Filters








64,128 Hits in 4.6 sec

Learning Interpretable Error Functions for Combinatorial Optimization Problem Modeling [article]

Florian Richoux, Jean-François Baffier
2021 arXiv   pre-print
Our method uses a variant of neural networks we named Interpretable Compositional Networks, allowing us to get interpretable results, unlike regular artificial neural networks.  ...  This is, to the best of our knowledge, the first attempt to automatically learn error functions for hard constraints.  ...  Compositional Networks, a variant of neural networks to get interpretable results, 3. to propose an architecture of Interpretable Compositional Network to learn error functions, and 4. to provide a proof  ... 
arXiv:2002.09811v4 fatcat:oilv5cgmcrhzfkf4ilu6z4xugq

Automatic well test interpretation based on convolutional neural network for a radial composite reservoir

Daolun LI, Xuliang LIU, Wenshu ZHA, Jinghai YANG, Detang LU
2020 Petroleum Exploration and Development  
An automatic well test interpretation method for radial composite reservoirs based on convolutional neural network (CNN) is proposed, and its effectiveness and accuracy are verified by actual field data  ...  In this paper, based on the data transformed by logarithm function and the loss function of mean square error (MSE), the optimal CNN is obtained by reducing the loss function to optimize the network with  ...  It can be seen that the proposed well test interpretation method based on CNN can interpret radial composite reservoir parameters with high accuracy and efficiency, and realize automatic interpretation  ... 
doi:10.1016/s1876-3804(20)60079-9 fatcat:fwu72dksbfggtdscffe34cs6j4

Neurosymbolic Programming

Swarat Chaudhuri, Kevin Ellis, Oleksandr Polozov, Rishabh Singh, Armando Solar-Lezama, Yisong Yue
2021 Foundations and Trends® in Programming Languages  
Like in classic machine learning, the goal here is to learn functions from data.  ...  Neurosymbolic representations are also, commonly, easier to interpret and formally verify than neural networks.  ...  Such programs are easy to write or automatically synthesize in many domains (Zhan et al., 2020; Sun et al., 2020) . When they are available, they can drastically reduce the cost of learning.  ... 
doi:10.1561/2500000049 fatcat:yf7hfvpborh73ht3ukw7ok7axm

Hybrid neural networks: An evolutionary approach with local search

Eduardo Masato Iyoda, Fernando J. Von Zuben
2002 Integrated Computer-Aided Engineering  
This paper is organized as follows: in Section 2, the traditional single hidden layer neural network with identical activation functions is presented.  ...  restriction: additive composition is the only way to combine the possibly distinct activation functions in order to produce the network output.  ...  Activation function the cascade of hybrid compositions) can be interpreted as particular cases of a generalized model of a neuron, presented in Fig. 4 .  ... 
doi:10.3233/ica-2002-9104 fatcat:242tnk65jnfdrpupygebascniy

Aesthetic Photo Collage with Deep Reinforcement Learning [article]

Mingrui Zhang, Mading Li, Li Chen, Jiahao Yu
2021 arXiv   pre-print
Photo collage aims to automatically arrange multiple photos on a given canvas with high aesthetic quality.  ...  In this paper, we propose a novel pipeline for automatic generation of aspect ratio specified collage and the reinforcement learning technique is introduced in collage for the first time.  ...  We decompose the collage generation into interpretable steps and model it as an reinforcement learning process, which to our knowledge, is the first work for directly applying deep learning in automatic  ... 
arXiv:2110.09775v1 fatcat:yixcot5ylff45jj3o4q53h42ii

Instructional design using component-based development and learning object classification

N. Pukkhem, W. Vatanawood
2005 Fifth IEEE International Conference on Advanced Learning Technologies (ICALT'05)  
Moreover, a composition technique of constructing the composite operations is presented by using Requirements Particle Networks.  ...  We demonstrate our formal specification method with a case study and the correctness of final specification from our method is proved with CafeOBJ interpreter.  ...  Moreover, pairs of modules with the high values of coupling measurement are located automatically.  ... 
doi:10.1109/icalt.2005.172 dblp:conf/icalt/PukkhemV05 fatcat:pd7lwnsygzhxnbnqarddek5gaa

MMLSpark: Unifying Machine Learning Ecosystems at Massive Scales [article]

Mark Hamilton, Sudarshan Raghunathan, Ilya Matiach, Andrew Schonhoffer, Anand Raman, Eli Barzilay, Karthik Rajendran, Dalitso Banda, Casey Jisoo Hong, Manon Knoertzer, Ben Brodsky, Minsoo Thigpen, Janhavi Suresh Mahajan (+3 others)
2019 arXiv   pre-print
Orchestration, Gradient Boosting, Model Interpretability, and other areas of modern computation.  ...  We introduce Microsoft Machine Learning for Apache Spark (MMLSpark), an ecosystem of enhancements that expand the Apache Spark distributed computing library to tackle problems in Deep Learning, Micro-Service  ...  If services reside on the same machine, one can use local networking capabilities to bypass internet data transfer costs and come closer to the latency of normal function dispatch.  ... 
arXiv:1810.08744v2 fatcat:5dpnnyfhczakjmkodv4tgcxzi4

Application of Deep Learning in Microbiome

Qiang Zhu, Ban Huo, Han Sun, Bojing Li, Xingpeng Jiang
2020 Journal of Artificial Intelligence for Medical Sciences  
Meanwhile, the rise of deep learning enables us to deal with these complex problems effectively.  ...  the accuracy and interpretability of the model.  ...  Compared with traditional machine learning methods, deep learning has advantages on pattern discovering automatically.  ... 
doi:10.2991/jaims.d.201028.001 fatcat:hnopfambffdlrcgbi4x4ud6phi

