Adaptive Stimulation Design for the Treatment of Epilepsy Joelle Pineau School of Computer Science, McGill University Montreal, QC CANADA Jointly with.

Slides:



Advertisements
Similar presentations
Pat Langley Computational Learning Laboratory Center for the Study of Language and Information Stanford University, Stanford, California
Advertisements

Introduction to Neural Networks
Short reading for Thursday Job talk at 1:30pm in ETRL 101 Kuka robotics –
Université du Québec École de technologie supérieure Face Recognition in Video Using What- and-Where Fusion Neural Network Mamoudou Barry and Eric Granger.
A machine learning perspective on neural networks and learning tools
Reinforcement learning
Lecture 18: Temporal-Difference Learning
NEU Neural Computing MSc Natural Computation Department of Computer Science University of York.
Lirong Xia Reinforcement Learning (2) Tue, March 21, 2014.
On-line learning and Boosting
Extraction and Transfer of Knowledge in Reinforcement Learning A.LAZARIC Inria “30 minutes de Science” Seminars SequeL Inria Lille – Nord Europe December.
Spike Train Statistics Sabri IPM. Review of spike train  Extracting information from spike trains  Noisy environment:  in vitro  in vivo  measurement.
FilterBoost: Regression and Classification on Large Datasets Joseph K. Bradley 1 and Robert E. Schapire 2 1 Carnegie Mellon University 2 Princeton University.
Treating Epilepsy via Adaptive Neurostimulation Joelle Pineau, PhD School of Computer Science, McGill University Congress of the Canadian Neurological.
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.
Failure Prediction in Hardware Systems Douglas Turnbull Neil Alldrin CSE 221: Operating System Final Project Fall
Generic Object Detection using Feature Maps Oscar Danielsson Stefan Carlsson
Reinforcement Learning Mitchell, Ch. 13 (see also Barto & Sutton book on-line)
Integrating POMDP and RL for a Two Layer Simulated Robot Architecture Presented by Alp Sardağ.
1 Hybrid Agent-Based Modeling: Architectures,Analyses and Applications (Stage One) Li, Hailin.
CSSE463: Image Recognition Day 31 Due tomorrow night – Project plan Due tomorrow night – Project plan Evidence that you’ve tried something and what specifically.
Chapter 6: Temporal Difference Learning
Statistical Learning: Pattern Classification, Prediction, and Control Peter Bartlett August 2002, UC Berkeley CIS.
For Better Accuracy Eick: Ensemble Learning
Artificial Intelligence (AI) Addition to the lecture 11.
Machine Learning CS 165B Spring 2012
Processing of large document collections Part 2 (Text categorization) Helena Ahonen-Myka Spring 2006.
Abstract The emergence of big data and deep learning is enabling the ability to automatically learn how to interpret EEGs from a big data archive. The.
CSSE463: Image Recognition Day 27 This week This week Last night: k-means lab due. Last night: k-means lab due. Today: Classification by “boosting” Today:
Machine Learning1 Machine Learning: Summary Greg Grudic CSCI-4830.
Functional Brain Signal Processing: EEG & fMRI Lesson 8 Kaushik Majumdar Indian Statistical Institute Bangalore Center M.Tech.
Using Neural Networks in Database Mining Tino Jimenez CS157B MW 9-10:15 February 19, 2009.
Benk Erika Kelemen Zsolt
Machine Learning.
1 Machine Learning 1.Where does machine learning fit in computer science? 2.What is machine learning? 3.Where can machine learning be applied? 4.Should.
Today Ensemble Methods. Recap of the course. Classifier Fusion
Ensemble Learning Spring 2009 Ben-Gurion University of the Negev.
Connections 2009 Biologically inspired stimulation for epilepsy control Supervisor: Berj Bardakjian Presenters: Sinisa Colic and Josh Dian.
Design Principles for Creating Human-Shapable Agents W. Bradley Knox, Ian Fasel, and Peter Stone The University of Texas at Austin Department of Computer.
Decision Making Under Uncertainty Lec #8: Reinforcement Learning UIUC CS 598: Section EA Professor: Eyal Amir Spring Semester 2006 Most slides by Jeremy.
Tony Jebara, Columbia University Advanced Machine Learning & Perception Instructor: Tony Jebara.
CSSE463: Image Recognition Day 33 This week This week Today: Classification by “boosting” Today: Classification by “boosting” Yoav Freund and Robert Schapire.
Team Dogecoin: An Experience in Predicting Hospital Readmissions Acknowledgements The Problem Hospitals in the UK must keep track of which patients, once.
Are worms more complex than humans? Rodrigo Quian Quiroga Sloan-Swartz Center for Theoretical Neurobiology. Caltech.
Neural Networks Presented by M. Abbasi Course lecturer: Dr.Tohidkhah.
COMP24111: Machine Learning Ensemble Models Gavin Brown
EEG processing based on IFAST system and Artificial Neural Networks for early detection of Alzheimer’s disease.
WHAT IS DATA MINING?  The process of automatically extracting useful information from large amounts of data.  Uses traditional data analysis techniques.
Pattern Recognition. What is Pattern Recognition? Pattern recognition is a sub-topic of machine learning. PR is the science that concerns the description.
Deep Learning and Deep Reinforcement Learning. Topics 1.Deep learning with convolutional neural networks 2.Learning to play Atari video games with Deep.
1 Azhari, Dr Computer Science UGM. Human brain is a densely interconnected network of approximately neurons, each connected to, on average, 10 4.
A recurring neurological disorder characterized by random firing of nerve cells in the brain which cause a temporary shutdown of normal brain function.
Adaboost (Adaptive boosting) Jo Yeong-Jun Schapire, Robert E., and Yoram Singer. "Improved boosting algorithms using confidence- rated predictions."
Combining Models Foundations of Algorithms and Machine Learning (CS60020), IIT KGP, 2017: Indrajit Bhattacharya.
Reinforcement Learning
Deep Reinforcement Learning
Advanced information retreival
Chapter 6: Temporal Difference Learning
Epileptic Seizure Prediction
COMP61011 : Machine Learning Ensemble Models
Vijay Srinivasan Thomas Phan
"Playing Atari with deep reinforcement learning."
Joelle Pineau: General info
Dr. Unnikrishnan P.C. Professor, EEE
October 6, 2011 Dr. Itamar Arel College of Engineering
Chapter 6: Temporal Difference Learning
Deep Reinforcement Learning
CS 188: Artificial Intelligence Fall 2008
Unsupervised Perceptual Rewards For Imitation Learning
David Kauchak CS158 – Spring 2019
Presentation transcript:

