DantEzio

DRL_GR_UDS

Combine DRL and GR for UDS real-time control

DRL_state_selection_cost

Interpretable-DRL-for-UDS

The second version of basic system

Knowledge-Guided_UDS_DRL

A DRL training framework based on knowledge of UDS RTC for flooding and CSO mitigation.

Koopman-Emulator-RL

Use Koopman emulator to train RLs

Koopman-RL-in-UDS

A model-free framework based on Koopman emulator for reinforcement learning in the context of UDS

Koopman_Measure_WQ

Data similarity strongly affects the generalization of DL models. The first step to validate this is to establish a measurement of the similarity of dynamic data.

NN_PDE

This project try to design an algorithm to solve PDE based on neural network. The basic theory is from Koopman operator theory and some research work of using neural network to solve PDE. All the reference paper can be found in SCI.

Seq_pred

ANN, RNN and LSTM were established to predict a time series data, Chl-a concentration over time. With the help of Transfer learning, these models were able to applied for a long period predict

WQ_Pre

Applying machine learning method on water quality prediction. Developed by Wenchong Tian and Xuan Wang in Tongji University

DantEzio/Koopman-Emulator-RL

Use Koopman emulator to train RLs

DantEzio/Seq_pred

ANN, RNN and LSTM were established to predict a time series data, Chl-a concentration over time. With the help of Transfer learning, these models were able to applied for a long period predict

DantEzio/NN_PDE

This project try to design an algorithm to solve PDE based on neural network. The basic theory is from Koopman operator theory and some research work of using neural network to solve PDE. All the reference paper can be found in SCI.

DantEzio/WQ_Pre

Applying machine learning method on water quality prediction. Developed by Wenchong Tian and Xuan Wang in Tongji University

DantEzio/Knowledge-Guided_UDS_DRL

A DRL training framework based on knowledge of UDS RTC for flooding and CSO mitigation.

DantEzio/Koopman_Measure_WQ

Data similarity strongly affects the generalization of DL models. The first step to validate this is to establish a measurement of the similarity of dynamic data.

DantEzio/DRL_GR_UDS

Combine DRL and GR for UDS real-time control

DantEzio/DRL_state_selection_cost

DantEzio/Interpretable-DRL-for-UDS

The second version of basic system

DantEzio/Koopman-RL-in-UDS

A model-free framework based on Koopman emulator for reinforcement learning in the context of UDS

DantEzio/Koopman_AAO_Dynamic

Use Koopman operator and deep learning for AAO dynamics prediction

DantEzio/Koopman_Lyapunov_Control

DantEzio/Multi-RL-Voting

Use different RLs for CSO and flooding mitigation

DantEzio/MultiBert_NLP

Use different bert NLP

DantEzio/NLP_Car_requirement_Trans

Use Transformer to build an NLP model to translate the requirement of car function into car designing

DantEzio/safe-RL-in-UDS