A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Filters
Distance-Penalized Active Learning Using Quantile Search
[article]
2017
arXiv
pre-print
Preserved Fulltext
Other Versions
quantile search. ...
We show that for one-dimensional threshold classifiers, a tradeoff between the number of samples taken and distance traveled can be achieved using a generalization of binary search, which we refer to as ...
Extending this idea to penalize distance traveled is a promising avenue for practical applications of quantile search. ...
arXiv:1509.08387v2
fatcat:myqkiqdgtzauhnhr2gydcjlqva
A decision-theoretic approach to the evaluation of machine learning algorithms in computational drug discovery
2019
Bioinformatics
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
However, small-molecule structure-activity data are high dimensional with low signal-to-noise ratios and proper validation of predictive methods is difficult. ...
It is poorly understood which, if any, of the currently available machine learning algorithms will best predict new candidate drugs. ...
Quantile bootstrap Decreasing the quantile-activity threshold for the training data from 1 (random partitioning described above) to 0.4 (only 40% of the data ordered by activity are used in the bootstrap ...
doi:10.1093/bioinformatics/btz293
pmid:31070704
pmcid:PMC6853675
fatcat:aeyp62iqinegzgptmyu7aftiju
A decision theoretic approach to model evaluation in computational drug discovery
[article]
2018
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
Partitioning based on quantiles of the activity distribution correctly penalizes models which can extrapolate onto structurally different molecules outside of the training data. ...
A new validation method, based on quantile splits on the activity distribution function, is proposed for the construction of training and testing sets. ...
This directly reflects how the different loss functions penalize predictive performance, with L γ min only penalizing the rank of the first active molecule. ...
arXiv:1807.08926v1
fatcat:eb5mznk3dbdrbeflbbpir7hyzm
Non-Parametric and Regularized Dynamical Wasserstein Barycenters for Time-Series Analysis
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Here, focusing on the univariate case where Wasserstein distances and barycenters can be computed in closed form, we extend [1] by discussing two challenges associated with learning a DWB model and two ...
We are particularly motivated by applications such as human activity analysis where the observed time-series contains segments representing distinct activities such as running or walking as well as segments ...
To counteract this diverging behavior, we propose a regularizer in the space of quantile functions that penalizes the Wasserstein distance of the pure state quantile functions P −1 q k from a reference ...
arXiv:2210.01918v2
fatcat:agtbiqua75b6hdahvueqzyy4zu
Lilikoi V2.0: a deep learning-enabled, personalized pathway-based R package for diagnosis and prognosis predictions using metabolomics data
2021
GigaScience
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
previously we developed Lilikoi, a personalized pathway-based method to classify diseases using metabolomics data. ...
The new Lilikoi v2.0 R package has implemented a deep learning method for classification, in addition to popular machine learning methods. ...
Lilikoi v2.0 supports users to run hyperparameter grid search on multiple deep learning models to achieve the best classification results. The activation functions are set as "Rectifier" or "Tanh." ...
doi:10.1093/gigascience/giaa162
pmid:33484242
pmcid:PMC7825009
fatcat:wleli42rpbbhzbfmpv5b44z5hu
Lilikoi V2.0: a deep-learning enabled, personalized pathway-based R package for diagnosis and prognosis predictions using metabolomics data
[article]
2020
bioRxiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Previously we developed Lilikoi, a personalized pathway-based method to classify diseases using metabolomics data. ...
The new Lilikoi v2.0 R package has implemented a deep-learning method for classification, in addition to popular machine learning methods. ...
Lilikoi v2.0 supports users to run hyperparameter grid search on multiple deep learning models to achieve the best classification results. The activation functions are set as "Rectifier" or "Tanh". ...
doi:10.1101/2020.07.09.195677
fatcat:op2s5dl6bzcypiwng2x4licy4i
Multivariate Probabilistic CRPS Learning with an Application to Day-Ahead Electricity Prices
[article]
2024
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2024; you can also visit the original URL.
The file type is application/pdf.
Other Versions
We discuss two smoothing methods: dimensionality reduction using Basis matrices and penalized smoothing. ...
It is based on Bernstein Online Aggregation (BOA) and yields optimal asymptotic learning properties. The procedure uses horizontal aggregation, i.e., aggregation across quantiles. ...
Therefore, we solely use the penalized smoothing approach for our learning task. We use 99 knots, i.e., one on each quantile. ...
arXiv:2303.10019v3
fatcat:pfla2jfsb5bjlezwvkt6xabdti
Monte Carlo Vehicle Routing
2020
European Conference on Artificial Intelligence
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Nested Rollout Policy Adaptation (NRPA) is a Monte Carlo search algorithm that learns a playout policy in order to solve a single player game. ...
