Partial Adversarial Training for Prediction Interval

Neural network (NN) based prediction or detection systems often perform excellently with easy problems without considering 1-5% difficult problems. This work proposes an adversarial NN training method for constructing the prediction interval (PI). The proposed training method considers adverse situations where traditional NN based PIs frequently fail. First, the conventional lower upper bound estimation (LUBE) method is applied in parallel for initial training of NNs with different initialization. Each NN based PI fails to cover a few samples. Input combinations of those samples are adversely changed by a small amount to generate the adverse samples. A new dataset is generated by appending adverse samples. Finally, an NN is trained with the adverse dataset. The method is applied to construct the NN for wind power prediction. According to the result analysis, the proposed method performs better in adverse situations.

[1]  Hemanshu R. Pota,et al.  Active power management in a low-voltage islanded microgrid , 2018 .

[2]  Dan Boneh,et al.  The Space of Transferable Adversarial Examples , 2017, ArXiv.

[3]  François Laviolette,et al.  Domain-Adversarial Training of Neural Networks , 2015, J. Mach. Learn. Res..

[4]  Sankaran Mahadevan,et al.  Reliability-based design optimization of multidisciplinary system under aleatory and epistemic uncertainty , 2016, Structural and Multidisciplinary Optimization.

[5]  R. Weron,et al.  Recent advances in electricity price forecasting: A review of probabilistic forecasting , 2016 .

[6]  Saeid Nahavandi,et al.  Evaluation and comparison of type reduction algorithms from a forecast accuracy perspective , 2013, 2013 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE).

[7]  Kais Zaman,et al.  Likelihood-based representation of epistemic uncertainty and its application in robustness-based design optimization , 2017 .

[8]  John Schulman,et al.  Concrete Problems in AI Safety , 2016, ArXiv.

[9]  Abbas Khosravi,et al.  Short-Term Load and Wind Power Forecasting Using Neural Network-Based Prediction Intervals , 2014, IEEE Transactions on Neural Networks and Learning Systems.

[10]  Saeid Nahavandi,et al.  Constructing Optimal Prediction Intervals by Using Neural Networks and Bootstrap Method , 2015, IEEE Transactions on Neural Networks and Learning Systems.

[11]  Christopher Meek,et al.  Adversarial learning , 2005, KDD '05.

[12]  Yee Whye Teh,et al.  Mondrian Forests for Large-Scale Regression when Uncertainty Matters , 2015, AISTATS.

[13]  Ananthram Swami,et al.  Practical Black-Box Attacks against Machine Learning , 2016, AsiaCCS.

[14]  Shujon Naha,et al.  LDA based paper currency recognition system using edge histogram descriptor , 2014, 2014 17th International Conference on Computer and Information Technology (ICCIT).

[15]  Charles Blundell,et al.  Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles , 2016, NIPS.

[16]  Fabio Roli,et al.  Randomized Prediction Games for Adversarial Machine Learning , 2016, IEEE Transactions on Neural Networks and Learning Systems.

[17]  Saeid Nahavandi,et al.  Prediction interval with examples of similar pattern and prediction strength , 2017, 2017 IEEE 30th Canadian Conference on Electrical and Computer Engineering (CCECE).

[18]  Hussain Mohammed Dipu Kabir,et al.  A frequency multiplier using three ambipolar graphene transistors , 2017, Microelectron. J..

[19]  John P. Parmigiani,et al.  Effect of incident angle on crack propagation at interfaces , 2016 .

[20]  John P. Parmigiani,et al.  An experimental assessment of methods to predict crack deflection at an interface , 2017 .

[21]  José Miguel Hernández-Lobato,et al.  Uncertainty Decomposition in Bayesian Neural Networks with Latent Variables , 2017, 1706.08495.

[22]  Swami Sankaranarayanan,et al.  Regularizing deep networks using efficient layerwise adversarial training , 2017, AAAI.

[23]  Md. Apel Mahmud,et al.  Multi-Agent Approach for Enhancing Security of Protection Schemes in Cyber-Physical Energy Systems , 2017, IEEE Transactions on Industrial Informatics.

[24]  Mansun Chan,et al.  Coil-Shaped Electrodes to Reduce the Current Variation of Drop-Casted OTFTs , 2017, IEEE Electron Device Letters.

[25]  Alex Kendall,et al.  What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision? , 2017, NIPS.

[26]  Susan T. Dumais,et al.  A Bayesian Approach to Filtering Junk E-Mail , 1998, AAAI 1998.

[27]  Saeid Nahavandi,et al.  An optimized mean variance estimation method for uncertainty quantification of wind power forecasts , 2014 .

[28]  Dan Boneh,et al.  Ensemble Adversarial Training: Attacks and Defenses , 2017, ICLR.

[29]  David A. Forsyth,et al.  NO Need to Worry about Adversarial Examples in Object Detection in Autonomous Vehicles , 2017, ArXiv.

[30]  Jonathon Shlens,et al.  Explaining and Harnessing Adversarial Examples , 2014, ICLR.

[31]  Peter A. Beling,et al.  Adversarial learning in credit card fraud detection , 2017, 2017 Systems and Information Engineering Design Symposium (SIEDS).

[32]  David A. Wagner,et al.  Audio Adversarial Examples: Targeted Attacks on Speech-to-Text , 2018, 2018 IEEE Security and Privacy Workshops (SPW).

[33]  H. R. Pota,et al.  A multi-agent approach for security of future power grid protection systems , 2016, 2016 IEEE Power and Energy Society General Meeting (PESGM).

[34]  Aki Vehtari,et al.  Chained Gaussian Processes , 2016, AISTATS.

[35]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[36]  Felipe Valencia,et al.  Prediction interval based on type-2 fuzzy systems for wind power generation and loads in microgrid control design , 2016, 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE).

[37]  David A. Forsyth,et al.  Standard detectors aren't (currently) fooled by physical adversarial stop signs , 2017, ArXiv.

[38]  Syed Bahauddin Alam,et al.  Non-linear Down-sampling and Signal Reconstruction, without Folding , 2010, 2010 Fourth UKSim European Symposium on Computer Modeling and Simulation.

[39]  Syed Bahauddin Alam,et al.  Watermarking with fast and highly secured encryption for real-time speech signals , 2010, 2010 IEEE International Conference on Information Theory and Information Security.

[40]  Iain Rice,et al.  Improved data visualisation through nonlinear dissimilarity modelling , 2018, Pattern Recognit..

[41]  Yu Tsao,et al.  Voice Conversion from Unaligned Corpora Using Variational Autoencoding Wasserstein Generative Adversarial Networks , 2017, INTERSPEECH.

[42]  G.J. Coram,et al.  How to (and how not to) write a compact model in Verilog-A , 2004, Proceedings of the 2004 IEEE International Behavioral Modeling and Simulation Conference, 2004. BMAS 2004..

[43]  Ke Zhu,et al.  A New Pearson-Type QMLE for Conditionally Heteroscedastic Models , 2015 .