4,883 Hits in 9.7 sec

Histogram-Equalized Hypercube Adaptive Linear Regression for Image Quality Assessment

N Balakrishnan, S P Shantharajah
2019 Sadhana (Bangalore)  
The efficiency of examination, reduction in the time taken for ROI localization by multiple passes and the quality of the image can be improved by the proposed method, Histogram-Equalized Hypercube Adaptive  ...  In this step, the features used to describe the quality of images are analysed using Histogram-Equalization-based Contrast Masking (HE-CM) model.  ...  ; the proposed FUIQA used two deep convolutional neural network models, L-CNN and C-CNN, respectively.  ... 
doi:10.1007/s12046-019-1148-3 fatcat:inyjq6icbvcmdf2heykgconb7e

Great Tits Learn Odors and Colors Equally Well, and Show No Predisposition for Herbivore-Induced Plant Volatiles

Diana Rubene, Utku Urhan, Velemir Ninkovic, Anders Brodin
2022 Frontiers in Ecology and Evolution  
Great tits showed no initial preference for HIPVs compared to neutral odors, and they learned all olfactory cues at a similar pace, except for methyl salicylate (MeSA), which they learned more slowly.  ...  Ability to efficiently localize productive foraging habitat is crucial for nesting success of insectivorous birds.  ...  For all analyses, we used generalized linear mixed models ACKNOWLEDGMENTS We would like to thank Dimitrije Markovic for assistance with production of odor pellets, Camilla Björklöv for lab supervision  ... 
doi:10.3389/fevo.2021.800057 fatcat:nrfj6yy7xfbqbgtjmskwgbijsy

Reliable Prediction Errors for Deep Neural Networks Using Test-Time Dropout [article]

Isidro Cortes-Ciriano, Andreas Bender
2019 arXiv   pre-print
framework, and to thereby generate reliable prediction errors for deep Neural Networks.  ...  with the confidence level) and efficient, as the predicted confidence intervals span a narrower set of values than those computed with Conformal Predictors generated using Random Forest (RF) models.  ...  The Conformal Predictors we generated using Test-Time Dropout are well calibrated (i.e., valid) for all dropout levels as well as for the RF-based models, as evidenced by the high correlation between the  ... 
arXiv:1904.06330v1 fatcat:eghzja34hbbqtfztlvq6oaj3bm

Concepts and Applications of Conformal Prediction in Computational Drug Discovery [article]

Isidro Cortés-Ciriano, Andreas Bender
2019 arXiv   pre-print
In this review, we summarize underlying concepts and practical applications of CP with a particular focus on virtual screening and activity modelling, and list open source implementations of relevant software  ...  Finally, we describe the current limitations in the field, and provide a perspective on future opportunities for CP in preclinical and clinical drug discovery.  ...  Conformal Prediction Using All Labelled Data for Learning A common disadvantage to all the CP modalities revisited so far is that not all data available for training are used for model fitting, as the  ... 
arXiv:1908.03569v1 fatcat:evr67kv32ve4dg6yd3iqg5ukd4

Ruling out pulmonary embolism across different subgroups of patients and healthcare settings: protocol for a systematic review and individual patient data meta-analysis (IPDMA)

G.-J. Geersing, N. Kraaijpoel, H. R. Büller, S. van Doorn, N. van Es, G. Le Gal, M. V. Huisman, C. Kearon, J. A. Kline, K. G. M. Moons, M. Miniati, M. Righini (+4 others)
2018 Diagnostic and Prognostic Research  
However, their predictive performance differs across different healthcare settings, patient subgroups, and clinical presentation, which are currently not accounted for in the available diagnostic approaches  ...  of each diagnostic strategy, and (iii) optimize and tailor the efficiency and safety of ruling out PE across a broad spectrum of patients with a new, patient-tailored clinical decision model that combines  ...  Funding GJG receives a Veni-grant for performing this project from the Netherlands Organization for Scientific Research (ZonMw 016.166.030).  ... 
doi:10.1186/s41512-018-0032-7 pmid:31093560 pmcid:PMC6460525 fatcat:gpsmvvwzfffyvl7dzj3gzqesxy

