, Daniel Roggen
2014 Proceedings of the 2014 ACM International Symposium on Wearable Computers - ISWC '14  
The fusion approach is agnostic to the sensor modalities and methods used for action recognition and localization.  ...  We present an unsupervised fusion method that takes advantage of this characteristic to enhance the recognition of location-related actions (e.g., open, close, switch, etc.).  ...  Performance Measures To evaluate the performance of LocAFusion, we compare the outputᾱ i of the algorithm with the ground truth labels α i , and check how the accuracy relates to the location-agnostic  ... 
doi:10.1145/2634317.2634323 dblp:conf/iswc/HardeggerNCTR14 fatcat:edpnfi7rsjewhk2xuimb2ytub4

Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI [article]

Alejandro Barredo Arrieta, Natalia Díaz-Rodríguez, Javier Del Ser, Adrien Bennetot, Siham Tabik, Alberto Barbado, Salvador García, Sergio Gil-López, Daniel Molina, Richard Benjamins, Raja Chatila, Francisco Herrera
2019 arXiv   pre-print
This literature analysis serves as the background for a series of challenges faced by XAI, such as the crossroads between data fusion and explainability.  ...  not present in the last hype of AI.  ...  Javier Del Ser also acknowledges funding support from the Consolidated Research Group MATHMODE (IT1294-19) granted by the Department of Education of the Basque Government.  ... 
arXiv:1910.10045v2 fatcat:hgoi7cvkazdd5jycsdfkn565di

Leveraging graphical models to improve accuracy and reduce privacy risks of mobile sensing

Abhinav Parate, Meng-Chieh Chiu, Deepak Ganesan, Benjamin M. Marlin
2013 Proceeding of the 11th annual international conference on Mobile systems, applications, and services - MobiSys '13  
temporal smoothing and fusion mechanisms to further boost performance by just connecting to our higher-level inference engine.  ...  To applications and users, CQue provides a query interface, allowing a) applications to obtain more accurate context results while remaining agnostic of what classifiers/sensors are used and when, and  ...  fusion mechanisms to improve performance.  ... 
doi:10.1145/2462456.2464457 dblp:conf/mobisys/ParateCGM13 fatcat:ddptfil5mjc2firizakwfwg4le

Hierarchical audio-visual cue integration framework for activity analysis in intelligent meeting rooms

Shankar T. Shivappa, Mohan M. Trivedi, Bhaskar D. Rao
2009 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops  
We exploit the fact that the output of one kind of human activity analysis task contains valuable information for another such block and by interconnecting them, a robust system results.  ...  The system performs the tasks of person tracking, head pose estimation, beamforming, speaker ID and speech recognition using audio and visual cues.  ...  In [4] , the authors develop a probabilistic integration framework for fusion of audio visual cues at the track and identity levels. This is an example of fusion at multiple levels of abstraction.  ... 
doi:10.1109/cvprw.2009.5204224 dblp:conf/cvpr/ShivappaTR09 fatcat:wywnweppbfahhbalu4wh3hzbri
« Previous Showing results 1 — 15 out of 1,059 results