a deep learning based approach
play

A deep learning based approach for genetic risk prediction Raquel - PowerPoint PPT Presentation

Sep 09, 2023 •598 likes •1k views

A deep learning based approach for genetic risk prediction Raquel Dias, PhD. Senior Staff Scientist Scripps Research Translational Institute raqueld@scripps.edu, @RaquelDiasSRTI Ali Torkamani, PhD. atorkama@scripps.edu, @ATorkamani Whole

  1. A deep learning based approach for genetic risk prediction Raquel Dias, PhD. Senior Staff Scientist Scripps Research Translational Institute raqueld@scripps.edu, @RaquelDiasSRTI Ali Torkamani, PhD. atorkama@scripps.edu, @ATorkamani

  2. Whole Genome Sequencing vs. Genotype array Full Data Sparse Data (whole genome sequencing) (genotype array) 0 0 1 0 0 0 0 1 1 1 1 0 1 1 0 1 0 0 ? ? ? 0 0 1 1 ? ? 0 ? ? ? ? 0 0 1 0 0 0 0 1 1 1 1 0 1 1 0 1 0 0 ? ? ? 0 0 1 1 ? ? 0 ? ? ? ? 0 1 0 1 0 1 0 1 1 1 1 0 1 1 0 1 0 1 ? ? ? 1 0 1 1 ? ? 0 ? ? ? ? 0 0 1 0 0 0 0 1 1 0 0 1 0 1 1 0 0 0 ? ? ? 0 0 1 1 ? ? 1 ? ? ? ? 0 1 0 1 0 1 0 1 1 1 1 0 1 1 0 1 0 1 ? ? ? 1 0 1 1 ? ? 0 ? ? ? ? 0 1 0 1 1 1 0 1 1 1 1 0 1 1 0 1 0 1 ? ? ? 1 0 1 1 ? ? 0 ? ? ? ?

  3. Whole Genome Sequencing vs. Genotype array Full Data Sparse Data ~80M genetic variants ~4 million genetic 0 0 1 0 0 0 0 1 1 1 1 0 1 1 0 1 0 0 ? ? ? 0 0 1 1 ? ? 0 ? ? ? ? 0 0 1 0 0 0 0 1 1 1 1 0 1 1 0 1 0 0 ? ? ? 0 0 1 1 ? ? 0 ? ? ? ? 0 1 0 1 0 1 0 1 1 1 1 0 1 1 0 1 0 1 ? ? ? 1 0 1 1 ? ? 0 ? ? ? ? 0 0 1 0 0 0 0 1 1 0 0 1 0 1 1 0 0 0 ? ? ? 0 0 1 1 ? ? 1 ? ? ? ? 0 1 0 1 0 1 0 1 1 1 1 0 1 1 0 1 0 1 ? ? ? 1 0 1 1 ? ? 0 ? ? ? ? 0 1 0 1 1 1 0 1 1 1 1 0 1 1 0 1 0 1 ? ? ? 1 0 1 1 ? ? 0 ? ? ? ?

  4. Genetic imputation problem ... ... 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1 1 ... ... ... ... 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 ... ... HapMap or Reference ... ... 0 1 0 1 0 1 0 1 0 1 0 1 1 0 0 1 1 ... ... 1,000 Genomes haplotypes … (whole genome) ... ... 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 ... ... Prediction 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 Cases and Study 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 Controls typed genotypes 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 Genotype array 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1

  5. A typical imputation approach ... ... 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1 1 ... ... Multiethnic ... ... 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 ... ... Haplotype Reference ... ... 0 1 0 1 0 1 0 1 0 1 0 1 1 0 0 1 1 ... ... Reference panel … Consortium ... ... 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 ... ... (HRC) Mapping 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 Linkage disequilibrium 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 Study (LD r 2 ) structure 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 genotypes 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 Prediction 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1 0 0 ? 0 0 1 1 ? 0 ? ? ? 0 ? 1 ? 1

  6. A typical imputation approach ... ... 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1 1 ... ... Muli-ethinic ... ... 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 ... ... Haplotype Reference ... ... 0 1 0 1 0 1 0 1 0 1 0 1 1 0 0 1 1 ... ... Reference panel … Consortium ... ... 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 ... ... (HRC) Mapping 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 Linkage disequilibrium 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 Study (LD r 2 ) structure 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 genotypes 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 Prediction 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1

  7. Polygenic Risk Score (PRS)

  8. Polygenic Risk Calculation w/ Trait* w/o Trait Σ Design Results Polygenic Risk Score 100,000+ subjects Millions of known variants Cumulative sum *Trait can often be heterogeneous e.g. coronary artery = heart attack, stroke, bypass surgery, etc.

