Supporting Information

Size: px
Start display at page:

Download "Supporting Information"

Transcription

1 Supporting Information Convolutional Embedding of Attributed Molecular Graphs for Physical Property Prediction Connor W. Coley a, Regina Barzilay b, William H. Green a, Tommi S. Jaakkola b, Klavs F. Jensen a a Department of Chemical Engineering, Massachusetts Institute of Technology; 77 Massachusetts Avenue, Cambridge, MA b Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology; 77 Massachusetts Avenue, Cambridge, MA Corresponding author, kfjensen@mit.edu S1. Full embedding algorithm The model architecture for convolutional embedding, which converts a molecular graph to a learned fingerprint vector, is as follows. 1. Define molecular tensor and initialize network a. For each molecule, define its molecular tensor as where is the combined number of atom and bond features and is the number of heavy atoms (i.e., not counting hydrogen), such that, 0,, 0, 0, connected otherwise, 1,, where are the atom-level features describing atom and are the bond-level features describing the bond between atoms and. b. Initialize the cumulative molecular fingerprint,, of length with all zeros. 0

2 c. Initialize the model weights and bias terms required at every stage of the convolution (i.e., depth or radius ), 0,1,,. Inner update weight matrices,, which are used to convert an atom and its neighbors representations into a new representation for that atom, are initialized to be close to the identity matrix. Outer update weight matrices,, which are used to convert an atom s representation to a longer atom fingerprint representation, are initialized to be close to zero. Small random uniform noise, is added to the weights and bias terms , , , 1,, 0.005, , , 1,, 2. Define initial attributes at radius zero, Define the initial attribute matrix at 0, that describes the initial feature vector corresponding to each atom.,, 1,, 1, Or, equivalently, using the molecular tensor,,, 1, N, 1, 3. Generate atom-level fingerprints and add to cumulative molecular fingerprint For each atom, calculate an atom-level fingerprint based on its current attributes by passing the current attribute vector through an output layer with weights (e.g., softmax). Add this fingerprint to the cumulative molecular fingerprint., bias, and non-linear activation function 4. Update the attributes associated with each atom Calculate new attributes for each atom based on its own attributes and those of its neighbors. This is done by passing those values through an internal update hidden layer using weights and bias. Information about neighbor features for each atom is conveniently stored in the th entry along the first dimension of. Because the rest of is filled with zeros when two atoms are not connected, we can sum over the second dimension of to aggregate neighbors; because the atom features of are recorded

3 along the diagonal,, this summation will include the features for itself. Next, we sum self and neighbor contributions before passing summed attributes through an inner hidden layer, 1, By design, has the same length along its first two dimensions and is initialized close to the identity matrix, so the updated attributes matrix is the same shape as the previous one,. 5. Update the molecular tensor Update the molecular tensor with new atom attributes. The attributes of an atom will appear in the diagonal of the molecular tensor and in any off-diagonal entry corresponding to a bond. The final index along the third dimension of is a placeholder feature that describes whether a bond is present.,,,, or,, otherwise 1, 1, 6. Repeat until target radius has been reached Repeat steps 3-5, updating the overall learned fingerprint to include fingerprint contributions at each radius, until the specified maximum radius, has been reached 1 Learning step For physical property prediction, the loss function used is the mean squared error; for toxicity prediction on each of the 12 targets, the loss function used is the binary crossentropy. Parameters (weights, biases) in the embedding architecture are trained at the same time as parameters (weights, biases) in the neural network regression architecture using the Adam or Adadelta update procedures, which operate similarly to gradient descent with a dynamic learning rate. The use of Theano operations in the network implementation enables automatic calculation of analytical gradients, i.e. the derivative of the loss function with respect to all network parameters. S2. Model hyperparameters

4 Table S1. Model hyperparameters. Common to all models are: = tanh, = softmax, = tanh, = 512, 5, 50, optimizer = adam, batch size = 1, loss = mean squared error.. CNN-De-aq-all e -epoch/30 Not CV; Train on full 100% CNN-Br-tm-all e -epoch/30 Not CV; Train on full 100% Learning rate Notes CNN-Ab-octrepresentative e -epoch/30 CNN-De-aq-representative e -epoch/30 Train on full 80% CNN-Br-tm-representative e -epoch/30 Train on full 80% CNN-Ab-oct-all e -epoch/30 Not CV; Train on full 100% CNN-Ab-oct-all-De-aqrepresentative CNN-De-aq-all-Ab-octrepresentative CNN-Br-tm-all-Ab-octrepresentative CNN-Ab-oct-De-aqrepresentative e -epoch/30 Train on full 80%; Initialized with CNN- Ab-oct-all e -epoch/30 Initialized with CNN-De-aq-all e -epoch/30 Initialized with CNN-Br-tm-all e -epoch/30 Multitarget; Train on full 80% Tox21-ST ab Adadelta default in Keras Tox21-ECFP4 abe Adadelta default in Keras Tox21-ECFP6 abe Adadelta default in Keras 80% training, 20% internal validation, separate test dataset 80% training, 20% internal validation, separate test dataset 80% training, 20% internal validation, separate test dataset a Activation of final output node = sigmoid, optimizer =Adadelta, loss = binary crossentropy b 100, 50 e Convolutional embedding replaced with fixed fingerprint representation S3. Baseline SVM model performance As an indication of how other QSAR/QSPR methods might perform on the same dataset, without pursuing a full optimization of those methods hyperparameters, we use Scikit-learn s SVM implementation to fit regressions for the Abraham, Delaney, and Bradley datasets. Default values of C (1.0) and epsilon (0.1) were used. For these tests, molecules are represented either by the Morgan fingerprint of radius 2 or of radius 3 (as calculated by RDKit with features, mimicking ECFP4 and ECFP6 fingerprints) folded to length 512. This length matches the learned fingerprint length in our convolutional approach. Three different kernel functions were used: linear, radial basis functions (RBF), and the Tanimoto similarity score defined for Boolean vector fingerprints and as the following:, and, or,

