probing neural networks in astronomy
play

Probing Neural Networks in Astronomy ARC CENTRE OF EXCELLENCE FOR - PowerPoint PPT Presentation

Feb 12, 2023 •100 likes •566 views

Colin Jacobs, Swinburne University of Technology Unimelb 26 August 2020 Probing Neural Networks in Astronomy ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D Deep learning - and its failures More and more applications in science (and

  1. Colin Jacobs, Swinburne University of Technology Unimelb 26 August 2020 Probing Neural Networks in Astronomy ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D

  2. Deep learning - and its failures More and more applications in science (and real life!) How can we find its weaknesses and know how it might fail? Can only know how well it will do on the data - we already have, may not be real world More sensitive to changes that would not fool - a human We might be blind to biases in the training set - These issues have consequences. Source: Wang 2017 For science : Hard to understand biases - Hasd to quantify errors - ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 2

  3. AI in science and society AI coming soon to your life : “The best minds of my generation Hiring and firing are thinking about how to make Financial access people click ads. (That sucks.)” University admission - Jeff Hammerbacher School rankings Legal system Advertising ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 3

  4. Fairness, transparency, accountability ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 4

  5. Bias in AI Challenges: Framing the problem Training data biased Lack of social context ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 5

  6. AI ethics ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 6

  7. Interpreting neural networks • Interpreting a trained ML model is vital to validate that the representation has accurately captured the general features of the data and not overfit. • High performance is mediated by generalisability. • An important step in ensuring the reproducibility of results. • Cars, medicine, courts, finance… urgent! Need something Explanatory and Interpretable SEE: Montavon, Samek and Muller (2018) and Lipton (2016) ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 7

  8. Neural networks - simple but complex Source: Veronez 2011 ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 8

  9. Convolutional neural networks - less simple but not too complex Source: Micheal Lanham 2018 ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 9

  10. ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 10

  11. What’s going on? Challenges with ANNs: Dimensionality of inputs enormous • Trainable weights ~10 6 - 10 9 • Hundreds of feature maps • Highly abstract and non-linear • Distribution of inputs, and gaps, hard to • comprehend Simonyan and Zisserman (2014) ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 11

  12. First attempt: Convolutional kernels ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 12

  13. Feature maps ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 13

  14. Input optimisation Take a trained model and train the inputs to maximise the activation for a particular class (maximise the output of a particular neuron). ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 14 Image: Varma and Das 2018

  15. Deep Dream ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 15 Pouff: https://www.youtube.com/watch?v=DgPaCWJL7XI

  16. Occlusion sensitivity Calculate the sensitivity to a particular pixel: i.e. d neuron/d pixel_i Very noisy! Smilkov et al 2017 ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 16

  17. Other attempts Deconvnet: Zeiler and Fergus 2014 Deconvolution : Zeiler and Fergus 2014 Guided backprop : Gradient of a particular neuron, through a ReLU. (Springenberg et al 2015). Springenberg et al 2015 ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 17

  18. Occlusion sensitivity Smoothgrad : Smilkov 2017 Adding noise to get more signal - sample an image many times (with added noise) and display the mean sensitivity map Smilkov et al 2017 ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 18

  19. Saliency mapping E.g. Grad-CAM (Selvaraju 2017) Take activations at last convolutional layer, determine importance to score Pool over feature maps -> importance Sum maps weighted by importance Upscale and project back onto input image. Selvaraju et al 2017 ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 19

  20. Saliency mapping: State of the art Integrated Gradients Occlusion Grad-CAM Input ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 20

  21. Saliency mapping Input LRP SmoothGrad Deconvnet ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 21 Guided Backprop PatternNet Pattern Attribution Deep Taylor

  22. Sensitivity Analysis How sensitive is the network to: A transformation of the data? • Dog: 97% Cat: 99% Some inherent property of the data? • Colour saturation: 50% Can we use this to identify weaknesses? Consider the correct-class probability as the key metric; could use another key measure. Dog: 93% Cat: 96% ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 22

  23. ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 23

  24. Sensie Automates sensitivity analysis - if you know what questions to ask! Available on Github Jacobs 2020 ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 24

