A review of modularization techniques in artificial neural networks

Predictive modelling of methane yield in biochar-amended cheese whey and septage co-digestion: Exploring synergistic effects using Gompertz and neural networks

This study performed bench scale studies on anaerobic co-digestion of cheese whey and septage mixed with biochar (BC) as additive at various dosages (0.5 g, 1 g, 2 g and 4 g) and total solids (TS) concentrations (5%, 7.5%, 10%,12.5% and 15%). The experimental results revealed 29.58% increase in methane yield (486 ± 11.32 mL/gVS) with 27% reduction in lag phase time at 10% TS concentration and 50 g/L of BC loading. The mechanistic investigations revealed that BC improved process stability by virtue of its robust buffering capacity and mitigated ammonia inhibition. Statistical analysis indicates BC dosage had a more pronounced effect (P < 0.0001) compared to the impact of TS concentrations. Additionally, the results were modelled using Gompertz model (GM) and artificial neural network (ANN) algorithm, which revealed the outperformance of ANN over GM with MSE 17.96, R2 value 0.9942 and error 0.27%. These findings validated the practicality of utilizing a high dosage of BC in semi-solid anaerobic digestion conditions.

Computer vision for plant pathology: A review with examples from cocoa agriculture

Plant pathogens can decimate crops and render the local cultivation of a species unprofitable. In extreme cases this has caused famine and economic collapse. Timing is vital in treating crop diseases, and the use of computer vision for precise disease detection and timing of pesticide application is gaining popularity. Computer vision can reduce labour costs, prevent misdiagnosis of disease, and prevent misapplication of pesticides. Pesticide misapplication is both financially costly and can exacerbate pesticide resistance and pollution. Here, we review the application and development of computer vision and machine learning methods for the detection of plant disease. This review goes beyond the scope of previous works to discuss important technical concepts and considerations when applying computer vision to plant pathology. We present new case studies on adapting standard computer vision methods and review techniques for acquiring training data, the use of diagnostic tools from biology, and the inspection of informative features. In addition to an in-depth discussion of convolutional neural networks (CNNs) and transformers, we also highlight the strengths of methods such as support vector machines and evolved neural networks. We discuss the benefits of carefully curating training data and consider situations where less computationally expensive techniques are advantageous. This includes a comparison of popular model architectures and a guide to their implementation.

Modelling and optimization of biodiesel production from waste fish oil using nano immobilized rPichiapastoris whole cell biocatalyst with response surface methodology and hybrid artificial neural network based approach

In this study, zinc oxide (ZnO) nano particle immobilized recombinant whole cell biocatalyst (rWCB) was used for bioconversion of waste fish oil in to biodiesel in a lab scale packed bed reactor (PBR). Central composite design and hybrid artificial neural network (ANN) models were explored to optimize the production of biodiesel. Developed rWCB exhibited maximum lipase activity at 15 % (v/v) of glutaraldehyde concentration and 6 % (w/v) of ZnO nanoparticles at pH of 7. Maximum biodiesel yield reached about 91.54 ± 1.86 % after 43 h in PBR using hybrid ANN model predicted process conditions of 13.2 % (w/v) of nano immobilized rWCB concentration and 4.7:1 of methanol to oil ratio at 33 °C. Importantly, developed nano immobilized rWCB was adequately stable for commercialization. Thus, production of biodiesel from waste fish oil using ZnO nano immobilized rWCB could become potential candidate for commercialization.

Miniaturized spectrometer with intrinsic long-term image memory

Miniaturized spectrometers have great potential for use in portable optoelectronics and wearable sensors. However, current strategies for miniaturization rely on von Neumann architectures, which separate the spectral sensing, storage, and processing modules spatially, resulting in high energy consumption and limited processing speeds due to the storage-wall problem. Here, we present a miniaturized spectrometer that utilizes a single SnS2/ReSe2 van der Waals heterostructure, providing photodetection, spectrum reconstruction, spectral imaging, long-term image memory, and signal processing capabilities. Interface trap states are found to induce a gate-tunable and wavelength-dependent photogating effect and a non-volatile optoelectronic memory effect. Our approach achieves a footprint of 19 μm, a bandwidth from 400 to 800 nm, a spectral resolution of 5 nm, and a > 104 s long-term image memory. Our single-detector computational spectrometer represents a path beyond von Neumann architectures.