5 Models were trained and tested using 5-fold CVs in triplicate, identical to the convolutional models. We acknowledge that this is by no means a definitive comparison to the full set of possible machine learning techniques, but it does provide a reference point for model performance using more traditional machine learning approaches. The results are shown below in Table S3. Performance is significantly worse than the convolutional models, although the Abraham SVM model using the ECFP4 fingerprint and Tanimoto kernel is better than the two worst-performing convolutional models. Table S3. Performance summary for baseline model runs using Morgan fingerprints of radius 2 or 3, using different kernel functions, run in triplicate. Units are log10(mol/l) for Abraham and Delaney solubility models, and degrees Celsius for Bradley melting point models. The best results for each dataset are bolded. = = Abraham octanol solubility Delaney aqueous solubility log(m) log(m) = Bradley melting point deg C Dataset Fingerprint Kernel MSE MAE SD Abraham Morgan (r=3) Linear p/m p/m p/m Abraham Morgan (r=3) RBF p/m p/m p/m Abraham Morgan (r=3) Tanimoto p/m p/m p/m Abraham Morgan (r=2) Linear p/m p/m p/m Abraham Morgan (r=2) RBF p/m p/m p/m Abraham Morgan (r=2) Tanimoto p/m p/m p/m Delaney Morgan (r=3) Linear p/m p/m p/m Delaney Morgan (r=3) RBF p/m p/m p/m Delaney Morgan (r=3) Tanimoto p/m p/m p/m Delaney Morgan (r=2) Linear p/m p/m p/m Delaney Morgan (r=2) RBF p/m p/m p/m Delaney Morgan (r=2) Tanimoto p/m p/m p/m Bradley Morgan (r=3) Linear p/m p/m p/m Bradley Morgan (r=3) RBF p/m p/m p/m Bradley Morgan (r=3) Tanimoto p/m p/m p/m Bradley Morgan (r=2) Linear p/m p/m p/m Bradley Morgan (r=2) RBF p/m p/m p/m Bradley Morgan (r=2) Tanimoto p/m p/m p/m S4. Code

6 Code and data used for model training/testing can be found at S5. Hyperparameters To examine the sensitivity to key hyperparameters (fingerprint depth during convolution, fingerprint length, and the number of hidden nodes in the regression layer ), we assessed performance on the Abraham octanol solubility dataset using a fixed 80%/20% training/testing split while varying only one parameter. The training schedule was arranged so that a small portion of the training data (~10%) was reserved for internal validation and early stopping. Results were noisy due to the limited number of samples in the Abraham dataset. Comparative boxplots are shown in Figures S4, S5, and S6 in terms of absolute error. Box bounds indicate the 25 th and 75 th percentiles, whiskers indicate the minimum and maximum, red lines indicate the median, and outliers are represented as red plus signs. Deeper convolutions, longer fingerprints, and more hidden nodes all increase the overall flexibility of the model. This is beneficial as long as long as trends are being accurately captured and as long as regularization can prevent overfitting. With smaller datasets the Abraham dataset in particular excessive flexibility will cause overfitting and poor generalization. We find that a fingerprint depth of 5, a fingerprint length of 512, and a hidden layer of size 50 provides a reasonable amount of flexibility without sacrificing test set performance, hence these values were used for models corresponding to the settings in Table S1. It is important to note that these settings were obtained based on the results of a randomized 20% test set from the Abraham dataset and then applied to all other models for the purposes of discussion only; quantitative results are also shown using a full hyperparameter grid search as described in the main text.

7 Figure S4. Effect of changing the fingerprint depth, i.e. the number of iterations used in the convolutional process, on model performance using the Abraham octanol solubility dataset 245 as a model dataset. Figure S5. Effect of changing the fingerprint length, i.e. the size of the embedded vector representation, on model performance using the Abraham octanol solubility dataset 245 as a model dataset.

8 Figure S6. Effect of changing the hidden layer size, i.e. the number of nodes in the hidden layer in the regression portion of the overall model architecture, on model performance using the Abraham octanol solubility dataset 245 as a model dataset. Performance does, however, depend strongly on the learning schedule. As a simple comparison, consider a shallower learning model which follows / and a steeper learning model which follows /. Both use a patience of 10 epochs (i.e., if the validation loss does not improve for 10 epochs, stop training). The loss (MSE) during training is shown in Figure S7 and a comparison of residuals in Figure S8. Due to the very small size of the internal validation dataset, training is prone to early stopping, so superior performance is observed with the more aggressive learning schedule. Figure S7. Comparison of mean squared error loss for the training and internal validation datasets during training for (left) a low, slow-decaying learning rate and (right) a high, fast-decaying learning rate.

9 Figure S8. Effect of changing the aggressiveness of the learning rate on model performance using the Abraham octanol solubility dataset 245 as a model dataset.

Convolutional Embedding of Attributed Molecular Graphs for Physical Property Prediction

Convolutional Embedding of Attributed Molecular Graphs for Physical Property Prediction Convolutional Embedding of Attributed Molecular Graphs for Physical Property Prediction The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters.

