Jingyuan Zhang 1, Bokai Cao 1, Sihong Xie 1, Chun-Ta Lu 1, Philip S. Yu 1,2, Ann B. Ragin 3 Identifying Connectivity Patterns for Brain Diseases via Multi-side-view.

Slides:

Advertisements

Similar presentations

Neural Networks in Financial Analysis

Advertisements

NEURAL NETWORKS Biological analogy

Introduction to Artificial Neural Networks

Advanced topics.

Stacking RBMs and Auto-encoders for Deep Architectures References:[Bengio, 2009], [Vincent et al., 2008] 2011/03/03 강병곤.

Presented by: Mingyuan Zhou Duke University, ECE September 18, 2009

An introduction to: Deep Learning aka or related to Deep Neural Networks Deep Structural Learning Deep Belief Networks etc,

Neural NetworksNN 11 Neural Networks Teacher: Elena Marchiori R4.47 Assistant: Kees Jong S2.22

Unsupervised Learning With Neural Nets Deep Learning and Neural Nets Spring 2015.

Neural Networks Basic concepts ArchitectureOperation.

1 Learning to Detect Objects in Images via a Sparse, Part-Based Representation S. Agarwal, A. Awan and D. Roth IEEE Transactions on Pattern Analysis and.

Connectionist models. Connectionist Models Motivated by Brain rather than Mind –A large number of very simple processing elements –A large number of weighted.

Slide 1 EE3J2 Data Mining EE3J2 Data Mining Lecture 15: Introduction to Artificial Neural Networks Martin Russell.

Power Systems Application of Artificial Neural Networks. (ANN)  Introduction  Brief history.  Structure  How they work  Sample Simulations. (EasyNN)

Visual Expertise Is a General Skill Maki Sugimoto University of California, San Diego November 20, 2000.

1 Automated Feature Abstraction of the fMRI Signal using Neural Network Clustering Techniques Stefan Niculescu and Tom Mitchell Siemens Medical Solutions,

Autoencoders Mostafa Heidarpour

MACHINE LEARNING AND ARTIFICIAL NEURAL NETWORKS FOR FACE VERIFICATION

Neural Networks in Data Mining “An Overview”

Presented by: Kamakhaya Argulewar Guided by: Prof. Shweta V. Jain

How to do backpropagation in a brain

Artificial Intelligence Lecture No. 28 Dr. Asad Ali Safi Assistant Professor, Department of Computer Science, COMSATS Institute of Information Technology.

1 st Neural Network: AND function Threshold(Y) = 2 X1 Y X Y.

IE 585 Introduction to Neural Networks. 2 Modeling Continuum Unarticulated Wisdom Articulated Qualitative Models Theoretic (First Principles) Models Empirical.

Authors : Ramon F. Astudillo, Silvio Amir, Wang Lin, Mario Silva, Isabel Trancoso Learning Word Representations from Scarce Data By: Aadil Hayat (13002)

Explorations in Neural Networks Tianhui Cai Period 3.

A shallow introduction to Deep Learning

Using Neural Networks to Predict Claim Duration in the Presence of Right Censoring and Covariates David Speights Senior Research Statistician HNC Insurance.

NEURAL NETWORKS FOR DATA MINING

Chapter 7 Neural Networks in Data Mining Automatic Model Building (Machine Learning) Artificial Intelligence.

Building high-level features using large-scale unsupervised learning Anh Nguyen, Bay-yuan Hsu CS290D – Data Mining (Spring 2014) University of California,

Methodology of Simulations n CS/PY 399 Lecture Presentation # 19 n February 21, 2001 n Mount Union College.

Xiangnan Kong,Philip S. Yu Multi-Label Feature Selection for Graph Classification Department of Computer Science University of Illinois at Chicago.

EE459 Neural Networks Examples of using Neural Networks Kasin Prakobwaitayakit Department of Electrical Engineering Chiangmai University.

Neural Networks Demystified by Louise Francis Francis Analytics and Actuarial Data Mining, Inc.

Neural Networks Presented by M. Abbasi Course lecturer: Dr.Tohidkhah.

Introduction to Deep Learning

Cognitive models for emotion recognition: Big Data and Deep Learning

Convolutional Restricted Boltzmann Machines for Feature Learning Mohammad Norouzi Advisor: Dr. Greg Mori Simon Fraser University 27 Nov

Artificial Neural Networks (ANN). Artificial Neural Networks First proposed in 1940s as an attempt to simulate the human brain’s cognitive learning processes.

Perceptrons Michael J. Watts

Deep Belief Network Training Same greedy layer-wise approach First train lowest RBM (h 0 – h 1 ) using RBM update algorithm (note h 0 is x) Freeze weights.

Deep Learning Overview Sources: workshop-tutorial-final.pdf

Network Management Lecture 13. MACHINE LEARNING TECHNIQUES 2 Dr. Atiq Ahmed Université de Balouchistan.

