CS8803: Statistical Techniques in Robotics Byron Boots. Predicting With Hilbert Space Embeddings

Size: px
Start display at page:

Download "CS8803: Statistical Techniques in Robotics Byron Boots. Predicting With Hilbert Space Embeddings"

Transcription

1 CS8803: Statistical Techniques in Robotics Byron Boots Predicting With Hilbert Space Embeddings 1

2 HSE: Gram/Kernel Matrices bc YX = 1 N bc XX = 1 N NX (y i )'(x i ) > = 1 N Y > X 2 R 1 1 i=1 NX i=1 '(x i )'(x i ) > = 1 N X > X 2 R 1 1 µ x = '(x) 2 R 1 1 µ Y = 1 N NX i=1 (y i ) 2 R 1 1 Would like to calculate: µ Y x = b C YX b C 1 XX µ x 2 R 1 1 CS8803: STR Fast Approximate Kernel Methods 2

3 HSE: Gram/Kernel Matrices µ Y x = b C YX b C 1 XX µ x ˆµ Y x = Y > X X > X + I 1 '(x) (Woodbury) Matrix Inversion Lemma = Y ( > X X + NI) 1 > X'(x) = Y (G XX + NI) 1 G XX (:,i) where G XX = > X X 2 R N N G XX (:,i)= > X'(x i ) 2 R N 1 CS8803: STR Fast Approximate Kernel Methods 3

4 Predicting via Kernel Regression E[f(Y ) x] =hf,µ Y x i E[f(Y ) x] = b f(y ), (Y ) b C 1 YY b C YX b C 1 XX µ x b b C CS8803: STR Fast Approximate Kernel Methods 4

5 A. B. x 10 x10 3 Accuracy (%) Avg. Prediction Err. 6 Avg. Prediction Err. Slot Car Slot Car 8 Hilbert Space Embeddings of Hidde Mean 7 advantage the other learned models by taking of the fact that Example: Slotcar Position Estimation Last x 10 6 A. B. 8 over 250 trials) with different estimated prediction (averaged 90 LDS 3 5 HMM Mean 7 85 (LE kernel). The LE kernel finds the best representation of th IMU RR-HMM Last 6 2 4LDS 80 Embedded HMM state IMU space (Gaussian RBF kernel), and the bottom g space. The top graph shows the Kalman filter state space ( models. Red line indicates true 2-d 2 1 position of the car over racetrack (bottom).0 (B) A comparison of training data embed unit 30 (IMU) (A)6 30 Figure 3: Slot car with inertial measurement Racetrack Prediction0 Horizon Racetrack Prediction Ho Figure 4. Slot car inertial measurement data. (A) The slot racetrack Figure 5 E[f (Y ) x] = hf, µy x i car platform and 4. the Slot IMU (top) and the racetrack (botfigure car inertial measurement data man vs. Racetrack tom). (B) Squared error for prediction with different estiembedde car platform and the IMU (top) and the r 0 mated models and baselines. different to learn a model of the noisy IMU data, and, after CS8803: STR Fast Approximate Kernel Methods Non-hu time we 5 MR oitacol rac tnereffid tom). (B) Squared error for prediction with this data while the slot car circled the trackeigenmap controlled Laplacian 11 seco mated models and baselines. by a constant policy. The goal of this experiment was 40

6 Predicting via Kernel Regression E[f(Y ) x] =hf,µ Y x i E[f(Y ) x] = b f(y ), (Y ) C b 1 YYC YXC b 1 XX µ x b C b f(y ) > > > Y ( Y Y + I) 1 Y X X > X + I 1 '(x) f(y )( > Y Y + I) 1 > Y Y ( > X X + I) 1 > X'(x) f(y )(G XX + NI) 1 G XX (:,i) CS8803: STR Fast Approximate Kernel Methods 6

