Explainable Deep Learning-Assisted Fluorescence Discrimination for Aminoglycoside Antibiotic Identification

Abstract

The complexity and multivariate analysis of biological systems and environment are the drawbacks of the current high-throughput sensing method and multianalyte identification. Deep learning (DL) algorithms contribute a big advantage in analyzing the nonlinear and multidimensional data. However, most DL models are datadriven black boxes suffering from nontransparent inner workings. In this work, we developed an explainable DL-assisted visualized fluorometric array-based sensing method. Based on a data set of 8496 fluorometric images of various target molecule fingerprint patterns, two typical DL algorithms and eight machine learning algorithms were investigated for the efficient qualitative and quantitative analysis of six aminoglycoside antibiotics (AGs). The convolutional neural network (CNN) approached 100% prediction accuracy and 1.34 ppm limit of detection of six AG analysis in domestic, industrial, medical, consumption, or aquaculture water. The class activation mapping assessment explicates how the CNN model assesses the importance of sensor elements and makes the discrimination decision. The feedback mechanism guides the sensor array evolution for less material using a simplified operation or efficient data acquisition. The explainable DL-assisted analysis method establishes an "end-to-end" strategy to resolve the black box of the DL algorithm, promote hardware design or principle optimization, and contribute facile indicators for environment monitoring, disease diagnosis, and even new scientific discovery.

Multi-analyte sensing strategies towards wearable and intelligent devices

“…9a). 128,129 For example, classification of six aminoglycoside antibiotics (AGs) were performed through three-dimensional LDA firstly, as shown in Fig. 9b.…”

Explainable Deep Learning-Assisted Photochromic Sensor for β-Lactam Antibiotic Identification

“…The explanatory artificial intelligence was recently achieved to point out the connection between input and output for the explanation of DL black-boxes' inner structure. 23,24,48,49 It offers feasibility to reveal the relationship between a DL model and chemical sensing information, an end-to-end approach.…”

“…Different targets can be identified by analyzing the sensing patterns with DL-assisted pattern recognition methods. ,,,, Due to massive and complicated parameter networks, most DL models are faced with a black-box dilemma: the nontransparency neural networks. The explanatory artificial intelligence was recently achieved to point out the connection between input and output for the explanation of DL black-boxes’ inner structure. ,,, It offers feasibility to reveal the relationship between a DL model and chemical sensing information, an end-to-end approach.…”

Explainable Deep-Learning-Assisted Sweat Assessment via a Programmable Colorimetric Chip

“…The "black box" may cause risk to the dilemma on how the network makes decision. 20 Empowering DL models to explain the connections between inputs and outputs will significantly help users to understand why accurate decision-making mechanisms are required and guide the sensor optimization. 21 In this study, we integrated the advantages of programmable colorimetric sensing and DL algorithm to develop a DLassisted programmable chip for the colorimetric detection of sweat biomarkers.…”

“…Although the application of AI in chemistry has made amazing progress, most DL models are facing an inexplicable dilemma: the inner working mechanisms of neural networks are still opaque. The “black box” may cause risk to the dilemma on how the network makes decision . Empowering DL models to explain the connections between inputs and outputs will significantly help users to understand why accurate decision-making mechanisms are required and guide the sensor optimization …”