This paper is concerned with a state-space approach to deep Gaussian process (DGP) regression. We construct the DGP by hierarchically putting transformed Gaussian process (GP) priors on the length scales and magnitudes of the next level of Gaussian processes in the hierarchy. The idea of the st...
In this work, we present DANCE as a deep learning library and benchmark platform to facilitate research and development for single-cell analysis. DANCE provides an end-to-end toolkit to facilitate single-cell analysis algorithm development and fair performance comparison on different benchmark datasets...
In comparison with traditional “shallow” techniques, deep learning has the ability to learn highly non-linear, complex relationships and correlations between the input and output data. For that reason, in the DR literature it is shown that deep learning methods usually outperform in prediction acc...
To make a head-to-head comparison with ProtTucker and these other methods, we trained a TM-Vec model on the same domains as ProtTucker’s model and evaluated TM-Vec on their test set of 219 domains (ProtTucker benchmark data). Across each level of the CATH hierarchy, TM-Vec ...
In providing a code capsule that reproduces this study, we hope to facilitate the development, testing and comparison of related methods. In particular, the development of interpretable (as opposed to ‘black box’) models that achieve a similar performance would be highly desirable, especially in...
deep fuzzy systems in two ways: (i) standard deep fuzzy systems, based on fundamental FLS; and (ii) hybrid deep fuzzy systems, with the combination of FLS and the conventional deep models discussed in Sect.3. Finally, Sect.5presents general discussions based on the state-of-the-art ...
DF-Net: Unsupervised Joint Learning of Depth and Flow using Cross-Task Consistency ECCV 2018 LineNet: a Zoomable CNN for Crowdsourced High Definition Maps Modeling in Urban Environments [mapping] Road-SLAM: Road Marking based SLAM with Lane-level Accuracy [Notes] [HD mapping] AVP-SLAM: Semantic...
combination methods have better prospects in the future. This inference is supported by the trend shown inFig. 6. Here we can see that during the years 2016-19, there is a shift towards deep learning and combination methods as a preferred approach in comparison with handcrafted feature-based ...
Illustration of the comparison between the pipelines of the conventional machine learning and deep learning Full size image Fortunately, due to the great success of deep learning in many vision tasks, hand-crafted features are no longer required to derive promising results. Instead, deep representation...
and stochastic interventions. We refer the interested reader to Supplementary section3for a definition of an intervention and the types used. The test results (AUC and AUPRC values) are summarized in Supplementary Tables8and9and support the notion that D2CL is robust to different types of noise...