  25. Sensie: Use case (MNIST) ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 25

  26. Sensie: Use case (CIFAR10) ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 26

  27. Querying an AI astronomer Jacobs+ 2019b ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 27

  28. Querying an AI astronomer ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 28

  29. False positives - Why? ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 29

  30. Feature activations ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 30

  31. ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 31

  32. Saliency mapping: Grad-CAM ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 32

  33. Grad-CAM - negative ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 33

  34. Probing with Sensie: Perturb test set ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 34

  35. Results: Colour ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 35

  36. Results: Blur (seeing) Effect on sims ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 36

  37. Effect on accuracy ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 37

  38. Results: Occlusion ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 38

  39. Results: PSF ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 39

  40. Results: Magnitude ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 40

  41. Results: Magnitude ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 41

  42. Results: Einstein Radius ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 42

  43. Conclusions Learned a few things: Good/expected: Not sensitive to Einstein radius - Robust to faint sources - - Sensitive to colour - physics? Some idea of a selection function - Bad: Sensitive to simulated PSF - Need to improve training set! github.com/coljac/sensie ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 43

  44. Further application: Redshifts ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 44

  45. REFERENCES Montavon, G., Samek, W. and Müller, K.R., 2018. Methods for interpreting and understanding deep neural ■ networks. Digital Signal Processing , 73 , pp.1-15. Lipton, Z.C., 2016. The mythos of model interpretability. arXiv preprint arXiv:1606.03490 . ■ Simonyan, K. and Zisserman, A., 2014. Very deep convolutional networks for large-scale image recognition. arXiv ■ preprint arXiv:1409.1556 . Greydanus, S., Kaul, A., Dodge, J. and Fern, A., 2017. Visualising and understanding atari agents. arXiv preprint ■ arXiv: 1711.00138. Zeiler, M. D., & Fergus, R. 2014, in Computer Vision – ECCV 2014, ed. D. Fleet, T. Pajdla, B. Schiele, & T. Tuytelaars, ■ Vol. 8689 (Cham: Springer International Publishing), 818–833 Selvaraju, R. R., Cogswell, M., Das, A., et al. 2017, in Proceedings of the IEEE International Conference on ■ Computer Vision, 618–626 Binder, A., Bach, S., Montavon, G., Müller, K.-R., & Samek, W. 2016, in Information Science and Applications (ICISA) ■ 2016, ed. K. J. Kim & N. Joukov, Lecture Notes in Electrical Engineering (Springer Singapore), 913–922 Smilkov, D., Thorat, N., Kim, B., Viégas, F., & Wattenberg, M. (2017), arXiv e-prints, arXiv:1706.03825. ■ ARC CENTRE OF EXCELLENCE FOR ALL SKY ASTROPHYSICS IN 3D 45

Recommend

Outline Evolution of neurocomputing Artificial neural networks Feed forward

Outline Evolution of neurocomputing Artificial neural networks Feed forward

Introduction to Neural Networks and Deep Learning Ling Guan References: S. Haykin, Neural Networks , 2 nd Edition, New Jersey: Prentice Hall, 1. 2004. C. M. Bishop, Neural Networks for Pattern Re cognition, New York: 2. Oxford University

451 views • 17 slides
Neural Networks 1. Introduction Fall 2017 Neural Networks are taking over! Neural networks

Neural Networks 1. Introduction Fall 2017 Neural Networks are taking over! Neural networks

Neural Networks 1. Introduction Fall 2017 Neural Networks are taking over! Neural networks have become one of the major thrust areas recently in various pattern recognition, prediction, and analysis problems In many problems they have

1.17k views • 91 slides
Neural Networks 0. Logistics Spring 2019 1 Neural Networks are taking over! Neural networks

Neural Networks 0. Logistics Spring 2019 1 Neural Networks are taking over! Neural networks

Neural Networks 0. Logistics Spring 2019 1 Neural Networks are taking over! Neural networks have become one of the major thrust areas recently in various pattern recognition, prediction, and analysis problems In many problems they have

852 views • 33 slides
CONVOLUTIONAL AND RECURRENT NEURAL NETWORKS Neural networks Fully connected networks

CONVOLUTIONAL AND RECURRENT NEURAL NETWORKS Neural networks Fully connected networks

CONVOLUTIONAL AND RECURRENT NEURAL NETWORKS Neural networks Fully connected networks Neuron Non-linearity Softmax layer DNN training Loss function and regularization SGD and backprop Learning rate Overfitting

1.8k views • 86 slides
Neural Networks 1. Introduction Spring 2020 1 Neural Networks are taking over! Neural

Neural Networks 1. Introduction Spring 2020 1 Neural Networks are taking over! Neural