7 A. B. x 10 x10 3 Accuracy (%) Avg. Prediction Err. 6 Avg. Prediction Err. Slot Car Slot Car 8 Hilbert Space Embeddings of Hidde Mean 7 advantage the other learned models by taking of the fact that Example: Slotcar Position Estimation Last x 10 6 A. B. 8 over 250 trials) with different estimated prediction (averaged 90 LDS 3 5 HMM Mean 7 85 (LE kernel). The LE kernel finds the best representation of th IMU RR-HMM Last 6 2 4LDS 80 Embedded HMM state IMU space (Gaussian RBF kernel), and the bottom g space. The top graph shows the Kalman filter state space ( models. Red line indicates true 2-d 2 1 position of the car over racetrack (bottom).0 (B) A comparison of training data embed unit 30 (IMU) (A)6 30 Figure 3: Slot car with inertial measurement Racetrack Prediction0 Horizon Racetrack Prediction Ho Figure 4. Slot car inertial measurement data. (A) The slot racetrack Figure 5 E[f (Y ) x] = hf, µy x i car platform and 4. the Slot IMU (top) and the racetrack (botfigure car inertial measurement data man vs. Racetrack tom). (B) Squared error for prediction with different estiembedde car platform and the IMU (top) and the r 0 mated models and baselines. different to learn a model of the noisy IMU data, and, after CS8803: STR Fast Approximate Kernel Methods Non-hu time we 7 MR oitacol rac tnereffid tom). (B) Squared error for prediction with this data while the slot car circled the trackeigenmap controlled Laplacian 11 seco mated models and baselines. by a constant policy. The goal of this experiment was 40

8 Example: Depth Camera and Manipulator Control Input (7-DOF) Kinect Sensor (1,228,800-dimensional observations) CS8803: STR Fast Approximate Kernel Methods 8

9 Example: Depth Camera and Manipulator Control Input (7-DOF) Kinect Sensor (1,228,800-dimensional observations) The Challenge: learn a full generative model CS8803: STR Fast Approximate Kernel Methods 8

10 Example: Depth Camera and Manipulator Control Input (7-DOF) Kinect Sensor (1,228,800-dimensional observations) continuous actions and observations highly nonlinear dynamics no prior knowledge of physics, kinematics, or geometry non-gaussian noise CS8803: STR Fast Approximate Kernel Methods 8

11 Example: Depth Camera and Manipulator Control Input (7-DOF) Kinect Sensor (1,228,800-dimensional observations) Training Data: high-dimensional observations 7-DOF continuous control 15 minutes (~60 GBs) of motor babbling Test Data: test set of arm motor babbling (completely new trajectories) CS8803: STR Fast Approximate Kernel Methods 8

12 Example: Depth Camera and Manipulator Filtering (tracking): Prediction (4 seconds): True Observation E[o t a 1:t,o 1:t 1 ] E[o t+120 a 1:t+120,o 1:t 1 ] o t+120 CS8803: STR Fast Approximate Kernel Methods 9

13 CS8803: Statistical Techniques in Robotics Byron Boots Fast Approximate Kernel Methods CS8803: STR Fast Approximate Kernel Methods 10

14 Kernel Methods are Expensive y Y (K + I) 1 K(:,x) Inversion is expensive For n data points, computing the posterior mean is O(n 3 ) storing the kernel is O(n 2 ) Can use approximation to reduce the computation and storage costs CS8803: STR Fast Approximate Kernel Methods 11

15 Methods for Approximating Kernel Machines Factorize the Kernel Matrix Random Features (Random or Greedy) Subset of Data CS8803: STR Fast Approximate Kernel Methods 12

16 Factorization for Approximating Kernel Machines y Y (K + I) 1 K(:,x) y Y (K + 1 I) 1 K(:,x) Y (R XR > X + 1 I) 1 R XR(x) > YR > X(R X R > X + 2 I) 1 R(x) Matrix Inversion Lemma Result: linear regression in finite feature space CS8803: STR Fast Approximate Kernel Methods 13

17 Factorization for Approximating Kernel Machines How do we factorize K? Lots of options in linear algebra: thin SVD, incomplete Cholesky, etc. For n data points, all the decompositions are approx O(n 3 ) However: incremental / approximate versions can be much faster incremental SVD: O(nd 3 ) incomplete Cholesky: O(nd 2 ) CS8803: STR Fast Approximate Kernel Methods 14

18 Factorization for Approximating Kernel Machines Factorization-based approaches essentially approximate an infinite dimensional feature space with a small number of basis vectors Is there a simpler, faster way to choose the basis vectors? CS8803: STR Fast Approximate Kernel Methods 15

19 Methods for Approximating Kernel Machines Factorize the Kernel Matrix Random Features (Random or Greedy) Subset of Data CS8803: STR Fast Approximate Kernel Methods 16

20 Random Features (Rahimi and Recht) Random features use randomly chosen basis vector to approximate the feature space What are the basis vectors? What type of randomness to use? CS8803: STR Fast Approximate Kernel Methods 17

21 Shift Invariant Kernels Kernel value only depends on the difference between two data points!(",#) =!(" #) =!(Δ) A shift invariant kernel!(δ) is the Fourier transformation of a non-negative measure Eg. CS8803: STR Fast Approximate Kernel Methods 18