  9. Objectives 1. More accurate and faster imputation 2. Find important genetic variants 3. Better polygenic risk score calculation

  10. Our proposed approach Encoding v Output Input Hidden layer layer layer 1 1 1 1 0 2 0 0 2 2 𝑥 𝑥′ Decoding

  11. Denoising autoencoder for image restoration Noise Mask Bigdeli, Siavash Arjomand, and Matthias Zwicker. "Image restoration using autoencoding priors." arXiv preprint arXiv:1703.09964 (2017). Wang, Ruxin, and Dacheng Tao. "Non-local auto-encoder with collaborative stabilization for image restoration." IEEE Transactions on Image Processing 25.5 (2016): 2117-2129.

  12. Genotype imputation case study example Ground truth Masked input (whole genome sequencing) (genotype array) 0 0 1 0 0 0 0 1 1 1 1 0 1 1 0 1 0 0 ? ? ? 0 0 1 1 ? ? 0 ? ? ? ? 0 0 1 0 0 0 0 1 1 1 1 0 1 1 0 1 0 0 ? ? ? 0 0 1 1 ? ? 0 ? ? ? ? 0 1 0 1 0 1 0 1 1 1 1 0 1 1 0 1 0 1 ? ? ? 1 0 1 1 ? ? 0 ? ? ? ? 0 0 1 0 0 0 0 1 1 0 0 1 0 1 1 0 0 0 ? ? ? 0 0 1 1 ? ? 1 ? ? ? ? 0 1 0 1 0 1 0 1 1 1 1 0 1 1 0 1 0 1 ? ? ? 1 0 1 1 ? ? 0 ? ? ? ? 0 1 0 1 1 1 0 1 1 1 1 0 1 1 0 1 0 1 ? ? ? 1 0 1 1 ? ? 0 ? ? ? ? Mask

  13. Case study: 9p21.3 region of the genome • Length: 59846 bp • 846 genetic variants in reference panel (whole genome data) • Approx. 200 common variants • Approx. 600 rare variants • Only 17-47 variants in genotype array!!! • Strong association to coronary artery disease (CAD) • Genotyped and sequenced in many studies

  14. Training on the reference panel: Data augmentation strategy ... ... 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1 1 ... ... Reference ... ... 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 ... ... Whole Genome ... ... 0 1 0 1 0 1 0 1 0 1 0 1 1 0 0 1 1 ... ... Mask ... ... 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1 1 ... ... Masked ... ... 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 ... ... input ... ... 0 1 0 1 0 1 0 1 0 1 0 1 1 0 0 1 1 ... ... Autoencoder ... ... 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1 1 ... ... Reconstructed ... ... 0 0 0 0 0 1 1 0 0 1 0 1 0 0 1 0 1 ... ... Output ... ... 0 1 0 1 0 1 0 1 0 1 0 1 1 0 0 1 1 ... ...

  15. Customized Sparsity Loss Function  Sparsity loss with Kullback-Leibler (KL) / cross entropy element: 𝜍 = 𝜍 ∗ log 𝜍 + 1 − 𝜍 ∗ log 1 − 𝜍 𝐸 𝐿𝑀 (𝜍| ො 𝜍 ො 1 − ො 𝜍  Customized loss adjusted for hidden activation sparsity: 𝑜 𝑚𝑝𝑡𝑡 = 𝑁𝑇𝐹 + 𝛾 ∗ ෍ 𝐸 𝐿𝑀(𝑗) 𝑗=1  Mean Squared Error: 𝑜 𝑁𝑇𝐹 = 1 𝑜 ෍ (𝑧 𝑗 − ො 𝑧 𝑗 ) 𝑗=1

  16. Hyper parameters to be optimized • b • r • Activation functions • L1/L2 regularizers • Learning rate • Batch size

  17. Parallel Grid Search Hyperparameter optimization approach 10000 10000 100 9999 9999 … … … 2 4 4 1 3 3 2 2 100 X grid 10000 X training 10000 X training 1 1 search samples Hyperparameter Trained model Grid samples Grid samples combinations performance 9 GPUs available: - 7 GTX 1080, 860 hours Accuracy, loss - 1 Titan V, (sequential run, Sparsity, MSE - 1 Titan Xp 100 epochs)