Distinguishing examples while building concepts in hippocampal and artificial networks

The hippocampal subfield CA3 is thought to function as an auto-associative network that stores experiences as memories. Information from these experiences arrives directly from the entorhinal cortex as well as indirectly through the dentate gyrus, which performs sparsification and decorrelation. The computational purpose for these dual input pathways has not been firmly established. We model CA3 as a Hopfield-like network that stores both dense, correlated encodings and sparse, decorrelated encodings. As more memories are stored, the former merge along shared features while the latter remain distinct. We verify our model's prediction in rat CA3 place cells, which exhibit more distinct tuning during theta phases with sparser activity. Finally, we find that neural networks trained in multitask learning benefit from a loss term that promotes both correlated and decorrelated representations. Thus, the complementary encodings we have found in CA3 can provide broad computational advantages for solving complex tasks.

Distinctive properties of biological neural networks and recent advances in bottom-up approaches toward a better biologically plausible neural network

Although it may appear infeasible and impractical, building artificial intelligence (AI) using a bottom-up approach based on the understanding of neuroscience is straightforward. The lack of a generalized governing principle for biological neural networks (BNNs) forces us to address this problem by converting piecemeal information on the diverse features of neurons, synapses, and neural circuits into AI. In this review, we described recent attempts to build a biologically plausible neural network by following neuroscientifically similar strategies of neural network optimization or by implanting the outcome of the optimization, such as the properties of single computational units and the characteristics of the network architecture. In addition, we proposed a formalism of the relationship between the set of objectives that neural networks attempt to achieve, and neural network classes categorized by how closely their architectural features resemble those of BNN. This formalism is expected to define the potential roles of top-down and bottom-up approaches for building a biologically plausible neural network and offer a map helping the navigation of the gap between neuroscience and AI engineering.

Assessment of long-term mangrove distribution using optimised machine learning algorithms and landscape pattern analysis

Mangrove ecosystems provide numerous benefits, including carbon storage, coastal protection and food for marine organisms. However, mapping and monitoring of mangrove status in some regions, such as the Red Sea area, has been hindered by a lack of data, accurate and precise maps and technical expertise. In this study, an advanced machine learning algorithm was proposed to produce an accurate and precise high-resolution land use map that includes mangroves in the Al Wajh Bank habitat in northeastern Saudi Arabia. To achieve this, high-resolution multispectral images were generated using an image fusion technique, and machine learning algorithms were applied, including artificial neural networks, random forests and support vector machine algorithms. The performance of the models was evaluated using various matrices, and changes in mangrove distribution and connectivity were assessed using the landscape fragmentation model and Getis-Ord statistics. The research gap that this study aims to address is the lack of accurate and precise mapping and assessment of mangrove status in the Red Sea area, particularly in data-scarce regions. Our study produced high-resolution mobile laser scanning (MLS) imagery of 15-m length for 2014 and 2022, and trained 5, 6 and 9 models for artificial neural networks, support vector machines and random forests (RF) to predict land use and land cover maps using 15-m and 30-m resolution MLS images. The best models were identified using error matrices, and it was found that RF outperformed other models. According to the 15-m resolution map of 2022 and the best models of RF, the mangrove cover in the Al Wajh Bank is 27.6 km², which increased to 34.99 km² in the case of the 30-m resolution image of 2022, and was 11.94 km² in 2014, indicating a doubling of the mangrove area. Landscape structure analysis revealed an increase in small core and hotspot areas, which were converted into medium core and very large hotspot areas in 2014. New mangrove areas were identified in the form of patches, edges, potholes and coldspots. The connectivity model showed an increase in connectivity over time, promoting biodiversity. Our study contributes to the promotion of the protection, conservation and planting of mangroves in the Red Sea area.