Robust Neurofuzzy Rule Base Knowledge Extraction and Estimation Using Subspace Decomposition Combined With Regularization and D-Optimality

X. Hong, C.J. Harris, S. Chen
2004 IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics)  
By using a weighting for the D-optimality cost function, the entire model construction procedure becomes automatic.  ...  Model rule bases are decomposed into orthogonal subspaces, so as to enhance model transparency with the capability of interpreting the derived rule base energy level.  ...  In recent studies [15] , [16] , the authors have outlined efficient learning algorithms, in which composite cost functions were introduced to optimize the model approximation ability using the forward  ... 
doi:10.1109/tsmcb.2003.817089 pmid:15369096 fatcat:2y3klf4eqzfubhozvy77oebvwu

Monitoring of Coral Reefs Using Artificial Intelligence: A Feasible and Cost-Effective Approach

González-Rivero, Beijbom, Rodriguez-Ramirez, EP Bryant, Ganase, Gonzalez-Marrero, Herrera-Reveles, V Kennedy, Kim, Lopez-Marcano, Markey, P Neal (+8 others)
2020 Remote Sensing  
Here, we evaluated the performance of Deep Learning Convolutional Neural Networks for automated image analysis, using a global coral reef monitoring dataset.  ...  Using this automated approach, data analysis and reporting can be accelerated by at least 200x and at a fraction of the cost (1%).  ...  Furthermore, the error of network estimations was comparable to the error associated with multiple LME, P = 0.691 Error = 2.9 ± 0.68 % Cost-Benefit Analysis of Implementing Deep Learning The cost of  ... 
doi:10.3390/rs12030489 fatcat:wbcloaudtvavflrlockthqgqbe

Interpretable policy derivation for reinforcement learning based on evolutionary feature synthesis

Hengzhe Zhang, Aimin Zhou, Xin Lin
2020 Complex & Intelligent Systems  
To deal with this problem, some researchers resort to the interpretable control policy generation algorithm.  ...  The experiment results reveal that evolutionary feature synthesis can achieve better performance than tree-based genetic programming to extract policy from the neural network with comparable interpretability  ...  The experimental results show that our algorithm achieves better performance than GP, DNN, and common interpretable machine learning algorithms with comparable interpretability.  ... 
doi:10.1007/s40747-020-00175-y fatcat:f2krj5owtbbgjhpdralrzxqkmq

One method of generating synthetic data to assess the upper limit of machine learning algorithms performance

Yan I. Kuchin, Ravil I. Muhkamediev, Kirill O. Yakunin, Duc Pham
2020 Cogent Engineering  
Information technology tools, such as predictive analytics with Supervised Machine Learning (SML) algorithms and Artificial Neural Networks (ANN) models, are nowadays widely used to automate geophysical  ...  Previous experiments showed an ANN accuracy of about 60% in the task of lithological interpretation of logging data.  ...  To adjust the weights θ of the neural network (network training), a cost function resembling the logistic regression cost function is used.  ... 
doi:10.1080/23311916.2020.1718821 fatcat:tote7arpjraebljxezox5zledy

Structure Discovery in Nonparametric Regression through Compositional Kernel Search [article]

David Duvenaud, James Robert Lloyd, Roger Grosse, Joshua B. Tenenbaum, Zoubin Ghahramani
2013 arXiv   pre-print
The learned structures can often decompose functions into interpretable components and enable long-range extrapolation on time-series datasets.  ...  Salakhutdinov & Hinton (2008) use a deep neural network to learn an embedding; this is a flexible approach to kernel learning but potentially less interpretable.  ...  Many architectures for learning complex functions, such as convolutional networks (LeCun et al., 1989) and sum-product networks (Poon & Domingos, 2011) , include units which compute AND-like and OR-like  ... 
arXiv:1302.4922v4 fatcat:qrtyoyzyajbkdnqp64sbirbfii

A Deep Learning Architecture for Image Representation, Visual Interpretability and Automated Basal-Cell Carcinoma Cancer Detection [chapter]

Angel Alfonso Cruz-Roa, John Edison Arevalo Ovalle, Anant Madabhushi, Fabio Augusto González Osorio
2013 Lecture Notes in Computer Science  
interpretability.  ...  A novel characteristic of this approach is that it extends the deep learning architecture to also include an interpretable layer that highlights the visual patterns that contribute to discriminate between  ...  This work was partially funded by "Automatic Annotation and Retrieval of Radiology Images Using Latent Semantic" project Colciencias 521/2010.  ... 
doi:10.1007/978-3-642-40763-5_50 fatcat:aqwjgw3iszfgnflzlyr4ve5qke

Deep materials informatics: Applications of deep learning in materials science

Ankit Agrawal, Alok Choudhary
2019 MRS Communications  
Deep learning Deep learning [23] refers to a family of techniques in AI and ML, and is essentially a rediscovery of neural networks that were algorithmically conceptualized back in the 1980s. [37, 38]  ...  The increasingly availability of materials databases and big data in general, along with groundbreaking advances in deep learning offers a lot of promise to accelerate the discovery, design, and deployment  ...  (ABX 3 ) with the interpretable model versus 0.099 eV/atom with the full model].  ... 
doi:10.1557/mrc.2019.73 fatcat:hzgwfbp2b5a4jlmylr34bbhxjy
« Previous Showing results 1 — 15 out of 64,128 results