  18. Grid Search Results: training accuracy

  19. Grid Search results: assessing best hyperparameter values

  20. Effect of hyper parameter values in training accuracy  b b Pearson correlation (r 2 )  r r  Learning rate

  21. Effect of hyper parameter values in training accuracy  b b Pearson correlation (r 2 )  r r  Learning rate

  22. Effect of hyper parameter values in training accuracy  b b Pearson correlation (r 2 )  r r  Learning rate

  23. Effect of hyper parameter values in training accuracy  b b Pearson correlation (r 2 )  r r  Learning rate

  24. Optimizing batch size: training accuracy Accuracy Loss Learning steps Learning steps 10 batches 50 batches 100 batches 1000 batches

  25. Optimizing batch size: training run time Accuracy Run time (hours) 10 batches 50 batches 100 batches 1000 batches

  26. Testing on multiple case studies • Atherosclerosis Risk in Communities (ARIC) • More than 3000 samples • Whole genome sequencing (846 variants, 0% mask, ground truth) • Affymetrix 6.0 genotype array (17 variants, 98% mask, input data) • Framingham Heart Study (FHS) • More than 500 samples • Whole genome sequencing (846 variants, 0% mask, ground truth) • Illumina 500K genotype array (47 variants, 95% mask, input data) • Illumina 5M (93 variants, 89% mask, input data)

  27. Accuracy in additional case studies: Proposed approach versus common statistic methodology Performance: all variants Performance: rare variants

  28. Accuracy in additional case studies: Proposed approach versus common statistic methodology Performance: all variants Performance: common variants

  29. Run time: Proposed approach versus common statistic methodology

  30. Linkage disequilibrium structure: ARIC Ground truth Linkage disequilibrium (LD) r 2 Prediction All variants Rare variants Common variants

  31. Linkage disequilibrium structure: FHS Ground truth Linkage disequilibrium (LD) r 2 Prediction All variants Rare variants Common variants

  32. Interpretability: identifying representative genetic variants Maximal information criteria

Recommend

A Hybrid Deep Learning Approach For Chaotic Time Series Prediction Based On Unsupervised Feature

A Hybrid Deep Learning Approach For Chaotic Time Series Prediction Based On Unsupervised Feature

A Hybrid Deep Learning Approach For Chaotic Time Series Prediction Based On Unsupervised Feature Learning Norbert Ayine Agana Advisor: Abdollah Homaifar Autonomous Control & Information Technology Institute (ACIT), Departmenent of

379 views • 35 slides
A Deep Learning-based approach to VM behavior identification in cloud systems Matteo Stefanini,

A Deep Learning-based approach to VM behavior identification in cloud systems Matteo Stefanini,

A Deep Learning-based approach to VM behavior identification in cloud systems Matteo Stefanini, Riccardo Lancellotti, Lorenzo Baraldi, Simone Calderara University of Modena and Reggio Emilia Department of engineering Enzo Ferrari CLOSER

180 views • 16 slides
An ENAS Based Approach for Constructing Deep Learning Models for Breast Cancer Recognition from

An ENAS Based Approach for Constructing Deep Learning Models for Breast Cancer Recognition from

An ENAS Based Approach for Constructing Deep Learning Models for Breast Cancer Recognition from Ultrasound Images Mohammed Ahmed, Hongbo Du, Alaa AlZoubi School of Computing, University of Buckingham, UK {1200526, Hongbo.du, alaa.alzoubi}

534 views • 5 slides
TOWARDS CREATING A KNOWLEDGE GAP FOR DEEP LEARNING BASED MEDICAL IMAGE ANALYSIS Dr. S.

TOWARDS CREATING A KNOWLEDGE GAP FOR DEEP LEARNING BASED MEDICAL IMAGE ANALYSIS Dr. S.

TOWARDS CREATING A KNOWLEDGE GAP FOR DEEP LEARNING BASED MEDICAL IMAGE ANALYSIS Dr. S. Kevin Zhou, Chinese Academy of Sciences Deep learning Input Algorithm: Output image variable Deep network X Y Y = f(X; W) Learning: arg min W S

685 views • 36 slides
Deep learning approach to describe and classify fungi microscopic images Medical Imaging with

Deep learning approach to describe and classify fungi microscopic images Medical Imaging with

Deep learning approach to describe and classify fungi microscopic images Medical Imaging with Deep Learning 2020 Bartosz Zieliski, Agnieszka Sroka-Oleksiak, Dawid Rymarczyk, Adam Piekarczyk, Monika Brzychczy-Woch Motivation We use a

181 views • 5 slides
Minjie Wang Deep Learning Deep Learning trend in the past 10 years Caffe State-of-art DL

Minjie Wang Deep Learning Deep Learning trend in the past 10 years Caffe State-of-art DL

