Open Source Software: Machine Learning, Solid Earth Evolution, and Reef Modelling

Solid Earth Evolution

1. BayesLands via MCMC (in Python)
( Download related paper )
2. Multi-Core Parallel Tempering Bayeslands (in Python)
( Download related paper )
3. Surrogate-assisted Bayeslands (in Python)
Download related paper [TBA] )
4. Bayeslands extended: region based precitipation and inference for initial topography with PNG application (in Python)
5. Continental Bayeslands: Large scale application to Australian continent (to be compelted by May 2019) (in Python)

Reef Modelling

1. BayesReef via MCMC (in Python)
( Download related paper )
2. Parallel tempering BayesReef (to be compelted by March 2019) (in Python)

Bayesian Neural Networks

1. Bayesian Neural Networks via MCMC (in Python)
2. Feeforward Neural Networks via MCMC Langevin Dynamics (in Python)
( Download related paper [3] )
3. Baysiean Multi-task learning in Neural Networks for Dynamic Time Series Prediction (in Python)
( Download related paper [TBA] )
4. Baysiean Neural Transfer learning (in Python)
( Download related paper [TBA] )

Parallel tempering for Bayesian Neural Networks

1. Parallel Tempering Baysiean Neural Networks (in Python)
( ( Download related paper )
2. Surrogate-assisted Baysiean Neural Networks (in Python)
( ( Download related paper )

Feedforward Neural Networks

1. Feeforward Neural Networks for Time Series Prediction (in Python)
2. Feeforward Neural Networks for Pattern Classification (in Python)
3. Feedforward Neural Networks in Pattern Classification (in C++)

Recurrent Neural Networks

1. Elman Recurrent Neural Networks (with BPTT) for Time Series Prediction (in C++)
( Download related paper [1] )
2. Elman Recurrent Neural Networks with Cooperative Coevolution (in C++)
( Download related paper [2] )

Bayesian Methods

1. Markov Chain Monte Carlo (MCMC) Random-Walk Sampler (in Python)
2. Approximate Bayesian Computation (in Python)
3. Parallel Tempering - Mixture Model (in Python)
4. Multi-Core Parallel Tempering - Mixture Model (in Python)

Coevolutionary Multi-task Learning (CMTL)

1. CMTL for Dynamic Time Series Prediction (in Matlab)
( Download related paper [4] )
2. CMTL for Modular Pattern Classification (in Matlab)
( Download related paper [5] )
3. CMTL with Predictive Recurrence for Multi-Step Ahead Prediction (in Matlab)
( Download related paper [6] )

Multi-task Modular Backpropataion for Misinformation

1. Modular Backpropataion Feedforward Network (in Python)
2. Multi-task Modular Backpropataion Feedforward Network (in Python)
( Download related paper [7] )

Evolutionary Algorithms for Optimisation

1. Real-Coded Genetic Algorithm (in C++)
2. Generalised Generation Gap with Parent Centric Crossover (in C++)
3. Cooperative Coevolution via G3-PCX (in C++)
4. Cooperative Coevolution via CMAES (in Matlab)

References

1. R. Chandra, "Towards prediction of rapid intensification in tropical cyclones with recurrent neural networks," CoRR abs/1701.04518 (2017)
2. R. Chandra, M. Zhang, "Cooperative coevolution of Elman recurrent neural networks for chaotic time series prediction," Neurocomputing 86: 116-123 (2012)
3. R. Chandra, L. Azizi, and S. Cripps, “Bayesian neural learning via Langevin dynamics for chaotic time series prediction,” International Conference on Neural Information Processing, China, November 2017. LNCS Springer (In Press).
4. R. Chandra, YS Ong, CK Goh, "Co-evolutionary multi-task learning for dynamic time series prediction," CoRR abs/1703.01887 (2017)
5. R. Chandra, “Co-evolutionary multi-task learning for modular pattern classification”, International Conference on Neural Information Processing, China, November 2017. LNCS Springer (In Press)
6. R. Chandra, YS Ong, CK Goh, "Co-evolutionary multi-task learning with predictive recurrence for multi-step chaotic time series prediction," Neurocomputing, 243: 21-34 (2017)
7. R. Chandra, “Multi-task modular backpropagation for feature-based pattern classification”, International Conference on Neural Information Processing, China, November 2017. LNCS Springer (In Press).