216 Hits in 5.9 sec

Embrace the Gap: VAEs Perform Independent Mechanism Analysis [article]

Patrik Reizinger, Luigi Gresele, Jack Brady, Julius von Kügelgen, Dominik Zietlow, Bernhard Schölkopf, Georg Martius, Wieland Brendel, Michel Besserve
2022 arXiv   pre-print
This allows VAEs to perform what has recently been termed independent mechanism analysis (IMA): it adds an inductive bias towards decoders with column-orthogonal Jacobians, which helps recovering the true  ...  at the expense of a gap to the exact (log-)marginal likelihood.  ...  More interestingly, the authors derive a formula for the ELBO gap in the linear case that is remarkably similar to the IMA objective. We show in Appx.  ... 
arXiv:2206.02416v1 fatcat:xzszp3m4dja3zcrcjl6hngo7iy

Variational Mutual Information Maximization Framework for VAE Latent Codes with Continuous and Discrete Priors [article]

Andriy Serdega, Dae-Shik Kim
2020 arXiv   pre-print
The objective acts as a regularizer that forces VAE to not ignore the latent variable and allows one to select particular components of it to be most informative with respect to the observations.  ...  On top of that, the proposed framework provides a way to evaluate mutual information between latent codes and observations for a fixed VAE model.  ...  We believe, that our work (with further analysis and improvements) have the potential to fill the gaps between previous theoretical insights for VAE from Information Theory perspective since ours empirical  ... 
arXiv:2006.02227v1 fatcat:trduf6cokfh5xeqx3t3xnqlnfi

Deep clustering with fusion autoencoder [article]

Shuai Chang
2022 arXiv   pre-print
Nevertheless, the plain VAE is insufficient to perceive the comprehensive latent features, leading to the deteriorative clustering performance.  ...  Embracing the deep learning techniques for representation learning in clustering research has attracted broad attention in recent years, yielding a newly developed clustering paradigm, viz. the deep clustering  ...  [27] aim to improve the performance of VAE by using deep-wise feature reconstruction loss for capturing perceptual features and deploying GAN training mechanism to keep the generated images on the manifold  ... 
arXiv:2201.04727v2 fatcat:glfdqqyfjvdpvor7f3tslegjqe

On Disentangled Representations Learned From Correlated Data [article]

Frederik Träuble, Elliot Creager, Niki Kilbertus, Francesco Locatello, Andrea Dittadi, Anirudh Goyal, Bernhard Schölkopf, Stefan Bauer
2021 arXiv   pre-print
The focus of disentanglement approaches has been on identifying independent factors of variation in data.  ...  In this work, we bridge the gap to real-world scenarios by analyzing the behavior of the most prominent disentanglement approaches on correlated data in a large-scale empirical study (including 4260 models  ...  To make our conclusion more sound we perform an empirical analysis of the pairwise entanglement metrics for the correlated pair vs. the median of all other pairs across the entire unsupervised study on  ... 
arXiv:2006.07886v3 fatcat:irkinvphx5drdep6fe2kul2iby

Hyperrealistic neural decoding: Linear reconstruction of face stimuli from fMRI measurements via the GAN latent space [article]

Thirza Dado, Yagmur Gucluturk, Luca Ambrogioni, Gabrielle Ras, Sander Erik Bosch, Marcel van Gerven, Umut Guclu
2020 bioRxiv   pre-print
To this end, we embrace the use of generative adversarial networks (GANs) at the earliest step of our neural decoding pipeline by acquiring functional magnetic resonance imaging data as subjects perceived  ...  Subsequently, we used a linear decoding approach to predict the latent state of the GAN from brain data.  ...  Next, we compared the performance of the HYPER framework to the state-of-the-art VAE-GAN approach [19] and the eigenface approach [3] .  ... 
doi:10.1101/2020.07.01.168849 fatcat:3ukydl3l5rbgdnjqethqueajki

Listener Modeling and Context-Aware Music Recommendation Based on Country Archetypes

Markus Schedl, Christine Bauer, Wolfgang Reisinger, Dominik Kowald, Elisabeth Lex
2021 Frontiers in Artificial Intelligence  
More precisely, it is a multi-layer generative model based on a variational autoencoder, in which contextual features can influence recommendations through a gating mechanism.  ...  To complement and extend these results, the article at hand delivers the following major contributions: First, using state-of-the-art unsupervized learning techniques, we identify and thoroughly investigate  ...  ACKNOWLEDGMENTS The authors would like to thank Peter Müllner from the Know-Center GmbH for providing the IDF calculations of the music tracks.  ... 
doi:10.3389/frai.2020.508725 pmid:33778483 pmcid:PMC7990101 fatcat:ddemjb7owfgnvnkgbmsg3wkgce

Deep Human-guided Conditional Variational Generative Modeling for Automated Urban Planning [article]

Dongjie Wang, Kunpeng Liu, Pauline Johnson, Leilei Sun, Bowen Du, Yanjie Fu
2021 arXiv   pre-print
Finally, we present extensive experiments to validate the enhanced performances of our method.  ...  To mitigate training data sparsity and improve model robustness, we introduce a variational Gaussian embedding mechanism.  ...  Inspired by [4] , the multi-head decoder in VAE can improve model performance by adding certain constraints.  ... 
arXiv:2110.07717v1 fatcat:mlzo5oqu4ze6zomb3xtjmpy5dy

Data augmentation for time series regression: Applying transformations, autoencoders and adversarial networks to electricity price forecasting