Tofu: Parallelizing Deep Learning Systems with Automatic Tiling Minjie Wang Deep Learning Deep Learning trend in the past 10 years Caffe State-of-art DL system is based on dataflow GPU#0 w1 w2 data g1 g2 Forward propagation

405 views • 38 slides
Approach to practical application of Deep Learning in manufacturer's production line Masahiro

Approach to practical application of Deep Learning in manufacturer's production line Masahiro

Approach to practical application of Deep Learning in manufacturer's production line Masahiro Kashiwagi, Hiroyuki Kusaka, Kiminori Kurosawa and Kenji Nishide 1 1 Applications Service Agriculture Advertisement Education Security Deep

504 views • 32 slides
CS7015 (Deep Learning) : Lecture 1 (Partial/Brief) History of Deep Learning Mitesh M. Khapra

CS7015 (Deep Learning) : Lecture 1 (Partial/Brief) History of Deep Learning Mitesh M. Khapra

CS7015 (Deep Learning) : Lecture 1 (Partial/Brief) History of Deep Learning Mitesh M. Khapra Department of Computer Science and Engineering Indian Institute of Technology Madras 1/81 Acknowledgements Most of this material is based on the

923 views • 81 slides
CS7015 (Deep Learning) : Lecture 1 (Partial/Brief) History of Deep Learning Mitesh M. Khapra

CS7015 (Deep Learning) : Lecture 1 (Partial/Brief) History of Deep Learning Mitesh M. Khapra

CS7015 (Deep Learning) : Lecture 1 (Partial/Brief) History of Deep Learning Mitesh M. Khapra Department of Computer Science and Engineering Indian Institute of Technology Madras 1/49 Acknowledgements Most of this material is based on the

1.55k views • 98 slides
Deep learning to diagnose the neutron star Kenji Fukushima The University of Tokyo Based on

Deep learning to diagnose the neutron star Kenji Fukushima The University of Tokyo Based on

Deep learning to diagnose the neutron star Kenji Fukushima The University of Tokyo Based on work with Yuki Fujimoto, Koichi Murase Deep Learning and Physics 2018 June 2, 2018 @ Osaka 1 Disclaimer I am a user of the deep

619 views • 24 slides
Deep Feedforward Networks Lecture slides for Chapter 6 of Deep Learning www.deeplearningbook.org

Deep Feedforward Networks Lecture slides for Chapter 6 of Deep Learning www.deeplearningbook.org

Deep Feedforward Networks Lecture slides for Chapter 6 of Deep Learning www.deeplearningbook.org Ian Goodfellow Last updated 2016-10-04 Roadmap Example: Learning XOR Gradient-Based Learning Hidden Units Architecture Design

483 views • 30 slides
Advanced Model-Based Reinforcement Learning CS 294-112: Deep Reinforcement Learning Sergey

Advanced Model-Based Reinforcement Learning CS 294-112: Deep Reinforcement Learning Sergey

Advanced Model-Based Reinforcement Learning CS 294-112: Deep Reinforcement Learning Sergey Levine Class Notes 1. Homework 3 is extended by one week, to Wednesday after next Todays Lecture 1. Managing overfitting in model-based RL

686 views • 40 slides
Towards Interpretable Deep Learning for Natural Language Processing Roy Schwartz University of

Towards Interpretable Deep Learning for Natural Language Processing Roy Schwartz University of

Towards Interpretable Deep Learning for Natural Language Processing Roy Schwartz University of Washington & Allen Institute for Artificial Intelligence December 2018 1 / 37 ( Deep-Learning -Based) AI Today 2 / 37 3 / 37 Deep Learning

1.63k views • 140 slides
Robust Deep Learning Based on Meta-learning Deyu Meng Xian Jiaotong University

Robust Deep Learning Based on Meta-learning Deyu Meng Xian Jiaotong University

Robust Deep Learning Based on Meta-learning Deyu Meng Xian Jiaotong University dymeng@mail.xjtu.edu.cn http://gr.xjtu.edu.cn/web/dymeng Deep Learning Robust Meta-learning The Success of Deep Learning Relies on

1k views • 46 slides
Hao Su July 6, 2017 Outline Overview of 3D deep learning 3D deep learning algorithms

Hao Su July 6, 2017 Outline Overview of 3D deep learning 3D deep learning algorithms

Deep 3D Representation Learning for Visual Computing Hao Su July 6, 2017 Outline Overview of 3D deep learning 3D deep learning algorithms Conclusion 2 Outline Overview of 3D deep learning Background 3D deep learning tasks 3D deep

1.66k views • 122 slides
GPU-BASED A3C FOR DEEP REINFORCEMENT LEARNING M. Babaeizadeh , , I.Frosio , S.Tyree