Neural Networks 1. Introduction Spring 2020 1 Neural Networks are taking over! Neural networks have become one of the major thrust areas recently in various pattern recognition, prediction, and analysis problems In many problems they

1.63k views • 119 slides
Neural Networks 1. Introduction Spring 2019 1 Neural Networks are taking over! Neural

Neural Networks 1. Introduction Spring 2019 1 Neural Networks are taking over! Neural

Neural Networks 1. Introduction Spring 2019 1 Neural Networks are taking over! Neural networks have become one of the major thrust areas recently in various pattern recognition, prediction, and analysis problems In many problems they

1.57k views • 124 slides
Relaxation and Hopfield Networks Neural Networks Neural Networks - Hopfield 1 Bibliography

Relaxation and Hopfield Networks Neural Networks Neural Networks - Hopfield 1 Bibliography

Relaxation and Hopfield Networks Neural Networks Neural Networks - Hopfield 1 Bibliography Hopfield, J. J., "Neural networks and physical systems with emergent collective computational abilities," Proceedings of the National Academy

367 views • 19 slides
Neural Networks and their Application to Go Neural Networks Learning Blackjack Theory Training

Neural Networks and their Application to Go Neural Networks Learning Blackjack Theory Training

Neural Networks and their Application to Go A. Bausch Neural Networks and their Application to Go Neural Networks Learning Blackjack Theory Training neural networks Problems AlphaGo Anne-Marie Bausch The Game of Go Policy Network

280 views • 24 slides
(Very) Brief Introduction to Neural Networks IITP-03 Algorithms for NLP 1 / 31 Learning

(Very) Brief Introduction to Neural Networks IITP-03 Algorithms for NLP 1 / 31 Learning

(Very) Brief Introduction to Neural Networks IITP-03 Algorithms for NLP 1 / 31 Learning objectives What are neural networks? What are deep neural networks? How do we train neural networks? What variants of neural network

331 views • 31 slides
How Neural Networks (NN) Biological Neuron: A . . . Can (Hopefully) Learn Artificial Neural . .

How Neural Networks (NN) Biological Neuron: A . . . Can (Hopefully) Learn Artificial Neural . .

Need for Machine . . . Neural Networks . . . How to Speed Up . . . Neural Networks: A . . . How Neural Networks (NN) Biological Neuron: A . . . Can (Hopefully) Learn Artificial Neural . . . Resulting Algorithm Faster by Taking Into How to

464 views • 17 slides
Outline Why model neural networks? Modeling Neural Networks A brief look at the neuron.

Outline Why model neural networks? Modeling Neural Networks A brief look at the neuron.

Outline Why model neural networks? Modeling Neural Networks A brief look at the neuron. A look at some current works. Adding an evolutionary strategy. Paul Nuytten CPSC 607 Why Model Neural Networks? The Neuron The nervous

697 views • 4 slides
Introduction to Deep Neural Networks 0. Logistics Spring 2020 1 Neural Networks are taking

Introduction to Deep Neural Networks 0. Logistics Spring 2020 1 Neural Networks are taking

Introduction to Deep Neural Networks 0. Logistics Spring 2020 1 Neural Networks are taking over! Neural networks have become one of the major thrust areas recently in various pattern recognition, prediction, and analysis problems In

987 views • 40 slides
Propagating Error Backward Hyperparameters for Neural Networks } Multi-layer (deep) neural

Propagating Error Backward Hyperparameters for Neural Networks } Multi-layer (deep) neural

Learning in Neural Networks w 1,3 w 3,5 1 3 5 w w 1,4 3,6 w w 2,3 4,5 2 4 6 w w 2,4 4,6 Class #18: Back-Propagation; } A neural network can learn a classification function by Tuning Hyper-Parameters adjusting its weights to

259 views • 4 slides
Neural Networks Hugo Larochelle ( @hugo_larochelle ) Google Brain 2 NEURAL NETWORKS What

Neural Networks Hugo Larochelle ( @hugo_larochelle ) Google Brain 2 NEURAL NETWORKS What

Neural Networks Hugo Larochelle ( @hugo_larochelle ) Google Brain 2 NEURAL NETWORKS What well cover ... f ( x ) computer vision architectures - convolutional networks - data augmentation 1 - residual networks ... ...