More information

Prediction of Organic Reaction Outcomes. Using Machine Learning

Prediction of Organic Reaction Outcomes. Using Machine Learning Prediction of Organic Reaction Outcomes Using Machine Learning Connor W. Coley, Regina Barzilay, Tommi S. Jaakkola, William H. Green, Supporting Information (SI) Klavs F. Jensen Approach Section S.. Data

More information

Holdout and Cross-Validation Methods Overfitting Avoidance

Holdout and Cross-Validation Methods Overfitting Avoidance Holdout and Cross-Validation Methods Overfitting Avoidance Decision Trees Reduce error pruning Cost-complexity pruning Neural Networks Early stopping Adjusting Regularizers via Cross-Validation Nearest

More information

Artificial Neural Networks D B M G. Data Base and Data Mining Group of Politecnico di Torino. Elena Baralis. Politecnico di Torino

Artificial Neural Networks D B M G. Data Base and Data Mining Group of Politecnico di Torino. Elena Baralis. Politecnico di Torino Artificial Neural Networks Data Base and Data Mining Group of Politecnico di Torino Elena Baralis Politecnico di Torino Artificial Neural Networks Inspired to the structure of the human brain Neurons as

More information

Introduction to Neural Networks

Introduction to Neural Networks CUONG TUAN NGUYEN SEIJI HOTTA MASAKI NAKAGAWA Tokyo University of Agriculture and Technology Copyright by Nguyen, Hotta and Nakagawa 1 Pattern classification Which category of an input? Example: Character

More information

Deep Learning & Artificial Intelligence WS 2018/2019

Deep Learning & Artificial Intelligence WS 2018/2019 Deep Learning & Artificial Intelligence WS 2018/2019 Linear Regression Model Model Error Function: Squared Error Has no special meaning except it makes gradients look nicer Prediction Ground truth / target

More information

CS 229 Project Final Report: Reinforcement Learning for Neural Network Architecture Category : Theory & Reinforcement Learning

CS 229 Project Final Report: Reinforcement Learning for Neural Network Architecture Category : Theory & Reinforcement Learning CS 229 Project Final Report: Reinforcement Learning for Neural Network Architecture Category : Theory & Reinforcement Learning Lei Lei Ruoxuan Xiong December 16, 2017 1 Introduction Deep Neural Network

More information

CS60010: Deep Learning

CS60010: Deep Learning CS60010: Deep Learning Sudeshna Sarkar Spring 2018 16 Jan 2018 FFN Goal: Approximate some unknown ideal function f : X! Y Ideal classifier: y = f*(x) with x and category y Feedforward Network: Define parametric

More information

CSE 417T: Introduction to Machine Learning. Final Review. Henry Chai 12/4/18

CSE 417T: Introduction to Machine Learning. Final Review. Henry Chai 12/4/18 CSE 417T: Introduction to Machine Learning Final Review Henry Chai 12/4/18 Overfitting Overfitting is fitting the training data more than is warranted Fitting noise rather than signal 2 Estimating! "#$

More information

Neural Networks and Deep Learning

Neural Networks and Deep Learning Neural Networks and Deep Learning Professor Ameet Talwalkar November 12, 2015 Professor Ameet Talwalkar Neural Networks and Deep Learning November 12, 2015 1 / 16 Outline 1 Review of last lecture AdaBoost

More information

Neural Networks, Computation Graphs. CMSC 470 Marine Carpuat

Neural Networks, Computation Graphs. CMSC 470 Marine Carpuat Neural Networks, Computation Graphs CMSC 470 Marine Carpuat Binary Classification with a Multi-layer Perceptron φ A = 1 φ site = 1 φ located = 1 φ Maizuru = 1 φ, = 2 φ in = 1 φ Kyoto = 1 φ priest = 0 φ

More information

Machine Learning for Large-Scale Data Analysis and Decision Making A. Neural Networks Week #6

Machine Learning for Large-Scale Data Analysis and Decision Making A. Neural Networks Week #6 Machine Learning for Large-Scale Data Analysis and Decision Making 80-629-17A Neural Networks Week #6 Today Neural Networks A. Modeling B. Fitting C. Deep neural networks Today s material is (adapted)

More information

CSC 578 Neural Networks and Deep Learning

CSC 578 Neural Networks and Deep Learning CSC 578 Neural Networks and Deep Learning Fall 2018/19 3. Improving Neural Networks (Some figures adapted from NNDL book) 1 Various Approaches to Improve Neural Networks 1. Cost functions Quadratic Cross

More information

Machine learning comes from Bayesian decision theory in statistics. There we want to minimize the expected value of the loss function.

Machine learning comes from Bayesian decision theory in statistics. There we want to minimize the expected value of the loss function. Bayesian learning: Machine learning comes from Bayesian decision theory in statistics. There we want to minimize the expected value of the loss function. Let y be the true label and y be the predicted

More information

Computational statistics

Computational statistics Computational statistics Lecture 3: Neural networks Thierry Denœux 5 March, 2016 Neural networks A class of learning methods that was developed separately in different fields statistics and artificial

More information

Introduction to Gaussian Process

Introduction to Gaussian Process Introduction to Gaussian Process CS 778 Chris Tensmeyer CS 478 INTRODUCTION 1 What Topic? Machine Learning Regression Bayesian ML Bayesian Regression Bayesian Non-parametric Gaussian Process (GP) GP Regression

More information

Machine Learning Basics

Machine Learning Basics Security and Fairness of Deep Learning Machine Learning Basics Anupam Datta CMU Spring 2019 Image Classification Image Classification Image classification pipeline Input: A training set of N images, each

More information

Convolutional Neural Networks

Convolutional Neural Networks Convolutional Neural Networks Books» http://www.deeplearningbook.org/ Books http://neuralnetworksanddeeplearning.com/.org/ reviews» http://www.deeplearningbook.org/contents/linear_algebra.html» http://www.deeplearningbook.org/contents/prob.html»

More information

Lecture 35: Optimization and Neural Nets

Lecture 35: Optimization and Neural Nets Lecture 35: Optimization and Neural Nets CS 4670/5670 Sean Bell DeepDream [Google, Inceptionism: Going Deeper into Neural Networks, blog 2015] Aside: CNN vs ConvNet Note: There are many papers that use