Machine Learning Strategies for the Retrieval of Leaf-Chlorophyll Dynamics: Model Choice, Sequential Versus Retraining Learning, and Hyperspectral Predictors

Yoseline Angel, Matthew F. McCabe
2022 Frontiers in Plant Science  
overlaps between relevant bands and VI predictors, highlighting a few decisive spectral ranges and indices useful for retrieving leaf-Chl levels.  ...  spectra and SPAD measurements, capturing temporal correlations, selecting relevant predictors, and retrieving accurate results under different conditions.  ...  Mousa and all the workers at the King Abdulaziz University Agricultural Research Station in Hada Al-Sham for their extensive assistance with field maintenance and harvesting.  ... 
doi:10.3389/fpls.2022.722442 pmid:35360313 pmcid:PMC8963469 fatcat:effkdbwannh2zozzp3ypaif4wy

Confidence Calibration for Deep Renal Biopsy Immunofluorescence Image Classification

Federico Pollastri, Juan Maronas, Federico Bolelli, Giulia Ligabue, Roberto Paredes, Riccardo Magistroni, Costantino Grana
2021 2020 25th International Conference on Pattern Recognition (ICPR)  
Experimental results demonstrate that the designed model yields good accuracy on the specific task, and that TS is able to provide reliable probabilities, which are highly valuable for such a task given  ...  Since modern neural networks often provide overconfident outputs, we stress the importance of having a reliable prediction, demonstrating that Temperature Scaling (TS), a recently introduced re-calibration  ...  This can be viewed as learning the posterior distribution assuming equal prior distributions for all classes.  ... 
doi:10.1109/icpr48806.2021.9412685 fatcat:tfmw735s25a7djf766d6x7nove

Efficient Conformal Prediction via Cascaded Inference with Expanded Admission [article]

Adam Fisch, Tal Schuster, Tommi Jaakkola, Regina Barzilay
2021 arXiv   pre-print
This is particularly pervasive in settings where the correct answer is not unique, and the number of total possible answers is high.  ...  We demonstrate the empirical effectiveness of our approach for multiple applications in natural language processing and computational chemistry for drug discovery.  ...  ACKNOWLEDGEMENTS We specially thank Ben Fisch for his helpful comments and discussion on this work, in addition to Kyle Swanson for help in running the chemprop models.  ... 
arXiv:2007.03114v3 fatcat:cjftzcc5zvdztltj6skc7jhnkq

Flow Prediction Using Remotely Sensed Soil Moisture in Irish Catchments

Chanyu Yang, Fiachra E. O'Loughlin
2020 Water  
Efficiency (NSE) and Coefficient of Determination (R2) for the calibration period.  ...  Using the conceptual hydrological model "Soil Moisture Accounting and Routing for Transport" (SMART), behavioural parameter sets (BPS) were selected using two different objective functions: the Nash Sutcliffe  ...  Across all the catchments calibrated using the R 2 BPS, the SMART model was able to produce simulated hydrographs that produced good NSE scores.  ... 
doi:10.3390/w12082202 fatcat:3225yo6f2vbhhnqdzdtpo2ijmy

Using spatial probability maps to highlight potential inaccuracies in deep learning-based contours – facilitating on-line adaptive radiotherapy

Ward van Rooij, Wilko F. Verbakel, Berend J. Slotman, Max Dahele
2021 Advances in Radiation Oncology  
Deep learning-based delineation (DLD) shows promise both in terms of quality and speed, but it does not yet perform perfectly. Because of that, manual checking of DLD is still recommended.  ...  In general, this technique demonstrated uncertainty in areas that (1) have lower contrast, (2) are less consistently contoured by clinicians, and (3) deviate from the anatomic norm.  ...  To retrieve an uncertainty score for the entire structure, the sum of the uncertainty score of all the voxels was divided by the sum of the object depicting the clinical structure because uncertainties  ... 
doi:10.1016/j.adro.2021.100658 pmid:33778184 pmcid:PMC7985281 fatcat:zei2bj7majax5avfc5ogrjoai4