1.29k views • 75 slides
Probing the Galactic Planetary Census Greg Laughlin -- UCSC Astronomy Tuesday, June 26, 12

Probing the Galactic Planetary Census Greg Laughlin -- UCSC Astronomy Tuesday, June 26, 12

Probing the Galactic Planetary Census Greg Laughlin -- UCSC Astronomy Tuesday, June 26, 12 Exoplanet News The finding was called exciting by from the AAS Dr. Kenneth Franklin of the American Museum-Hayden Planetarium. He meeting

1.01k views • 21 slides
Neural Networks Oskar Taubert (SCC) SCC 1 15.01.2020 Oskar Taubert - Neural Networks SCC

Neural Networks Oskar Taubert (SCC) SCC 1 15.01.2020 Oskar Taubert - Neural Networks SCC

Neural Networks Oskar Taubert (SCC) SCC 1 15.01.2020 Oskar Taubert - Neural Networks SCC www.kit.edu KIT The Research University in the Helmholtz Association Neural Network Concept (very) remotely brain inspired computational system

479 views • 25 slides
Introduction to Artificial Intelligence Neural Networks - Deep Learning for NLP Janyl Jumadinova

Introduction to Artificial Intelligence Neural Networks - Deep Learning for NLP Janyl Jumadinova

Introduction to Artificial Intelligence Neural Networks - Deep Learning for NLP Janyl Jumadinova November 21, 2016 Neural Networks 2/20 Neural Networks 3/20 Neural Networks Neural computing requires a number of neurons , to be connected

813 views • 21 slides
7. Artificial neural networks Introduction to neural networks Despite struggling to understand

7. Artificial neural networks Introduction to neural networks Despite struggling to understand

7. Artificial neural networks Introduction to neural networks Despite struggling to understand intricacies of protein, cell, and network function within the brain, neuroscientists would agree on the following simplistic description of how the

157 views • 11 slides
Neural networks Chapter 20 Chapter 20 1 Outline Brains Neural networks Perceptrons

Neural networks Chapter 20 Chapter 20 1 Outline Brains Neural networks Perceptrons

Neural networks Chapter 20 Chapter 20 1 Outline Brains Neural networks Perceptrons Multilayer networks Applications of neural networks Chapter 20 2 Brains 10 11 neurons of > 20 types, 10 14 synapses, 1ms10ms cycle

804 views • 47 slides
Modeling Neural Networks Paul Nuytten CPSC 607 Outline Why model neural networks? A

Modeling Neural Networks Paul Nuytten CPSC 607 Outline Why model neural networks? A

Modeling Neural Networks Paul Nuytten CPSC 607 Outline Why model neural networks? A brief look at the neuron. A look at some current works. Adding an evolutionary strategy. Why Model Neural Networks? The nervous system is a

1.22k views • 23 slides
Modular Neural Networks CPSC 533 Franco Lee Ian Ko Modular Neural Networks What is it ? Dif

Modular Neural Networks CPSC 533 Franco Lee Ian Ko Modular Neural Networks What is it ? Dif

Modular Neural Networks CPSC 533 Franco Lee Ian Ko Modular Neural Networks What is it ? Dif f erent models of neural net w orks combined int o a single syst em. Each single net w ork is made int o a module t hat can be f reely int

289 views • 25 slides
Probing Features in the Primordial Power Spectrum Arman Shafieloo Korea Astronomy and Space

Probing Features in the Primordial Power Spectrum Arman Shafieloo Korea Astronomy and Space

Probing Features in the Primordial Power Spectrum Arman Shafieloo Korea Astronomy and Space Science Institute (KASI) & University of Science and Technology (UST) General Relativity - The Next Generation February 19 - 23, 2018 YITP, Kyoto

438 views • 42 slides
Convolutional Neural Networks (CNNs) Recurrent Neural Networks (RNNs) L1 Scalar Processor L0

Convolutional Neural Networks (CNNs) Recurrent Neural Networks (RNNs) L1 Scalar Processor L0

Convolutional Neural Networks (CNNs) Recurrent Neural Networks (RNNs) L1 Scalar Processor L0 L0 M*V Processor N N N N N N Project Brainwave Pretrained DNN Model Scalable DNN Hardware Neural Processing Unit Microservice Extract

525 views • 30 slides
Computa(on through dynamics Using recurrent neural networks to unveil mechanism in neural

Computa(on through dynamics Using recurrent neural networks to unveil mechanism in neural

Computa(on through dynamics Using recurrent neural networks to unveil mechanism in neural circuits David Sussillo with Valerio Mante and Bill Newsome Table of contents Introduc(on Training recurrent neural networks(RNNs) Understanding how

1.57k views • 105 slides

More recommend