Adaptive Stimulation Design for the Treatment of Epilepsy Joelle Pineau School of Computer Science, McGill University Montreal, QC CANADA Jointly with Robert Vincent, Aaron Courville, Massimo Avoli SAMSI Program on Challenges in Dynamic Treatment Regimes and Multistage Decision-Making June 21, 2007

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Background Vagus nerve and deep brain stimulation are used to treat various neurological disorders, including epilepsy. Images from and

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Project goal Problem: Existing devices offer limited control and do not adapt to the patients condition. Idea: Create an improved class of devices with closed-loop control.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy A reinforcement formulation Objective: Minimize occurrence of seizures and overall amount of stimulation. The MDP model: –States, s t : recordings of electrical activity –Actions, a t : stimulation (frequency, voltage, location) –Transitions, P(s t |s t-1, a t ) : unknown –Rewards, r t : large cost for seizures, small cost for stimulation s t-1 atat a t+1 … r t-1 stst rtrt s t+1 r t+1 a t-1

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Technical challenges 1.Investigate supervised learning methods for automatic seizure detection to inform choice of good state representation. 2.Design a computational (generative) model of epilepsy. 3.Run reinforcement learning using online data from the computational model. 4.Run reinforcement learning using batch data from an in-vitro model.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Recordings of electrical activity Recorded from single sensing electrode in in-vitro model of epilepsy. Raw data: 4096-sample frames windowed, normalized, FFT x t = 83 real-valued features: mean, range, energy, 80 FFT magnitudes y t ={normal, spike, seizure} hand-labeled for each frame.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Seizure detection Long literature on seizure detection using supervised learning (e.g. RBFs, wavelets, neural networks, energy methods). »Usually uninterpretable results. Related literature in time-series prediction (e.g. HMMs, CRFs). »Requires feature selection. Instead, focus on ensemble boosting methods: –Extend standard multi-class Adaboost [Schapire&Singer, 1999] to a recurrent formulation, in which y t = f (x t, y t-1, y t-2, …, y t-k ). »Relatively interpretable results (assuming simple learners). »Requires no prior model of the distribution of features.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Classification accuracy y t = f (x t, y t-1 )