More information

Learning Molecular Fingerprints from the Graph Up

Learning Molecular Fingerprints from the Graph Up Learning Molecular Fingerprints from the Graph Up David Duvenaud, Dougal Maclaurin, Jorge Aguilera-Iparraguirre, Rafael Gómez-Bombarelli, Timothy Hirzel, Alán Aspuru-Guzik, Ryan P. Adams Motivation Want

More information

NONLINEAR CLASSIFICATION AND REGRESSION. J. Elder CSE 4404/5327 Introduction to Machine Learning and Pattern Recognition

NONLINEAR CLASSIFICATION AND REGRESSION. J. Elder CSE 4404/5327 Introduction to Machine Learning and Pattern Recognition NONLINEAR CLASSIFICATION AND REGRESSION Nonlinear Classification and Regression: Outline 2 Multi-Layer Perceptrons The Back-Propagation Learning Algorithm Generalized Linear Models Radial Basis Function

More information

Deep Feedforward Networks. Lecture slides for Chapter 6 of Deep Learning Ian Goodfellow Last updated

Deep Feedforward Networks. Lecture slides for Chapter 6 of Deep Learning  Ian Goodfellow Last updated 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

More information

Final Examination CS 540-2: Introduction to Artificial Intelligence

Final Examination CS 540-2: Introduction to Artificial Intelligence Final Examination CS 540-2: Introduction to Artificial Intelligence May 7, 2017 LAST NAME: SOLUTIONS FIRST NAME: Problem Score Max Score 1 14 2 10 3 6 4 10 5 11 6 9 7 8 9 10 8 12 12 8 Total 100 1 of 11

More information

Comments. Assignment 3 code released. Thought questions 3 due this week. Mini-project: hopefully you have started. implement classification algorithms

Comments. Assignment 3 code released. Thought questions 3 due this week. Mini-project: hopefully you have started. implement classification algorithms Neural networks Comments Assignment 3 code released implement classification algorithms use kernels for census dataset Thought questions 3 due this week Mini-project: hopefully you have started 2 Example:

More information

Sample questions for Fundamentals of Machine Learning 2018

Sample questions for Fundamentals of Machine Learning 2018 Sample questions for Fundamentals of Machine Learning 2018 Teacher: Mohammad Emtiyaz Khan A few important informations: In the final exam, no electronic devices are allowed except a calculator. Make sure

More information

Lecture 5: Logistic Regression. Neural Networks

Lecture 5: Logistic Regression. Neural Networks Lecture 5: Logistic Regression. Neural Networks Logistic regression Comparison with generative models Feed-forward neural networks Backpropagation Tricks for training neural networks COMP-652, Lecture

More information

Nonlinear Classification

Nonlinear Classification Nonlinear Classification INFO-4604, Applied Machine Learning University of Colorado Boulder October 5-10, 2017 Prof. Michael Paul Linear Classification Most classifiers we ve seen use linear functions

More information

Lecture 5 Neural models for NLP

Lecture 5 Neural models for NLP CS546: Machine Learning in NLP (Spring 2018) http://courses.engr.illinois.edu/cs546/ Lecture 5 Neural models for NLP Julia Hockenmaier juliahmr@illinois.edu 3324 Siebel Center Office hours: Tue/Thu 2pm-3pm

More information

The Perceptron Algorithm

The Perceptron Algorithm The Perceptron Algorithm Greg Grudic Greg Grudic Machine Learning Questions? Greg Grudic Machine Learning 2 Binary Classification A binary classifier is a mapping from a set of d inputs to a single output

More information

DEEP LEARNING AND NEURAL NETWORKS: BACKGROUND AND HISTORY

DEEP LEARNING AND NEURAL NETWORKS: BACKGROUND AND HISTORY DEEP LEARNING AND NEURAL NETWORKS: BACKGROUND AND HISTORY 1 On-line Resources http://neuralnetworksanddeeplearning.com/index.html Online book by Michael Nielsen http://matlabtricks.com/post-5/3x3-convolution-kernelswith-online-demo

More information

ECE521 Lectures 9 Fully Connected Neural Networks

ECE521 Lectures 9 Fully Connected Neural Networks ECE521 Lectures 9 Fully Connected Neural Networks Outline Multi-class classification Learning multi-layer neural networks 2 Measuring distance in probability space We learnt that the squared L2 distance

More information

Deep Neural Networks (3) Computational Graphs, Learning Algorithms, Initialisation

Deep Neural Networks (3) Computational Graphs, Learning Algorithms, Initialisation Deep Neural Networks (3) Computational Graphs, Learning Algorithms, Initialisation Steve Renals Machine Learning Practical MLP Lecture 5 16 October 2018 MLP Lecture 5 / 16 October 2018 Deep Neural Networks

More information

Introduction to Machine Learning Midterm Exam

Introduction to Machine Learning Midterm Exam 10-701 Introduction to Machine Learning Midterm Exam Instructors: Eric Xing, Ziv Bar-Joseph 17 November, 2015 There are 11 questions, for a total of 100 points. This exam is open book, open notes, but

More information

Neural networks and support vector machines

Neural networks and support vector machines Neural netorks and support vector machines Perceptron Input x 1 Weights 1 x 2 x 3... x D 2 3 D Output: sgn( x + b) Can incorporate bias as component of the eight vector by alays including a feature ith

More information

Deep Feedforward Networks

Deep Feedforward Networks Deep Feedforward Networks Liu Yang March 30, 2017 Liu Yang Short title March 30, 2017 1 / 24 Overview 1 Background A general introduction Example 2 Gradient based learning Cost functions Output Units 3

More information

Neural Networks. Yan Shao Department of Linguistics and Philology, Uppsala University 7 December 2016

Neural Networks. Yan Shao Department of Linguistics and Philology, Uppsala University 7 December 2016 Neural Networks Yan Shao Department of Linguistics and Philology, Uppsala University 7 December 2016 Outline Part 1 Introduction Feedforward Neural Networks Stochastic Gradient Descent Computational Graph

More information

Numerical Learning Algorithms

Numerical Learning Algorithms Numerical Learning Algorithms Example SVM for Separable Examples.......................... Example SVM for Nonseparable Examples....................... 4 Example Gaussian Kernel SVM...............................