Efficacy of Bayesian Neural Networks in Active Learning [article]

Vineeth Rakesh, Swayambhoo Jain
2021 arXiv   pre-print
By performing a comprehensive set of experiments, we show that Bayesian neural networks are more efficient than ensemble based techniques in capturing uncertainty.  ...  In this paper, we explore the efficacy of Bayesian neural networks for active learning, which naturally models uncertainty by learning distribution over the weights of neural networks.  ...  recall is defined as the fraction of the total relevant instances (i.e., both true positives and true negatives) that were retrieved.  ... 
arXiv:2104.00896v2 fatcat:5r2t72bcgnfjnceb3dd2452fuq

Spoken Content Retrieval—Beyond Cascading Speech Recognition with Text Retrieval

Lin-shan Lee, James Glass, Hung-yi Lee, Chun-an Chan
2015 IEEE/ACM Transactions on Audio Speech and Language Processing  
to retrieve spoken content that is semantically related to the query, but not necessarily including the query terms themselves; 5) Interactive Retrieval and Efficient Presentation of the Retrieved Objects  ...  This potentially eliminates the requirement of producing text descriptions for multimedia content for indexing and retrieval purposes, and is able to precisely locate the exact time the desired information  ...  The relevance scores in (17) obtained with all pattern sets based on different model granularities are then averaged, and the average scores are used in ranking all the documents for retrieval.  ... 
doi:10.1109/taslp.2015.2438543 fatcat:hwrwmwtlkzfbfagox7bazu5r6a

Detection and Captioning with Unseen Object Classes [article]

Berkan Demirel, Ramazan Gokberk Cinbis
2021 arXiv   pre-print
In order to improve the detection component, we jointly define a class-to-class similarity based class representation and a practical score calibration mechanism.  ...  For this problem, we propose a detection-driven approach based on a generalized zero-shot detection model and a template-based sentence generation model.  ...  In this work, we aim to generate captions that can include classes that are not seen in the supervised training set, where retrieval-based approaches are not directly suitable.  ... 
arXiv:2108.06165v1 fatcat:k4riatvu6vcknggyiwq2kbjcze

Low-Shot Validation: Active Importance Sampling for Estimating Classifier Performance on Rare Categories [article]

Fait Poms, Vishnu Sarukkai, Ravi Teja Mullapudi, Nimit S. Sohoni, William R. Mark, Deva Ramanan, Kayvon Fatahalian
2021 arXiv   pre-print
We propose a statistical validation algorithm that accurately estimates the F-score of binary classifiers for rare categories, where finding relevant examples to evaluate on is particularly challenging  ...  Our key insight is that simultaneous calibration and importance sampling enables accurate estimates even in the low-sample regime (< 300 samples).  ...  Since ImageNet50K is 20× smaller than Im- to ACIS because most of the relevant samples for the F1 ageNet1M there are fewer relevant samples to find for a score have been labeled.  ... 
arXiv:2109.05720v1 fatcat:adylmf6hvvcuxhj5z4e7lq5ryy

Constructing High Precision Knowledge Bases with Subjective and Factual Attributes [article]

Ari Kobren, Pablo Barrio, Oksana Yakhnenko, Johann Hibschman, Ian Langmore
2019 arXiv   pre-print
The results demonstrate that our learned models are well-calibrated and thus can successfully be used to control the KB's precision.  ...  Moreover, when constrained to maintain 95% precision, the best consensus model matches the F-score of a baseline that models each entity-attribute pair as a binary variable and does not support tunable  ...  Second, we measure model calibration. Model-based Attribute Predictor The primary responsibility of the KB is to retrieve all locations that exhibit a queried attribute.  ... 
arXiv:1905.12807v3 fatcat:ikgr3rrs3fhkvf7phbvqqurtn4
« Previous Showing results 1 — 15 out of 4,883 results