22 Random Fourier Features What basis should we use? $ %& (" #) can be replaced by cos(&(" #)) since both!(" #) and ((&) are real functions cos(&(" #)) =cos(&") cos(&#) +sin(&") sin(&#) For each &, use feature [cos(&"), sin(&") ] What type of randomness? Randomly draw & from ((&) Eg. Gaussian RBF kernel, drawn from Gaussian CS8803: STR Fast Approximate Kernel Methods 19

23 Random Fourier Features Random'features'usually'need'more'feature'dimensions'than' factoriza4on'approaches'to'achieve'the'same'approxima4on'accuracy CS8803: STR Fast Approximate Kernel Methods 20

24 Methods for Approximating Kernel Machines Factorize the Kernel Matrix Random Features (Random or Greedy) Subsets of Data CS8803: STR Fast Approximate Kernel Methods 21

25 Nystrom s Method Use sub-block of the kernel matrix to approximate the entire kernel matrix G W C G C W + CT CS8803: STR Fast Approximate Kernel Methods 22

26 Nystrom s Method Use sub-block of the kernel matrix to approximate the entire kernel matrix G CS8803: STR Fast Approximate Kernel Methods 23

27 Summary y Y (K + I) 1 K(:,x) y Y (K + 1 I) 1 K(:,x) Y (R XR > X + 1 I) 1 R XR(x) > YR > X(R X R > X + 2 I) 1 R(x) Matrix Inversion Lemma Result: linear regression in finite feature space CS8803: STR Fast Approximate Kernel Methods 24

Incremental Learning of Robot Dynamics using Random Features

Incremental Learning of Robot Dynamics using Random Features Incremental Learning of Robot Dynamics using Random Features Arjan Gijsberts, Giorgio Metta Cognitive Humanoids Laboratory Dept. of Robotics, Brain and Cognitive Sciences Italian Institute of Technology

More information

08 An Introduction to Dense Continuous Robotic Mapping

08 An Introduction to Dense Continuous Robotic Mapping NAVARCH/EECS 568, ROB 530 - Winter 2018 08 An Introduction to Dense Continuous Robotic Mapping Maani Ghaffari March 14, 2018 Previously: Occupancy Grid Maps Pose SLAM graph and its associated dense occupancy

More information

Divide and Conquer Kernel Ridge Regression

Divide and Conquer Kernel Ridge Regression Divide and Conquer Kernel Ridge Regression Yuchen Zhang John Duchi Martin Wainwright University of California, Berkeley COLT 2013 Yuchen Zhang (UC Berkeley) Divide and Conquer KRR COLT 2013 1 / 15 Problem

More information

Inverse Kinematics. Given a desired position (p) & orientation (R) of the end-effector

Inverse Kinematics. Given a desired position (p) & orientation (R) of the end-effector Inverse Kinematics Given a desired position (p) & orientation (R) of the end-effector q ( q, q, q ) 1 2 n Find the joint variables which can bring the robot the desired configuration z y x 1 The Inverse

More information

random fourier features for kernel ridge regression: approximation bounds and statistical guarantees

random fourier features for kernel ridge regression: approximation bounds and statistical guarantees random fourier features for kernel ridge regression: approximation bounds and statistical guarantees Haim Avron, Michael Kapralov, Cameron Musco, Christopher Musco, Ameya Velingker, and Amir Zandieh Tel

More information

CPSC 340: Machine Learning and Data Mining. Kernel Trick Fall 2017

CPSC 340: Machine Learning and Data Mining. Kernel Trick Fall 2017 CPSC 340: Machine Learning and Data Mining Kernel Trick Fall 2017 Admin Assignment 3: Due Friday. Midterm: Can view your exam during instructor office hours or after class this week. Digression: the other

More information

Apprenticeship Learning for Reinforcement Learning. with application to RC helicopter flight Ritwik Anand, Nick Haliday, Audrey Huang

Apprenticeship Learning for Reinforcement Learning. with application to RC helicopter flight Ritwik Anand, Nick Haliday, Audrey Huang Apprenticeship Learning for Reinforcement Learning with application to RC helicopter flight Ritwik Anand, Nick Haliday, Audrey Huang Table of Contents Introduction Theory Autonomous helicopter control

More information

Robotics. Chapter 25. Chapter 25 1

Robotics. Chapter 25. Chapter 25 1 Robotics Chapter 25 Chapter 25 1 Outline Robots, Effectors, and Sensors Localization and Mapping Motion Planning Chapter 25 2 Mobile Robots Chapter 25 3 Manipulators P R R R R R Configuration of robot