More information

10-701/ Machine Learning, Fall

10-701/ Machine Learning, Fall 0-70/5-78 Machine Learning, Fall 2003 Homework 2 Solution If you have questions, please contact Jiayong Zhang .. (Error Function) The sum-of-squares error is the most common training

More information

Retrieval of Cloud Top Pressure

Retrieval of Cloud Top Pressure Master Thesis in Statistics and Data Mining Retrieval of Cloud Top Pressure Claudia Adok Division of Statistics and Machine Learning Department of Computer and Information Science Linköping University

More information

COMPUTATIONAL INTELLIGENCE (INTRODUCTION TO MACHINE LEARNING) SS16

COMPUTATIONAL INTELLIGENCE (INTRODUCTION TO MACHINE LEARNING) SS16 COMPUTATIONAL INTELLIGENCE (INTRODUCTION TO MACHINE LEARNING) SS6 Lecture 3: Classification with Logistic Regression Advanced optimization techniques Underfitting & Overfitting Model selection (Training-

More information

Machine Learning Concepts in Chemoinformatics

Machine Learning Concepts in Chemoinformatics Machine Learning Concepts in Chemoinformatics Martin Vogt B-IT Life Science Informatics Rheinische Friedrich-Wilhelms-Universität Bonn BigChem Winter School 2017 25. October Data Mining in Chemoinformatics

More information

Introduction to Deep Learning CMPT 733. Steven Bergner

Introduction to Deep Learning CMPT 733. Steven Bergner Introduction to Deep Learning CMPT 733 Steven Bergner Overview Renaissance of artificial neural networks Representation learning vs feature engineering Background Linear Algebra, Optimization Regularization

More information

Recurrent Neural Networks with Flexible Gates using Kernel Activation Functions

Recurrent Neural Networks with Flexible Gates using Kernel Activation Functions 2018 IEEE International Workshop on Machine Learning for Signal Processing (MLSP 18) Recurrent Neural Networks with Flexible Gates using Kernel Activation Functions Authors: S. Scardapane, S. Van Vaerenbergh,

More information

Deep Feedforward Networks. Seung-Hoon Na Chonbuk National University

Deep Feedforward Networks. Seung-Hoon Na Chonbuk National University Deep Feedforward Networks Seung-Hoon Na Chonbuk National University Neural Network: Types Feedforward neural networks (FNN) = Deep feedforward networks = multilayer perceptrons (MLP) No feedback connections

More information

Introduction to Convolutional Neural Networks 2018 / 02 / 23

Introduction to Convolutional Neural Networks 2018 / 02 / 23 Introduction to Convolutional Neural Networks 2018 / 02 / 23 Buzzword: CNN Convolutional neural networks (CNN, ConvNet) is a class of deep, feed-forward (not recurrent) artificial neural networks that

More information

PATTERN CLASSIFICATION

PATTERN CLASSIFICATION PATTERN CLASSIFICATION Second Edition Richard O. Duda Peter E. Hart David G. Stork A Wiley-lnterscience Publication JOHN WILEY & SONS, INC. New York Chichester Weinheim Brisbane Singapore Toronto CONTENTS

More information

Neural Networks. Advanced data-mining. Yongdai Kim. Department of Statistics, Seoul National University, South Korea

Neural Networks. Advanced data-mining. Yongdai Kim. Department of Statistics, Seoul National University, South Korea Neural Networks Advanced data-mining Yongdai Kim Department of Statistics, Seoul National University, South Korea What is Neural Networks? One of supervised learning method using one or more hidden layer.

More information

Serious limitations of (single-layer) perceptrons: Cannot learn non-linearly separable tasks. Cannot approximate (learn) non-linear functions

Serious limitations of (single-layer) perceptrons: Cannot learn non-linearly separable tasks. Cannot approximate (learn) non-linear functions BACK-PROPAGATION NETWORKS Serious limitations of (single-layer) perceptrons: Cannot learn non-linearly separable tasks Cannot approximate (learn) non-linear functions Difficult (if not impossible) to design

More information

Evaluation. Andrea Passerini Machine Learning. Evaluation

Evaluation. Andrea Passerini Machine Learning. Evaluation Andrea Passerini passerini@disi.unitn.it Machine Learning Basic concepts requires to define performance measures to be optimized Performance of learning algorithms cannot be evaluated on entire domain

More information

A Reservoir Sampling Algorithm with Adaptive Estimation of Conditional Expectation

A Reservoir Sampling Algorithm with Adaptive Estimation of Conditional Expectation A Reservoir Sampling Algorithm with Adaptive Estimation of Conditional Expectation Vu Malbasa and Slobodan Vucetic Abstract Resource-constrained data mining introduces many constraints when learning from

More information

epochs epochs

epochs epochs Neural Network Experiments To illustrate practical techniques, I chose to use the glass dataset. This dataset has 214 examples and 6 classes. Here are 4 examples from the original dataset. The last values

More information

Machine Learning and Data Mining. Linear regression. Kalev Kask

Machine Learning and Data Mining. Linear regression. Kalev Kask Machine Learning and Data Mining Linear regression Kalev Kask Supervised learning Notation Features x Targets y Predictions ŷ Parameters q Learning algorithm Program ( Learner ) Change q Improve performance

