A comparison review of transfer learning and self-supervised learning: Definitions, applications, advantages and limitations

Expert Systems with Applications, Journal Year: 2023, Volume and Issue: 242, P. 122807 - 122807

Published: Dec. 2, 2023

Deep learning has emerged as a powerful tool in various domains, revolutionising machine research. However, one persistent challenge is the scarcity of labelled training data, which hampers performance and generalisation deep models. To address this limitation, researchers have developed innovative methods to overcome data enhance model capabilities. Two prevalent techniques that gained significant attention are transfer self-supervised learning. Transfer leverages knowledge learned from pre-training on large-scale dataset, such ImageNet, applies it target task with limited data. This approach allows models benefit representations effectively new tasks, resulting improved generalisation. On other hand, focuses using pretext tasks do not require manual annotation, allowing them learn valuable large amounts unlabelled These can then be fine-tuned for downstream mitigating need extensive In recent years, found applications fields, including medical image processing, video recognition, natural language processing. approaches demonstrated remarkable achievements, enabling breakthroughs areas disease diagnosis, object understanding. while these offer numerous advantages, they also limitations. For example, may face domain mismatch issues between requires careful design ensure meaningful representations. review paper explores fields within past three years. It delves into advantages limitations each approach, assesses employing techniques, identifies potential directions future By providing comprehensive current methods, article offers guidance selecting best technique specific issue.

Language: Английский

---

Deep learning for water quality

Nature Water, Journal Year: 2024, Volume and Issue: 2(3), P. 228 - 241

Published: March 12, 2024

Language: Английский

---

Improving daily streamflow simulations for data-scarce watersheds using the coupled SWAT-LSTM approach

Journal of Hydrology, Journal Year: 2023, Volume and Issue: 622, P. 129734 - 129734

Published: May 30, 2023

Language: Английский

---

Prediction modelling framework comparative analysis of dissolved oxygen concentration variations using support vector regression coupled with multiple feature engineering and optimization methods: A case study in China

Ecological Indicators, Journal Year: 2023, Volume and Issue: 146, P. 109845 - 109845

Published: Jan. 2, 2023

Dissolved oxygen (DO) is an essential indicator for assessing water quality and managing aquatic environments, but it still a challenging topic to accurately understand predict the spatiotemporal variation of DO concentrations under complex effects different environmental factors. In this study, practical prediction framework was proposed based on support vector regression (SVR) model coupling multiple intelligence techniques (i.e., four data denoising techniques, three feature selection rules, hyperparameter optimization methods). The holistic tested using matrix (17,532 observation in total) 12 indicators from vital monitoring stations longest inter-basin diversion project world Middle-Route South-to-North Water Diversion Project China), during year 2017 2020 period. results showed that we advocated could successfully concentration variations geographical locations. used "wavelet analysis–LASSO regression–random search–SVR" combination Waihuanhe station has best performance, with Root Mean Square Error (RMSE), (MSE), Absolute (MAE), coefficient determination (R2) values 0.251, 0.063, 0.190, 0.911, respectively. combined methods can significantly promote robustness accuracy provide new universal way investigating understanding drivers variations. For management department, comprehensive also identify reveal key parameters should be concerned monitored factors change. More studies terms potential integrated risk multi-indicators mega projects and/or similar bodies are required future.

Language: Английский

---

Handling missing data in near real-time environmental monitoring: A system and a review of selected methods

Future Generation Computer Systems, Journal Year: 2021, Volume and Issue: 128, P. 63 - 72

Published: Oct. 11, 2021

High-frequency water quality monitoring systems provide valuable measurements for predicting the trend of quality, warning abnormal activities or operating hydrological models. However, missing values are prevalent due to network miscommunication, device replacement failure. Applying datasets with can lead biased results in statistical analysis modelling work. We develop a cloud-based data processing system combining advanced algorithms impute near real-time. The provides high compatibility supporting different variables, imputation and extensive scalability support numerous streams. Based on proposed approach, we review various methods which be applied data. Overall, this work systematic design system, explores advantages limitations selected analyses performance two real-time located both USA Australia. practical guidelines applications

Language: Английский

---

Prediction model of drinking water source quality with potential industrial-agricultural pollution based on CNN-GRU-Attention

Journal of Hydrology, Journal Year: 2022, Volume and Issue: 610, P. 127934 - 127934

Published: May 17, 2022

Language: Английский

---

A comparative study of data-driven models for runoff, sediment, and nitrate forecasting

Journal of Environmental Management, Journal Year: 2023, Volume and Issue: 341, P. 118006 - 118006

Published: May 8, 2023

Language: Английский

---

Multi-sensor and multi-platform retrieval of water chlorophyll a concentration in karst wetlands using transfer learning frameworks with ASD, UAV, and Planet CubeSate reflectance data

The Science of The Total Environment, Journal Year: 2023, Volume and Issue: 901, P. 165963 - 165963

Published: Aug. 4, 2023

Language: Английский

---

A data-driven model for water quality prediction in Tai Lake, China, using secondary modal decomposition with multidimensional external features

Journal of Hydrology Regional Studies, Journal Year: 2023, Volume and Issue: 47, P. 101435 - 101435

Published: May 30, 2023

Tai Lake, the third largest freshwater lake in China, with a history of serious ecological pollution incidents. Lake water quality prediction techniques are essential to ensure an early emergency response capability for sustainable management. Herein, effective data-driven ensemble model was developed predicting dissolved oxygen (DO) based on meteorological factors, indicators and spatial information. First, variation mode decomposition (VMD) used decompose data into multiple modal components classify them feature terms self terms. The were combined relevant external features multivariate by convolutional neural network (CNN) bi-directional long short-term memory (BiLSTM) attention mechanism (AT), as well using whale optimization algorithm (WOA) optimize hyperparameters. form secondary model. Finally, groupings linearly summed obtain outcome. proposed has highest accuracy best effect 0.5 days period. This research also establishes stepwise temperature regulation mechanism, where output target DO content value is achieved changing magnitude combining it this model, thereby strengthening protection resources management fishery production.

Language: Английский

---

Long short-term memory models of water quality in inland water environments

Water Research X, Journal Year: 2023, Volume and Issue: 21, P. 100207 - 100207

Published: Nov. 16, 2023

Water quality is substantially influenced by a multitude of dynamic and interrelated variables, including climate conditions, landuse seasonal changes. Deep learning models have demonstrated predictive power water due to the superior ability automatically learn complex patterns relationships from variables. Long short-term memory (LSTM), one deep for prediction, type recurrent neural network that can account longer-term traits time-dependent data. It most widely applied used predict time series First, we reviewed applications standalone LSTM discussed its calculation time, prediction accuracy, good robustness with process-driven numerical other machine learning. This review was expanded into model data pre-processing techniques, Complete Ensemble Empirical Mode Decomposition Adaptive Noise method Synchrosqueezed Wavelet Transform. The then focused on coupling convolutional network, attention transfer coupled networks their performance over model. We also emphasized influence static variables in transformation dataset. Outlook further challenges were addressed. outlook research application hydrology concludes review.

Language: Английский

---