Modeling and optimization of photo-fermentation biohydrogen production from co-substrates basing on response surface methodology and artificial neural network integrated genetic algorithm

The main aim of the present study was to establish a relationship model between bio-hydrogen yield and the key operating parameters affecting photo-fermentation hydrogen production (PFHP) from co-substrates. Central composite design-response surface methodology (CCD-RSM) and artificial neural network-genetic algorithm (ANN-GA) models were used to optimize the hydrogen production performance from co-substrates. Compared to CCD-RSM, the ANN-GA had higher determination coefficient (R² = 0.9785) and lower mean square error (MSE = 9.87), average percentage deviation (APD = 2.72) and error (4.3%), indicating the ANN-GA was more suitable, reliable and accurate in predicting biohydrogen yield from co-substrates by PFHP. The highest biohydrogen yield (99.09 mL/g) predicted by the ANN-GA model at substrate concentration 35.62 g/L, temperature 30.94 °C, initial pH 7.49 and inoculation ratio 32.98 %(v/v), which was 4.20 % higher than the CCD-RSM model (95.10 mL/g).

A Computational Intelligence Hybrid Algorithm Based on Population Evolutionary and Neural Network Learning for the Crude Oil Spot Price Prediction

This research attempts to reinforce the cultivating expression of radial basis function neural network (RBFnet) through computational intelligence (CI) and swarm intelligence (SI) learning methods. Consequently, the artificial immune system (AIS) and ant colony optimization (ACO) approaches are utilized to cultivate RBFnet for function approximation issue. The proposed hybridization of AIS and ACO approaches optimization (HIAO) algorithm combines the complementarity of exploitation and exploration to realize problem solving. It allows the solution domain having the advantages of intensification and diversification, which further avoids the situation of immature convergence. In addition, the empirical achievements have confirmed that the HIAO algorithm not only obtained the best accurate function approximation for theoretically standard nonlinear problems, it can be further applied on the instance solving for practical crude oil spot price prediction.

Construction and Research of Constructive English Teaching Model Applying Multimodal Neural Network Algorithm

This article draws on previous research on constructive English teaching models and uses multimodal neural network algorithm theory and constructive English teaching as the theoretical basis, experimental research method, questionnaire survey method, and evaluation method. In this article, we propose a multimodal neural network consisting of a multiscale FCN module and an LSTM module. The network focuses on both the multiscale geometric spatial features and the numerical time-dependent features of the time series curves, and with the comprehensive knowledge of their characteristics, it can better distinguish the classes to which the series belong. A large-scale perceptual field is achieved by null convolution in the model to ensure that the training pressure does not increase significantly. A series of experiments on the UCR dataset verifies the effectiveness and superiority of the model. Simulation experiments were conducted to build the proposed constructive English teaching model based on a multimodal neural network algorithm, and a test environment was built for use case testing. The experimental results showed that the algorithm can be better applied to constructive English teaching and has better adaptability and accuracy in various scenarios. At the end of the experiment, a posttest of grammar level was conducted in two classes to test whether the constructive English teaching model based on the multimodal neural network model could help students improve their English grammar skills. The results of the data analysis showed that the mean score of the experimental class was significantly higher than that of the control class, and the experimental class showed a more significant improvement, indicating that this new constructive English teaching model was beneficial to improving students' English grammar skills. The interaction strategy proposed under the constructive English teaching model can effectively improve the interaction between teachers and students. This positive feedback effect can provide front-line teachers with corresponding teaching references.

A transfer learning approach for predictive modeling of bioprocesses using small data

Predictive modeling of new biochemical systems with small data is a great challenge. To fill this gap, transfer learning, a subdomain of machine learning that serves to transfer knowledge from a generalized model to a more domain-specific model, provides a promising solution. While transfer learning has been used in natural language processing, image analysis, and chemical engineering fault detection, its application within biochemical engineering has not been systematically explored. In this study, we demonstrated the benefits of transfer learning when applied to predict dynamic behaviors of new biochemical processes. Two different case studies were presented to investigate the accuracy, reliability, and advantage of this innovative modeling approach. We thoroughly discussed the different transfer learning strategies and the effects of topology on transfer learning, comparing the performance of the transfer learning models against benchmark kinetic and data-driven models. Furthermore, strong connections between the underlying process mechanism and the transfer learning model's optimal structure were highlighted, suggesting the interpretability of transfer learning to enable more accurate prediction than a naive data-driven modeling approach. Therefore, this study shows a novel approach to effectively combining data from different resources for bioprocess simulation.