More information

Speaker Representation and Verification Part II. by Vasileios Vasilakakis

Speaker Representation and Verification Part II. by Vasileios Vasilakakis Speaker Representation and Verification Part II by Vasileios Vasilakakis Outline -Approaches of Neural Networks in Speaker/Speech Recognition -Feed-Forward Neural Networks -Training with Back-propagation

More information

Neural Reasoning for Chemical-Chemical Interaction

Neural Reasoning for Chemical-Chemical Interaction Neural Reasoning for Chemical-Chemical Interaction Trang Pham, Truyen Tran, and Svetha Venkatesh Applied AI Institute, Deakin University, Geelong, Australia {phtra,truyen.tran,svetha.venkatesh}@deakin.edu.au

More information

Based on the original slides of Hung-yi Lee

Based on the original slides of Hung-yi Lee Based on the original slides of Hung-yi Lee Google Trends Deep learning obtains many exciting results. Can contribute to new Smart Services in the Context of the Internet of Things (IoT). IoT Services

More information

Article from. Predictive Analytics and Futurism. July 2016 Issue 13

Article from. Predictive Analytics and Futurism. July 2016 Issue 13 Article from Predictive Analytics and Futurism July 2016 Issue 13 Regression and Classification: A Deeper Look By Jeff Heaton Classification and regression are the two most common forms of models fitted

More information

Evaluation requires to define performance measures to be optimized

Evaluation requires to define performance measures to be optimized Evaluation Basic concepts Evaluation requires to define performance measures to be optimized Performance of learning algorithms cannot be evaluated on entire domain (generalization error) approximation

More information

More on Neural Networks

More on Neural Networks More on Neural Networks Yujia Yan Fall 2018 Outline Linear Regression y = Wx + b (1) Linear Regression y = Wx + b (1) Polynomial Regression y = Wφ(x) + b (2) where φ(x) gives the polynomial basis, e.g.,

More information

Deep learning / Ian Goodfellow, Yoshua Bengio and Aaron Courville. - Cambridge, MA ; London, Spis treści

Deep learning / Ian Goodfellow, Yoshua Bengio and Aaron Courville. - Cambridge, MA ; London, Spis treści Deep learning / Ian Goodfellow, Yoshua Bengio and Aaron Courville. - Cambridge, MA ; London, 2017 Spis treści Website Acknowledgments Notation xiii xv xix 1 Introduction 1 1.1 Who Should Read This Book?

More information

Supplementary Figure 1: Chemical compound space. Errors depending on the size of the training set for models with T = 1, 2, 3 interaction passes

Supplementary Figure 1: Chemical compound space. Errors depending on the size of the training set for models with T = 1, 2, 3 interaction passes 9 8 7 6 5 4 3 2 1 0 10 3 10 4 10 5 6 5 4 3 2 1 0 10000 25000 50000 100000 Supplementary Figure 1: Chemical compound space. Errors depending on the size of the training set for models with T = 1, 2, 3 interaction

More information

EPL442: Computational

EPL442: Computational EPL442: Computational Learning Systems Lab 2 Vassilis Vassiliades Department of Computer Science University of Cyprus Outline Artificial Neuron Feedforward Neural Network Back-propagation Algorithm Notes

More information

ESS2222. Lecture 4 Linear model

ESS2222. Lecture 4 Linear model ESS2222 Lecture 4 Linear model Hosein Shahnas University of Toronto, Department of Earth Sciences, 1 Outline Logistic Regression Predicting Continuous Target Variables Support Vector Machine (Some Details)

More information

Neural Networks with Applications to Vision and Language. Feedforward Networks. Marco Kuhlmann

Neural Networks with Applications to Vision and Language. Feedforward Networks. Marco Kuhlmann Neural Networks with Applications to Vision and Language Feedforward Networks Marco Kuhlmann Feedforward networks Linear separability x 2 x 2 0 1 0 1 0 0 x 1 1 0 x 1 linearly separable not linearly separable

More information

Neural Network Tutorial & Application in Nuclear Physics. Weiguang Jiang ( 蒋炜光 ) UTK / ORNL

Neural Network Tutorial & Application in Nuclear Physics. Weiguang Jiang ( 蒋炜光 ) UTK / ORNL Neural Network Tutorial & Application in Nuclear Physics Weiguang Jiang ( 蒋炜光 ) UTK / ORNL Machine Learning Logistic Regression Gaussian Processes Neural Network Support vector machine Random Forest Genetic

More information

Source localization in an ocean waveguide using supervised machine learning

Source localization in an ocean waveguide using supervised machine learning Source localization in an ocean waveguide using supervised machine learning Haiqiang Niu, Emma Reeves, and Peter Gerstoft Scripps Institution of Oceanography, UC San Diego Part I Localization on Noise09

More information

Need for Deep Networks Perceptron. Can only model linear functions. Kernel Machines. Non-linearity provided by kernels

Need for Deep Networks Perceptron. Can only model linear functions. Kernel Machines. Non-linearity provided by kernels Need for Deep Networks Perceptron Can only model linear functions Kernel Machines Non-linearity provided by kernels Need to design appropriate kernels (possibly selecting from a set, i.e. kernel learning)

More information

Neural Networks. CSE 6363 Machine Learning Vassilis Athitsos Computer Science and Engineering Department University of Texas at Arlington

Neural Networks. CSE 6363 Machine Learning Vassilis Athitsos Computer Science and Engineering Department University of Texas at Arlington Neural Networks CSE 6363 Machine Learning Vassilis Athitsos Computer Science and Engineering Department University of Texas at Arlington 1 Perceptrons x 0 = 1 x 1 x 2 z = h w T x Output: z x D A perceptron

More information

Logistic Regression & Neural Networks

Logistic Regression & Neural Networks Logistic Regression & Neural Networks CMSC 723 / LING 723 / INST 725 Marine Carpuat Slides credit: Graham Neubig, Jacob Eisenstein Logistic Regression Perceptron & Probabilities What if we want a probability