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Feature extraction = insight into state design In all recurrent examples, first weak hypothesis recruited was frequency band 62 or 63 (= Hz). –High value normal –Low value spike Frequency bands 6-8 (=7-10 Hz) also consistently recruited early. –High value spike –Low value normal Often recruited in the first 20 rounds is Energy. –High value spike In recurrent Adaboost, prior label often recruited early and acts as memory. –High spike prior spike etc.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Technical challenges 1.Investigate supervised learning methods for automatic seizure detection to inform choice of good state representation. 2.Design a computational (generative) model of epilepsy. 3.Run reinforcement learning using online data from the computational model. 4.Run reinforcement learning using batch data from an in-vitro model.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy A computational model of epilepsy Aims of computational modeling: –To understand basic mechanisms of epilepsy sufficiently to design a good state representation. –To understand the appropriate class of policies to consider. –To have an inexpensive testing environment for RL algorithms. A word of caution: –Epilepsy is a complex disease. Many animal models are found in the literature. We focus on model by [ Avoli et al., 2002], which we will use for online exploration in-vitro.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Model overview (a) Assume a simple stochastic neuron model (leaky integrate-and-fire). (b) Connect many neurons in small-world network configuration. (a) Leaky integrate-and-fire neuron model(b) Network structure

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Model overview (contd) (c) Define a sensor model: –Voltage measurement over a patch of adjacent neurons (neuron contribution falls off in inverse-square relationship with distance from the patch center). (d) Define the stimulation model: –Input current applied uniformely to a patch of neural units. i stim v sensor

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Traces from the computational model # neurons firing sensor voltage

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Results from fixed stimulation strategies 0 Hz 0.5 Hz 1.0 Hz 2.0 Hz 4.0 Hz 5.0 Hz

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Technical challenges 1.Investigate supervised learning methods for automatic seizure detection to inform choice of good state representation. 2.Design a computational (generative) model of epilepsy. 3.Run reinforcement learning using online data from the computational model. 4.Run reinforcement learning using batch data from an in-vitro model.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Reinforcement learning agent States: s t = {sensor reading} x {# time steps since last stimulation} discretized into small number of (independent) features Actions: a 1 = Stimulator on for 1 time-step. a 2 = Stimulator off for 1 time-step. Rewards: R(s,a) = -100for a seizure (i.e. firing count > 50) R(s,a) = -10for a stimulation R(s,a) = 0otherwise Training details: »Online data, finite horizon (60 x 100-sec traces), -greedy exploration, Sarsa( ) with eligibility

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Results: Learning (a) After 1 learning episode(b) After 10 learning episodes

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Results: Expected return

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Results: Policy

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Technical challenges 1.Investigate supervised learning methods for automatic seizure detection to inform choice of good state representation. 2.Design a computational (generative) model of epilepsy. 3.Run reinforcement learning using online data from the computational model. 4.Run reinforcement learning using batch data from an in-vitro model.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Learning from batch in-vitro data We have data of the type used in the classification task, but which includes fixed-policy stimulation. States: 83 real-valued features (electrical signal sampled + FFT) Actions: {0 Hz, 0.5 Hz, 1.0 Hz, 2.0 Hz, 5.0 Hz} Training: batch (~30 x 60-sec traces), tree-based regression. No results

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Wrap-up Preliminary evidence for: –ability to detect seizures using in-vitro data. –generation of synthetic data with epileptiform behavior. –controllability of the stimulator in the computational model. Results apply to a specific model of epilepsy - generalization to other models is unknown. Issues I did not discuss today: –Designing the reward function. –Using the features found in Adaboost within the RL agent. –Transferring what we have learned in the computational model to the biological model. –Learning from very few data points.

SAMSI 2007: Adaptive Stimulation Design for the Treatment of Epilepsy Available data Recordings used for classification (no stimulation) – Recordings used for batch reinforcement learning (fixed policy stimulation) –Subject to approval by neuroscientists (1-2 weeks delay). Generative model of epilepsy –Still under development, but publicly available soon (1-2 months).