Sumeyra Demir, Krystof Mincev, Koen Kok, Nikolaos G. Paterakis
2021 Applied Energy  
published version features the final layout of the paper including the volume, issue and page numbers.  ...  If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the "Taverne" license above, please follow below link for the End User  ...  All authors approved the version of the manuscript to be published. Acronyms  ... 
doi:10.1016/j.apenergy.2021.117695 fatcat:iyubwp4grneyzkbdlcmwqq5yfu

Reconsidering Generative Objectives For Counterfactual Reasoning

Danni Lu, Chenyang Tao, Junya Chen, Fan Li, Feng Guo, Lawrence Carin
2020 Neural Information Processing Systems  
By appealing to the Robinson decomposition, we derive a reformulated variational bound that explicitly targets the causal effect estimation rather than specific predictive goals.  ...  Our procedure acknowledges the uncertainties in representation and solves a Fenchel mini-max game to resolve the representation imbalance for better counterfactual generalization, justified by new theory.The  ...  The work at Virginia Tech was supported by the National Surface Transportation Safety Center for Excellence.  ... 
dblp:conf/nips/LuTCLGC20 fatcat:6tr7m7datnfxvgvsaml3ql3zfy

Materials Representation and Transfer Learning for Multi-Property Prediction [article]

Shufeng Kong, Dan Guevarra, Carla P. Gomes, John M. Gregoire
2021 arXiv   pre-print
The best performance is obtained using H-CLMP with Transfer learning (H-CLMP(T)) wherein a generative adversarial network is trained on computational density of states data and deployed in the target domain  ...  Hurdles limiting full capitalization of recent advancements in machine learning include the limited development of methods to learn the underlying interactions of multiple elements, as well as the relationships  ...  RESULTS The presentation of results commences with a description of the H-CLMP model and introduction of the settings for which it is deployed, followed by analysis of multi-target regression performance  ... 
arXiv:2106.02225v3 fatcat:cwclgy534rglxpwvsarzib3c6i

Interpretable and Explainable Machine Learning for Materials Science and Chemistry [article]

Felipe Oviedo, Juan Lavista Ferres, Tonio Buonassisi, Keith Butler
2021 arXiv   pre-print
While the uptake of data-driven approaches for materials science and chemistry is at an exciting, early stage, to realise the true potential of machine learning models for successful scientific discovery  ...  The predictions and inner workings of models should provide a certain degree of explainability by human experts, permitting the identification of potential model issues or limitations, building trust on  ...  This worked was supported by the National Research Foundation (NRF), the Singapore Massachusetts Institute of Technology (MIT) Alliance for Research and Technology's Low Energy Electronic Systems research  ... 
arXiv:2111.01037v2 fatcat:hirrciqrlbfcfdmo4bi4sv4lxm

Causal Representation Learning for Out-of-Distribution Recommendation

Wenjie Wang, Xinyu Lin, Fuli Feng, Xiangnan He, Min Lin, Tat-Seng Chua
2022 Proceedings of the ACM Web Conference 2022  
Towards the learning objectives, we embrace causal modeling of the generation procedure from user features to interactions.  ...  We further perform counterfactual inference to mitigate the effect of out-of-date interactions.  ...  Once learned, we can perform post-intervention inference by feeding the latest user features e ′ 1 to the VAE.  ... 
doi:10.1145/3485447.3512251 fatcat:zqbzdbqbevgrrcyp7oxo2u6l7y

Data Requirements for Electronic Surveillance of Healthcare-Associated Infections

Keith F. Woeltje, Michael Y. Lin, Michael Klompas, Marc Oliver Wright, Gianna Zuccotti, William E. Trick
2014 Infection control and hospital epidemiology  
This white paper reviews different approaches to electronic surveillance, discusses the specific data elements required for performing surveillance, and considers important issues of data validation.  ...  Optimal implementation of electronic surveillance requires that specific information be available to the surveillance systems.  ...  There are also important gaps in the current knowledge of how best to implement electronic surveillance systems.  ... 
doi:10.1086/677623 pmid:25111915 fatcat:n3k4kaykerd2djyroqzd2xakza

Machine-Learning-Assisted De Novo Design of Organic Molecules and Polymers: Opportunities and Challenges

Guang Chen, Zhiqiang Shen, Akshay Iyer, Umar Farooq Ghumman, Shan Tang, Jinbo Bi, Wei Chen, Ying Li
2020 Polymers  
We anticipate that ML-assisted materials design for organic molecules and polymers will be the driving force in the near future, to meet the tremendous demand of new materials with tailored properties  ...  We further discuss the relevant challenges to solve in order to fully realize the potential of ML-assisted materials design for organic molecules and polymers.  ...  Author Contributions: Y.L. and W.C. conceived and designed this study; G.C. and Z.S. performed the case studies on ML-assisted materials design and discussions; A.I. wrote the BO approach for inverse materials  ... 
doi:10.3390/polym12010163 pmid:31936321 pmcid:PMC7023065 fatcat:rptoxcbvsfg4toacpbxedwnq5m

Introduction to machine and deep learning for medical physicists

Sunan Cui, Huan‐Hsin Tseng, Julia Pakela, Randall K. Ten Haken, Issam El Naqa
2020 Medical Physics (Lancaster)  
Data processing, which is a crucial step for model stability and precision, should be performed before training the model.  ...  Embracing the current big data era, medical physicists equipped with these state-of-the-art tools should be able to solve pressing problems in modern radiation oncology.  ...  ACKNOWLEDGMENTS This work was supported in part by the National Institutes of  ... 
doi:10.1002/mp.14140 pmid:32418339 fatcat:b6jc2fta6zdp7pv7bjqw2st6zm
« Previous Showing results 1 — 15 out of 216 results