NRPA gives better result than the algorithm previously used by EDF. ...
end if
The Quantile Heuristic The objective of the Quantile Heuristic is to penalize movements from bad solutions. ...
dblp:conf/ecai/CazenaveLKT20
fatcat:jegsv2ceubcfzbnvoddugtaog4
Optimal Search with Neural Networks: Challenges and Approaches
2022
Proceedings of the International Symposium on Combinatorial Search
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
This paper looks at challenges in using deep learning as a part of optimal search, including what is feasible using current public frameworks, and what barriers exist for further adoption. ...
Several approaches are described, with the most successful approach being based on learning a heuristic as a classifier and then adjusting the quantile used with the classifier to ensure heuristic admissibility ...
A breadth-first search is used to compute exact distances in φ(V ), which are then stored in a table and used as heuristics in the original state space. ...
doi:10.1609/socs.v15i1.21758
fatcat:qu6qokashbbfxi6jzu7jsaiype
A Novel Interval Energy-Forecasting Method for Sustainable Building Management Based on Deep Learning
2022
Sustainability
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
The methodology is based on the long short-term memory (LSTM) model and penalized quantile regression (PQR). ...
A comprehensive analysis for a time period of a year is conducted using the proposed method, and back propagation neural networks (BPNN), support vector machine (SVM), and random forest are applied as ...
with penalized quantile regression, as quantile regression has so far not been adequately studied in the field of building energy consumption prediction; (2) verify the advantages of deep learning in ...
doi:10.3390/su14148584
fatcat:yi5yawdpjbfx5cs6krloggf4su
A Data-Mining Scheme for Identifying Peptide Structural Motifs Responsible for Different MS/MS Fragmentation Intensity Patterns
2008
Journal of Proteome Research
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2016; you can also visit the original URL.
The file type is application/pdf.
Keywords: data mining • cluster analysis • K-means algorithm • penalized K-means algorithm • CART • statistical analysis • quantile map • peptide • MS/MS • intensity • ion trap • CID • fragmentation pattern ...
Fragmentation intensity information from 28 330 ion trap peptide MS/MS spectra of different charge states and sequences went through unsupervised clustering using a penalized K-means algorithm. ...
An adjusted Euclidean distance is used, which accounts for missing values but does not penalize them. ...
doi:10.1021/pr070106u
pmid:18052120
pmcid:PMC2464298
fatcat:znkfrgmg5fhdvchafrszcmfsyu
Learnability of Timescale Graphical Event Models
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
I propose and evaluate different heuristics to determine hyper-parameters during the structure learning algorithm and refine an existing distance measure. ...
The proposed learning algorithm follows two steps: in the Forward search edges are added and timescales are refined, whereas the Backward search tries to simplify the model. ...
search approach. ...
arXiv:2005.12186v1
fatcat:cl7qhghdgbb5xddtf773erouam
Autoregressive Quantile Networks for Generative Modeling
[article]
2018
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We introduce autoregressive implicit quantile networks (AIQN), a fundamentally different approach to generative modeling than those commonly used, that implicitly captures the distribution using quantile ...
In this work we extend the PixelCNN model with AIQN and demonstrate results on CIFAR-10 and ImageNet using Inception score, FID, non-cherry-picked samples, and inpainting results. ...
The evaluation metric used for the hyperparameter search was the Fréchet Inception Distance (FID) (Heusel et al., 2017) , see Appendix for details. ...
arXiv:1806.05575v1
fatcat:s5c7xt4afnftfbiw2mwlswwwg4
Leadership, Public Health Messaging, and Containment of Mobility in Mexico During the COVID-19 Pandemic
2021
Social Science Research Network
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Our findings, using a LASSO penalized quantile regression, indicate that when President AMLO's discourse is positive, mobility tends to decrease, but when the Health Minister's discourse is positive, mobility ...
Using quantile regressions, sentiment analysis and topic modelling, we shed light on the dynamics between political discourse and containment of mobility. ...
Quantile Regression Results Figures 11 to 14 present the coefficient estimations of equation 4, the LASSO penalized quantile regression. ...
doi:10.2139/ssrn.3786398
fatcat:rfgfirppffcbvnuwgqusthjixy
Calibrated Forecasts of Quasi-Periodic Climate Processes with Deep Echo State Networks and Penalized Quantile Regression
[article]
2023
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf.
This work aims at showing how 1) data-driven, stochastic machine learning approaches provide an affordable yet flexible means to forecast these processes; 2) the associated uncertainty can be properly ...
1} and w = 1, . . . , n w from the DESN in equation ( 5 ) for w = 1, . . . , n w do • Calculate the residuals for the forecasts R w from equation ( 6 ) • Fit the residuals using a penalized quantile ...
Denote the fitted quantiles as μ(k, q 1 ) and μ(k, q 2 ) • Calculate the distance between the median q = 0.5 and each fitted quantile curve μ(k, q 1 ) and μ(k, q 2 ). ...
arXiv:2308.04391v1
fatcat:e7bs6zy4pjdfleyq6e2jwabgyi
« Previous
Showing results 1 — 15 out of 1,914 results