Decentralized control and local information for robust and adaptive decentralized Deep Reinforcement Learning

Decentralization is a central characteristic of biological motor control that allows for fast responses relying on local sensory information. In contrast, the current trend of Deep Reinforcement Learning (DRL) based approaches to motor control follows a centralized paradigm using a single, holistic controller that has to untangle the whole input information space. This motivates to ask whether decentralization as seen in biological control architectures might also be beneficial for embodied sensori-motor control systems when using DRL. To answer this question, we provide an analysis and comparison of eight control architectures for adaptive locomotion that were derived for a four-legged agent, but with their degree of decentralization varying systematically between the extremes of fully centralized and fully decentralized. Our comparison shows that learning speed is significantly enhanced in distributed architectures-while still reaching the same high performance level of centralized architectures-due to smaller search spaces and local costs providing more focused information for learning. Second, we find an increased robustness of the learning process in the decentralized cases-it is less demanding to hyperparameter selection and less prone to becoming trapped in poor local minima. Finally, when examining generalization to uneven terrains-not used during training-we find best performance for an intermediate architecture that is decentralized, but integrates only local information from both neighboring legs. Together, these findings demonstrate beneficial effects of distributing control into decentralized units and relying on local information. This appears as a promising approach towards more robust DRL and better generalization towards adaptive behavior.

Continual learning classification method for time-varying data space based on artificial immune system

Classification methods play an important role in many fields. However, they cannot effectively classify the samples from sample spaces that are varying with time, for they lack continual learning ability. A continual learning classification method for time-varying data space based on artificial immune system, CLCMTVD, is proposed. It is inspired by the intelligent mechanism that memory cells of the biological immune system can recognize and eliminate previous invaders when they attack again very fast and more efficiently, and these memory cells can evolve with the evolution of previous invaders. Memory cells were continuously updated by learning testing data during the testing stage, thus realize the self-improvement of classification performance. CLCMTVD changes a linearly inseparable spatial problem into many classification problems of several different times, and it degenerates into a common supervised learning classification method when all data independent of time. To assess the performance and possible advantages of CLCMTVD, the experiments on well-known datasets from UCI repository, synthetic data and XJTU-SY rolling element bearing accelerated life test datasets were performed. Results show that CLCMTVD has better classification performance for time-invariant data, and outperforms the other methods for time-varying data space.

Deep Learning Strategies for Survival Prediction in Prophylactic Resection Patients

Human race is looking forward to an era where science and technology can wipeout the threats laid by lethal diseases. Major statistics shows that about 10 million people die from various forms of cancer annually. Every sixth death in the world is caused by cancer. Treatment to cancer always depend on its type and spread. Treatment includes single or combination of surgery, chemotherapy and radiation therapy. In this paper, survival prediction in prophylactic resection patients are carried out using various deep learning methods. Prophylactic resection has been found to be very effective in colon cancer, breast cancer and ovarian cancer. In this paper, we try to validate the results in a test environment using multi layered deep neural network. Classical Navie Bayer’s algorithm has been used to classify the dataset and convolution neural network (CNN) has been used to create the survival prediction model. Results affirm better survival results in prophylactic resection patients.

Optimization of Feedforward Neural Networks Using an Improved Flower Pollination Algorithm for Short-Term Wind Speed Prediction

It is well known that the inherent instability of wind speed may jeopardize the safety and operation of wind power generation, consequently affecting the power dispatch efficiency in power systems. Therefore, accurate short-term wind speed prediction can provide valuable information to solve the wind power grid connection problem. For this reason, the optimization of feedforward (FF) neural networks using an improved flower pollination algorithm is proposed. First of all, the empirical mode decomposition method is devoted to decompose the wind speed sequence into components of different frequencies for decreasing the volatility of the wind speed sequence. Secondly, a back propagation neural network is integrated with the improved flower pollination algorithm to predict the changing trend of each decomposed component. Finally, the predicted values of each component can get into an overlay combination process and achieve the purpose of accurate prediction of wind speed. Compared with major existing neural network models, the performance tests confirm that the average absolute error using the proposed algorithm can be reduced up to 3.67%.