GPU-BASED A3C FOR DEEP REINFORCEMENT LEARNING M. Babaeizadeh , , I.Frosio , S.Tyree

GPU-BASED A3C FOR DEEP REINFORCEMENT LEARNING M. Babaeizadeh , , I.Frosio , S.Tyree , J. Clemons , J.Kautz University of Illinois at Urbana-Champaign, USA NVIDIA, USA An ICLR 2017 paper A github project GPU-BASED

1.16k views • 48 slides
ACCELERATE DEEP LEARNING WITH NVIDIA'S DEEP LEARNING PLATFORM | STEPHEN JONES | GTC16 DEEP

ACCELERATE DEEP LEARNING WITH NVIDIA'S DEEP LEARNING PLATFORM | STEPHEN JONES | GTC16 DEEP

ACCELERATE DEEP LEARNING WITH NVIDIA'S DEEP LEARNING PLATFORM | STEPHEN JONES | GTC16 DEEP LEARNING EVERYWHERE INTERNET & CLOUD MEDICINE & BIOLOGY MEDIA & ENTERTAINMENT SECURITY & DEFENSE AUTONOMOUS MACHINES Image

308 views • 26 slides
Learning and Generating Distributed Routing Protocols Using Graph-Based Deep Learning Fabien

Learning and Generating Distributed Routing Protocols Using Graph-Based Deep Learning Fabien

Chair of Network Architectures and Services Department of Informatics Technical University of Munich Learning and Generating Distributed Routing Protocols Using Graph-Based Deep Learning Fabien Geyer, Georg Carle Monday 20 th August, 2018 ACM

819 views • 28 slides
GTC 2018 Silicon Valley, California Predictive Learning of Factor Based Strategies using Deep

GTC 2018 Silicon Valley, California Predictive Learning of Factor Based Strategies using Deep

GTC 2018 Silicon Valley, California Predictive Learning of Factor Based Strategies using Deep Neural Networks for Investment and Risk Management Yigal Jhirad and Blay Tarnoff March 27, 2018 GTC 2018: Table of Contents I. Deep Learning in

340 views • 31 slides
Excitation Deep Learning Model Abdul Qayyum ImViA laboratory, University of Bourgogne

Excitation Deep Learning Model Abdul Qayyum ImViA laboratory, University of Bourgogne

Medi dical cal Imagin ging g with th Deep ep Learni ning ng 2020 20 Short t Paper er MID IDL L 2020 20 Segmentation of the Myocardium on Late-Gadolinium Enhanced MRI based on 2.5 D Residual Squeeze and Excitation Deep Learning

203 views • 5 slides
Towards a learning approach for abbreviation detection and resolution Klaar Vanopstal, Bart

Towards a learning approach for abbreviation detection and resolution Klaar Vanopstal, Bart

Background Annotation Pattern-based approach Learning-based approach Conclusions and future work Towards a learning approach for abbreviation detection and resolution Klaar Vanopstal, Bart Desmet, V eronique Hoste LT 3 , Language and

1.42k views • 84 slides
Paper number: 136 A deep learning approach to segmentation of the developing cortex in fetal

Paper number: 136 A deep learning approach to segmentation of the developing cortex in fetal

Paper number: 136 A deep learning approach to segmentation of the developing cortex in fetal brain MRI with minimal manual labeling AE Fetit , A Alansary, L Cordero-Grande, J Cupitt, AB Davidson, AD Edwards, JV Hajnal, E Hughes, K Kamnitsas, V

676 views • 11 slides
Deep Learning on GPUs March 2016 What is Deep Learning? GPUs and DL AGENDA DL in practice

Deep Learning on GPUs March 2016 What is Deep Learning? GPUs and DL AGENDA DL in practice

Deep Learning on GPUs March 2016 What is Deep Learning? GPUs and DL AGENDA DL in practice Scaling up DL 2 What is Deep Learning? 3 DEEP LEARNING EVERYWHERE INTERNET & CLOUD MEDICINE & BIOLOGY SECURITY & DEFENSE MEDIA &

1.45k views • 39 slides
Reproducibility and Replicability in Deep Reinforcement Learning (and Other Deep Learning

Reproducibility and Replicability in Deep Reinforcement Learning (and Other Deep Learning

Reproducibility and Replicability in Deep Reinforcement Learning (and Other Deep Learning Methods) Peter Henderson Statistical Society of Canada Annual Meeting 2018 Contributors Riashat Islam Phil Bachman Doina Precup David Meger Joelle

850 views • 37 slides

More recommend