اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدNeural Network Attribution Methods for Problems in Geoscience: A Novel Synthetic Benchmark Dataset
81
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Despite the increasingly successful application of neural networks to many problems in the geosciences, their complex and nonlinear structure makes the interpretation of their predictions difficult, which limits model trust and does not allow scientists to gain physical insights about the problem at hand. Many different methods have been introduced in the emerging field of eXplainable Artificial Intelligence (XAI), which aim at attributing the networks prediction to specific features in the input domain. XAI methods are usually assessed by using benchmark datasets (like MNIST or ImageNet for image classification), or through deletion/insertion techniques. In either case, however, an objective, theoretically-derived ground truth for the attribution is lacking, making the assessment of XAI in many cases subjective. Also, benchmark datasets for problems in geosciences are rare. Here, we provide a framework, based on the use of additively separable functions, to generate attribution benchmark datasets for regression problems for which the ground truth of the attribution is known a priori. We generate a long benchmark dataset and train a fully-connected network to learn the underlying function that was used for simulation. We then compare estimated attribution heatmaps from different XAI methods to the ground truth in order to identify examples where specific XAI methods perform well or poorly. We believe that attribution benchmarks as the ones introduced herein are of great importance for further application of neural networks in the geosciences, and for accurate implementation of XAI methods, which will increase model trust and assist in discovering new science.
قيم البحث
اقرأ أيضاً
The space of synthesizable molecules is greater than $10^{60}$, meaning only a vanishingly small fraction of these molecules have ever been realized in the lab. In order to prioritize which regions of this space to explore next, synthetic chemists ne
ed access to accurate molecular property predictions. While great advances in molecular machine learning have been made, there is a dearth of benchmarks featuring properties that are useful for the synthetic chemist. Focussing directly on the needs of the synthetic chemist, we introduce the Photoswitch Dataset, a new benchmark for molecular machine learning where improvements in model performance can be immediately observed in the throughput of promising molecules synthesized in the lab. Photoswitches are a versatile class of molecule for medical and renewable energy applications where a molecules efficacy is governed by its electronic transition wavelengths. We demonstrate superior performance in predicting these wavelengths compared to both time-dependent density functional theory (TD-DFT), the incumbent first principles quantum mechanical approach, as well as a panel of human experts. Our baseline models are currently being deployed in the lab as part of the decision process for candidate synthesis. It is our hope that this benchmark can drive real discoveries in photoswitch chemistry and that future benchmarks can be introduced to pivot learning algorithm development to benefit more expansive areas of synthetic chemistry.
In this paper, we introduce MedLane -- a new human-annotated Medical Language translation dataset, to align professional medical sentences with layperson-understandable expressions. The dataset contains 12,801 training samples, 1,015 validation sampl
es, and 1,016 testing samples. We then evaluate one naive and six deep learning-based approaches on the MedLane dataset, including directly copying, a statistical machine translation approach Moses, four neural machine translation approaches (i.e., the proposed PMBERT-MT model, Seq2Seq and its two variants), and a modified text summarization model PointerNet. To compare the results, we utilize eleven metrics, including three new measures specifically designed for this task. Finally, we discuss the limitations of MedLane and baselines, and point out possible research directions for this task.
In recent years, deep neural networks have achieved great success in the field of computer vision. However, it is still a big challenge to deploy these deep models on resource-constrained embedded devices such as mobile robots, smart phones and so on
. Therefore, network compression for such platforms is a reasonable solution to reduce memory consumption and computation complexity. In this paper, a novel channel pruning method based on genetic algorithm is proposed to compress very deep Convolution Neural Networks (CNNs). Firstly, a pre-trained CNN model is pruned layer by layer according to the sensitivity of each layer. After that, the pruned model is fine-tuned based on knowledge distillation framework. These two improvements significantly decrease the model redundancy with less accuracy drop. Channel selection is a combinatorial optimization problem that has exponential solution space. In order to accelerate the selection process, the proposed method formulates it as a search problem, which can be solved efficiently by genetic algorithm. Meanwhile, a two-step approximation fitness function is designed to further improve the efficiency of genetic process. The proposed method has been verified on three benchmark datasets with two popular CNN models: VGGNet and ResNet. On the CIFAR-100 and ImageNet datasets, our approach outperforms several state-of-the-art methods. On the CIFAR-10 and SVHN datasets, the pruned VGGNet achieves better performance than the original model with 8 times parameters compression and 3 times FLOPs reduction.
Optical neural networks, employing optical fields and photonic tools to perform artificial neural network computations, are rapidly advancing and are generating a broad interest and sparking new applications. We propose a nascent approach for realizi
ng the optical neural network utilizing a single resonator network, where the arrival times of optical pulses are interconnected to construct a synthetic temporal dimension. The set of pulses in each roundtrip therefore provides the sites in each layer in the optical neural network, and can be linearly transformed with splitters and delay lines, including the phase modulators, when pulses circulate inside the network. Such linear transformation can be arbitrarily controlled by applied modulation phases, which serve as the building block of the neural network together with a nonlinear component for pulses. We validate the functionality of the proposed optical neural network using an example of a complicated wine classification problem. This proof of principle demonstration opens up an opportunity to develop a photonics-based machine learning in a single ring network utilizing the concept of synthetic dimensions. Our approach holds flexibility and easiness of reconfiguration with potentially complex functionality in achieving desired optical tasks, pointing towards promisingly perform on-chip optical computations with further miniaturization.
The review is conceived to help the reader to interpret present geoneutrino results in the framework of Earths energetics and composition. Starting from the comprehension of antineutrino production, propagation, and detection, the status of the geone
utrino field is presented through the description of the experimental and technological features of the Borexino and KamLAND ongoing experiments. The current understanding of the energetical, geophysical and geochemical traits of our planet is examined in a critical analysis of the currently available models. By combining theoretical models and experimental results, the mantle geoneutrino signal extracted from the results of the two experiments demonstrates the effectiveness in investigating deep earth radioactivity through geoneutrinos from different sites. The obtained results are discussed and framed in the puzzle of the diverse classes of formulated Bulk Silicate Earth models, analyzing their implications on planetary heat budget and composition. As a final remark, we present the engaging technological challenges and the future experiments envisaged for the next decade in the geoneutrino field.
الأسئلة المقترحة
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