Artificial Neural Networks By: Steve Kidos. Outline Artificial Neural Networks: An Introduction Frank Rosenblatt’s Perceptron Multi-layer Perceptron Dot.

Xintao Wu University of Arkansas Introduction to Deep Learning 1.

Business Intelligence and Decision Support Systems (9 th Ed., Prentice Hall) Chapter 6: Artificial Neural Networks for Data Mining.

Big data classification using neural network

A presentation to El Paso del Norte Software Association

Data Mining, Neural Network and Genetic Programming

Supervised Training of Deep Networks

Deep Learning Qing LU, Siyuan CAO.

Unsupervised Learning and Autoencoders

Deep Learning Workshop

Introduction to Neural Networks

Towards Understanding the Invertibility of Convolutional Neural Networks Anna C. Gilbert1, Yi Zhang1, Kibok Lee1, Yuting Zhang1, Honglak Lee1,2 1University.

شبکه عصبی تنظیم: بهروز نصرالهی-فریده امدادی استاد محترم: سرکار خانم کریمی دانشگاه آزاد اسلامی واحد شهرری.

Artificial Intelligence Methods

Deep learning Introduction Classes of Deep Learning Networks

network of simple neuron-like computing elements

Introduction to Neural Networks And Their Applications - Basics

Example of a simple deep network architecture.

Prepared by: Mahmoud Rafeek Al-Farra

III. Introduction to Neural Networks And Their Applications - Basics

An introduction to: Deep Learning aka or related to Deep Neural Networks Deep Structural Learning Deep Belief Networks etc,

Deep Neural Networks as Scientific Models

Example of a simple deep network architecture.

Peng Cui Tsinghua University

Visual Grounding.

Presentation transcript:

Jingyuan Zhang 1, Bokai Cao 1, Sihong Xie 1, Chun-Ta Lu 1, Philip S. Yu 1,2, Ann B. Ragin 3 Identifying Connectivity Patterns for Brain Diseases via Multi-side-view Guided Deep Architectures 1. University of Illinois at Chicago 2. Tsinghua University 3. Northwestern University

2 Outline Introduction & Motivation Data Analysis Methodology Experiments Conclusions

3 Introduction Mining neuroimage data  Detect neurological disorders Problems  Connectivities of brain regions  Interpretation of disease mechanisms

4 Introduction Mining brain networks  Discover clinically meaningful patterns

5 Introduction Brain networks  Nodes: brain regions e.g., insula, hippocampus  Edges: connections of brain regions fMRI: correlations between brain regions in functional activity DTI: number of neural fibers connecting brain regions

6 Introduction Mining brain networks  Connected subgraph patterns Problems  Interplay among patterns

7 Motivation Deep learning architectures  Non-linear representations of patterns Difficulties  Limited subjects & high costs Multiple side views  Other sources of auxiliary data  Clinical, serologic, immunologic, etc.

8 Motivation Tackle the bias of scarce data  Incorporate multiple side views  Guide the representation learning Why not a simple concatenation?

9 Motivation A simple concatenation  Mixture features  Difficulty of interpretation Our basic idea  Pairwise similarities from side views  Patterns w.r.t the geometry of side views Are the multiple side views really useful in guiding deep learning?

10 Data Analysis Multi-side-view guidance  Consistency between side views and pre-specified label information Two-sample one-tail t-test  Similar instances under side views  Similar labels, both positive or negative Significance level 0.05

11 Methodology MVAE  Multi-side-View guided AutoEncoder

12 Methodology The classic autoencoder (AE)  Self-reconstruction criterion  Unsupervised feedforward NN  Fit input with the reconstructed output

13 Methodology Multi-side-view guidance  Encapsulate side views into AE  Pairwise relationship of learned features  Consistent with that of the side views

14 Methodology Objective function of MVAE Guiding parameter

15 Methodology Build deep architectures  Stacked MVAE  Greedy layer-wise pre-training  Supervised fine-tuning

16 Experiments Data collection  Chicago Early HIV Infection Study  56 HIV and 21 normal controls Brain networks  From fMRI and DTI images Seven side views  Neuropsychological tests  Flow cytometry  Plasma luminex ...

17 Experiments Baselines 1. gSSC 2. Freq 3. Conf 4. Ratio 5. Gtest 6. HSIC gMSV AE-1 MVAE-1 No side-view guidance Side-view guidance No side-view guidance Side-view guidance

18 Experiments

19 Experiments Performance of stacked MVAE  MVAE-2 performs the best  MVAE-3 may over fit the limited data

20 Experiments Performance with each side view  Brain volumetry  Plasma luminex

21 Experiments Analysis of the guiding parameter  Relative importance of side views

22 Experiments Visualization on fMRI data  Each neuron in the first hidden layer  A discriminative connectivity pattern  Two most significant connectivity patterns

23 Conclusions Identify brain connectivity patterns  Utilize a deep architecture  Explore multiple side views  Benefit not only classification tasks  But also clinical interpretations

24

25

26