More information

Machine Learning for Computer Vision 8. Neural Networks and Deep Learning. Vladimir Golkov Technical University of Munich Computer Vision Group

Machine Learning for Computer Vision 8. Neural Networks and Deep Learning. Vladimir Golkov Technical University of Munich Computer Vision Group Machine Learning for Computer Vision 8. Neural Networks and Deep Learning Vladimir Golkov Technical University of Munich Computer Vision Group INTRODUCTION Nonlinear Coordinate Transformation http://cs.stanford.edu/people/karpathy/convnetjs/

More information

Linear Regression (continued)

Linear Regression (continued) Linear Regression (continued) Professor Ameet Talwalkar Professor Ameet Talwalkar CS260 Machine Learning Algorithms February 6, 2017 1 / 39 Outline 1 Administration 2 Review of last lecture 3 Linear regression

More information

From CDF to PDF A Density Estimation Method for High Dimensional Data

From CDF to PDF A Density Estimation Method for High Dimensional Data From CDF to PDF A Density Estimation Method for High Dimensional Data Shengdong Zhang Simon Fraser University sza75@sfu.ca arxiv:1804.05316v1 [stat.ml] 15 Apr 2018 April 17, 2018 1 Introduction Probability

More information

Need for Deep Networks Perceptron. Can only model linear functions. Kernel Machines. Non-linearity provided by kernels

Need for Deep Networks Perceptron. Can only model linear functions. Kernel Machines. Non-linearity provided by kernels Need for Deep Networks Perceptron Can only model linear functions Kernel Machines Non-linearity provided by kernels Need to design appropriate kernels (possibly selecting from a set, i.e. kernel learning)

More information

The exam is closed book, closed notes except your one-page (two sides) or two-page (one side) crib sheet.

The exam is closed book, closed notes except your one-page (two sides) or two-page (one side) crib sheet. CS 189 Spring 013 Introduction to Machine Learning Final You have 3 hours for the exam. The exam is closed book, closed notes except your one-page (two sides) or two-page (one side) crib sheet. Please

More information

Regression. Goal: Learn a mapping from observations (features) to continuous labels given a training set (supervised learning)

Regression. Goal: Learn a mapping from observations (features) to continuous labels given a training set (supervised learning) Linear Regression Regression Goal: Learn a mapping from observations (features) to continuous labels given a training set (supervised learning) Example: Height, Gender, Weight Shoe Size Audio features

More information

CSC242: Intro to AI. Lecture 21

CSC242: Intro to AI. Lecture 21 CSC242: Intro to AI Lecture 21 Administrivia Project 4 (homeworks 18 & 19) due Mon Apr 16 11:59PM Posters Apr 24 and 26 You need an idea! You need to present it nicely on 2-wide by 4-high landscape pages

More information

Regression. Goal: Learn a mapping from observations (features) to continuous labels given a training set (supervised learning)

Regression. Goal: Learn a mapping from observations (features) to continuous labels given a training set (supervised learning) Linear Regression Regression Goal: Learn a mapping from observations (features) to continuous labels given a training set (supervised learning) Example: Height, Gender, Weight Shoe Size Audio features

More information

Support Vector Machines

Support Vector Machines Support Vector Machines INFO-4604, Applied Machine Learning University of Colorado Boulder September 28, 2017 Prof. Michael Paul Today Two important concepts: Margins Kernels Large Margin Classification

More information

CS 6501: Deep Learning for Computer Graphics. Basics of Neural Networks. Connelly Barnes

CS 6501: Deep Learning for Computer Graphics. Basics of Neural Networks. Connelly Barnes CS 6501: Deep Learning for Computer Graphics Basics of Neural Networks Connelly Barnes Overview Simple neural networks Perceptron Feedforward neural networks Multilayer perceptron and properties Autoencoders

More information

Reading Group on Deep Learning Session 1

Reading Group on Deep Learning Session 1 Reading Group on Deep Learning Session 1 Stephane Lathuiliere & Pablo Mesejo 2 June 2016 1/31 Contents Introduction to Artificial Neural Networks to understand, and to be able to efficiently use, the popular

More information

Day 3 Lecture 3. Optimizing deep networks

Day 3 Lecture 3. Optimizing deep networks Day 3 Lecture 3 Optimizing deep networks Convex optimization A function is convex if for all α [0,1]: f(x) Tangent line Examples Quadratics 2-norms Properties Local minimum is global minimum x Gradient

More information

Final Examination CS540-2: Introduction to Artificial Intelligence

Final Examination CS540-2: Introduction to Artificial Intelligence Final Examination CS540-2: Introduction to Artificial Intelligence May 9, 2018 LAST NAME: SOLUTIONS FIRST NAME: Directions 1. This exam contains 33 questions worth a total of 100 points 2. Fill in your

More information

Neural Nets Supervised learning

Neural Nets Supervised learning 6.034 Artificial Intelligence Big idea: Learning as acquiring a function on feature vectors Background Nearest Neighbors Identification Trees Neural Nets Neural Nets Supervised learning y s(z) w w 0 w

More information

CS 1674: Intro to Computer Vision. Final Review. Prof. Adriana Kovashka University of Pittsburgh December 7, 2016

CS 1674: Intro to Computer Vision. Final Review. Prof. Adriana Kovashka University of Pittsburgh December 7, 2016 CS 1674: Intro to Computer Vision Final Review Prof. Adriana Kovashka University of Pittsburgh December 7, 2016 Final info Format: multiple-choice, true/false, fill in the blank, short answers, apply an

More information

Neural Networks. Nethra Sambamoorthi, Ph.D. Jan CRMportals Inc., Nethra Sambamoorthi, Ph.D. Phone:

