Dual Accuracy-Quality-Driven Neural Network for Prediction Interval Generation.

Accurate uncertainty quantification is necessary to enhance the reliability of deep learning (DL) models in real-world applications. In the case of regression tasks, prediction intervals (PIs) should be provided along with the deterministic predictions of DL models. Such PIs are useful or "high-quality (HQ)" as long as they are sufficiently narrow and capture most of the probability density. In this article, we present a method to learn PIs for regression-based neural networks (NNs) automatically in addition to the conventional target predictions. In particular, we train two companion NNs: one that uses one output, the target estimate, and another that uses two outputs, the upper and lower bounds of the corresponding PI. Our main contribution is the design of a novel loss function for the PI-generation network that takes into account the output of the target-estimation network and has two optimization objectives: minimizing the mean PI width and ensuring the PI integrity using constraints that maximize the PI probability coverage implicitly. Furthermore, we introduce a self-adaptive coefficient that balances both objectives within the loss function, which alleviates the task of fine-tuning. Experiments using a synthetic dataset, eight benchmark datasets, and a real-world crop yield prediction dataset showed that our method was able to maintain a nominal probability coverage and produce significantly narrower PIs without detriment to its target estimation accuracy when compared to those PIs generated by three state-of-the-art neural-network-based methods. In other words, our method was shown to produce higher quality PIs.

Improved Yield Prediction of Winter Wheat Using a Novel Two-Dimensional Deep Regression Neural Network Trained via Remote Sensing.

In recent years, the use of remotely sensed and on-ground observations of crop fields, in conjunction with machine learning techniques, has led to highly accurate crop yield estimations. In this work, we propose to further improve the yield prediction task by using Convolutional Neural Networks (CNNs) given their unique ability to exploit the spatial information of small regions of the field. We present a novel CNN architecture called Hyper3DNetReg that takes in a multi-channel input raster and, unlike previous approaches, outputs a two-dimensional raster, where each output pixel represents the predicted yield value of the corresponding input pixel. Our proposed method then generates a yield prediction map by aggregating the overlapping yield prediction patches obtained throughout the field. Our data consist of a set of eight rasterized remotely-sensed features: nitrogen rate applied, precipitation, slope, elevation, topographic position index (TPI), aspect, and two radar backscatter coefficients acquired from the Sentinel-1 satellites. We use data collected during the early stage of the winter wheat growing season (March) to predict yield values during the harvest season (August). We present leave-one-out cross-validation experiments for rain-fed winter wheat over four fields and show that our proposed methodology produces better predictions than five compared methods, including Bayesian multiple linear regression, standard multiple linear regression, random forest, an ensemble of feedforward networks using AdaBoost, a stacked autoencoder, and two other CNN architectures.

Extensive transcriptomic study emphasizes importance of vesicular transport in C9orf72 expansion carriers.

The majority of the clinico-pathological variability observed in patients harboring a repeat expansion in the C9orf72-SMCR8 complex subunit (C9orf72) remains unexplained. This expansion, which represents the most common genetic cause of frontotemporal lobar degeneration (FTLD) and motor neuron disease (MND), results in a loss of C9orf72 expression and the generation of RNA foci and dipeptide repeat (DPR) proteins. The C9orf72 protein itself plays a role in vesicular transport, serving as a guanine nucleotide exchange factor that regulates GTPases. To further elucidate the mechanisms underlying C9orf72-related diseases and to identify potential disease modifiers, we performed an extensive RNA sequencing study. We included individuals for whom frontal cortex tissue was available: FTLD and FTLD/MND patients with (n = 34) or without (n = 44) an expanded C9orf72 repeat as well as control subjects (n = 24). In total, 6706 genes were differentially expressed between these groups (false discovery rate [FDR] < 0.05). The top gene was C9orf72 (FDR = 1.41E-14), which was roughly two-fold lower in C9orf72 expansion carriers than in (disease) controls. Co-expression analysis revealed groups of correlated genes (modules) that were enriched for processes such as protein folding, RNA splicing, synaptic signaling, metabolism, and Golgi vesicle transport. Within our cohort of C9orf72 expansion carriers, machine learning uncovered interesting candidates associated with clinico-pathological features, including age at onset (vascular endothelial growth factor A [VEGFA]), C9orf72 expansion size (cyclin dependent kinase like 1 [CDKL1]), DPR protein levels (eukaryotic elongation factor 2 kinase [EEF2K]), and survival after onset (small G protein signaling modulator 3 [SGSM3]). Given the fact that we detected a module involved in vesicular transport in addition to a GTPase activator (SGSM3) as a potential modifier, our findings seem to suggest that the presence of a C9orf72 repeat expansion might hamper vesicular transport and that genes affecting this process may modify the phenotype of C9orf72-linked diseases.