Neural Networks. Nethra Sambamoorthi, Ph.D. Jan CRMportals Inc., Nethra Sambamoorthi, Ph.D. Phone: Neural Networks Nethra Sambamoorthi, Ph.D Jan 2003 CRMportals Inc., Nethra Sambamoorthi, Ph.D Phone: 732-972-8969 Nethra@crmportals.com What? Saying it Again in Different ways Artificial neural network

More information

Active and Semi-supervised Kernel Classification

Active and Semi-supervised Kernel Classification Active and Semi-supervised Kernel Classification Zoubin Ghahramani Gatsby Computational Neuroscience Unit University College London Work done in collaboration with Xiaojin Zhu (CMU), John Lafferty (CMU),

More information

Jakub Hajic Artificial Intelligence Seminar I

Jakub Hajic Artificial Intelligence Seminar I Jakub Hajic Artificial Intelligence Seminar I. 11. 11. 2014 Outline Key concepts Deep Belief Networks Convolutional Neural Networks A couple of questions Convolution Perceptron Feedforward Neural Network

More information

Ch.6 Deep Feedforward Networks (2/3)

Ch.6 Deep Feedforward Networks (2/3) Ch.6 Deep Feedforward Networks (2/3) 16. 10. 17. (Mon.) System Software Lab., Dept. of Mechanical & Information Eng. Woonggy Kim 1 Contents 6.3. Hidden Units 6.3.1. Rectified Linear Units and Their Generalizations

More information

Introduction to Machine Learning Spring 2018 Note Neural Networks

Introduction to Machine Learning Spring 2018 Note Neural Networks CS 189 Introduction to Machine Learning Spring 2018 Note 14 1 Neural Networks Neural networks are a class of compositional function approximators. They come in a variety of shapes and sizes. In this class,

More information

CSC321 Lecture 2: Linear Regression

CSC321 Lecture 2: Linear Regression CSC32 Lecture 2: Linear Regression Roger Grosse Roger Grosse CSC32 Lecture 2: Linear Regression / 26 Overview First learning algorithm of the course: linear regression Task: predict scalar-valued targets,

More information

SPSS, University of Texas at Arlington. Topics in Machine Learning-EE 5359 Neural Networks

SPSS, University of Texas at Arlington. Topics in Machine Learning-EE 5359 Neural Networks Topics in Machine Learning-EE 5359 Neural Networks 1 The Perceptron Output: A perceptron is a function that maps D-dimensional vectors to real numbers. For notational convenience, we add a zero-th dimension

More information

Demand and Trip Prediction in Bike Share Systems

Demand and Trip Prediction in Bike Share Systems Demand and Trip Prediction in Bike Share Systems Team members: Zhaonan Qu SUNet ID: zhaonanq December 16, 2017 1 Abstract 2 Introduction Bike Share systems are becoming increasingly popular in urban areas.

More information

Regression Adjustment with Artificial Neural Networks

Regression Adjustment with Artificial Neural Networks Regression Adjustment with Artificial Neural Networks Age of Big Data: data comes in a rate and in a variety of types that exceed our ability to analyse it Texts, image, speech, video Real motivation:

More information

Pattern Recognition and Machine Learning

Pattern Recognition and Machine Learning Christopher M. Bishop Pattern Recognition and Machine Learning ÖSpri inger Contents Preface Mathematical notation Contents vii xi xiii 1 Introduction 1 1.1 Example: Polynomial Curve Fitting 4 1.2 Probability

More information

Introduction to Machine Learning Midterm Exam Solutions

Introduction to Machine Learning Midterm Exam Solutions 10-701 Introduction to Machine Learning Midterm Exam Solutions Instructors: Eric Xing, Ziv Bar-Joseph 17 November, 2015 There are 11 questions, for a total of 100 points. This exam is open book, open notes,

More information

Black-box α-divergence Minimization

Black-box α-divergence Minimization Black-box α-divergence Minimization José Miguel Hernández-Lobato, Yingzhen Li, Daniel Hernández-Lobato, Thang Bui, Richard Turner, Harvard University, University of Cambridge, Universidad Autónoma de Madrid.

More information

A General Method for Combining Predictors Tested on Protein Secondary Structure Prediction

A General Method for Combining Predictors Tested on Protein Secondary Structure Prediction A General Method for Combining Predictors Tested on Protein Secondary Structure Prediction Jakob V. Hansen Department of Computer Science, University of Aarhus Ny Munkegade, Bldg. 540, DK-8000 Aarhus C,

More information

Midterm: CS 6375 Spring 2015 Solutions

Midterm: CS 6375 Spring 2015 Solutions Midterm: CS 6375 Spring 2015 Solutions The exam is closed book. You are allowed a one-page cheat sheet. Answer the questions in the spaces provided on the question sheets. If you run out of room for an

More information

ECE521 week 3: 23/26 January 2017

ECE521 week 3: 23/26 January 2017 ECE521 week 3: 23/26 January 2017 Outline Probabilistic interpretation of linear regression - Maximum likelihood estimation (MLE) - Maximum a posteriori (MAP) estimation Bias-variance trade-off Linear

More information

Neural Networks. David Rosenberg. July 26, New York University. David Rosenberg (New York University) DS-GA 1003 July 26, / 35

Neural Networks. David Rosenberg. July 26, New York University. David Rosenberg (New York University) DS-GA 1003 July 26, / 35 Neural Networks David Rosenberg New York University July 26, 2017 David Rosenberg (New York University) DS-GA 1003 July 26, 2017 1 / 35 Neural Networks Overview Objectives What are neural networks? How

More information

The exam is closed book, closed notes except your one-page cheat sheet.

The exam is closed book, closed notes except your one-page cheat sheet. CS 189 Fall 2015 Introduction to Machine Learning Final Please do not turn over the page before you are instructed to do so. You have 2 hours and 50 minutes. Please write your initials on the top-right

More information