More information

Basis Functions. Volker Tresp Summer 2017

Basis Functions. Volker Tresp Summer 2017 Basis Functions Volker Tresp Summer 2017 1 Nonlinear Mappings and Nonlinear Classifiers Regression: Linearity is often a good assumption when many inputs influence the output Some natural laws are (approximately)

More information

CS8803: Statistical Techniques in Robotics Byron Boots. Thoughts on Machine Learning and Robotics. (how to apply machine learning in practice)

CS8803: Statistical Techniques in Robotics Byron Boots. Thoughts on Machine Learning and Robotics. (how to apply machine learning in practice) CS8803: Statistical Techniques in Robotics Byron Boots Thoughts on Machine Learning and Robotics (how to apply machine learning in practice) 1 CS8803: Statistical Techniques in Robotics Byron Boots Thoughts

More information

L15. POSE-GRAPH SLAM. NA568 Mobile Robotics: Methods & Algorithms

L15. POSE-GRAPH SLAM. NA568 Mobile Robotics: Methods & Algorithms L15. POSE-GRAPH SLAM NA568 Mobile Robotics: Methods & Algorithms Today s Topic Nonlinear Least Squares Pose-Graph SLAM Incremental Smoothing and Mapping Feature-Based SLAM Filtering Problem: Motion Prediction

More information

Overview. EECS 124, UC Berkeley, Spring 2008 Lecture 23: Localization and Mapping. Statistical Models

Overview. EECS 124, UC Berkeley, Spring 2008 Lecture 23: Localization and Mapping. Statistical Models Introduction ti to Embedded dsystems EECS 124, UC Berkeley, Spring 2008 Lecture 23: Localization and Mapping Gabe Hoffmann Ph.D. Candidate, Aero/Astro Engineering Stanford University Statistical Models

More information

Isometries. 1 Identifying Isometries

Isometries. 1 Identifying Isometries Isometries 1 Identifying Isometries 1. Modeling isometries as dynamic maps. 2. GeoGebra files: isoguess1.ggb, isoguess2.ggb, isoguess3.ggb, isoguess4.ggb. 3. Guessing isometries. 4. What can you construct

More information

Monte Carlo Integration

Monte Carlo Integration Lecture 11: Monte Carlo Integration Computer Graphics and Imaging UC Berkeley CS184/284A, Spring 2016 Reminder: Quadrature-Based Numerical Integration f(x) Z b a f(x)dx x 0 = a x 1 x 2 x 3 x 4 = b E.g.

More information

GTSAM 4.0 Tutorial Theory, Programming, and Applications

GTSAM 4.0 Tutorial Theory, Programming, and Applications GTSAM 4.0 Tutorial Theory, Programming, and Applications GTSAM: https://bitbucket.org/gtborg/gtsam Examples: https://github.com/dongjing3309/gtsam-examples Jing Dong 2016-11-19 License CC BY-NC-SA 3.0

More information

Probabilistic Robotics

Probabilistic Robotics Probabilistic Robotics Bayes Filter Implementations Discrete filters, Particle filters Piecewise Constant Representation of belief 2 Discrete Bayes Filter Algorithm 1. Algorithm Discrete_Bayes_filter(

More information

CVPR 2014 Visual SLAM Tutorial Efficient Inference

CVPR 2014 Visual SLAM Tutorial Efficient Inference CVPR 2014 Visual SLAM Tutorial Efficient Inference kaess@cmu.edu The Robotics Institute Carnegie Mellon University The Mapping Problem (t=0) Robot Landmark Measurement Onboard sensors: Wheel odometry Inertial

More information

Kernel Methods for Topological Data Analysis

Kernel Methods for Topological Data Analysis Kernel Methods for Topological Data Analysis Toward Topological Statistics Kenji Fukumizu The Institute of Statistical Mathematics (Tokyo Japan) Joint work with Genki Kusano and Yasuaki Hiraoka (Tohoku

More information

Joint Feature Distributions for Image Correspondence. Joint Feature Distribution Matching. Motivation

Joint Feature Distributions for Image Correspondence. Joint Feature Distribution Matching. Motivation Joint Feature Distributions for Image Correspondence We need a correspondence model based on probability, not just geometry! Bill Triggs MOVI, CNRS-INRIA, Grenoble, France http://www.inrialpes.fr/movi/people/triggs

More information

Learning Inverse Dynamics: a Comparison

Learning Inverse Dynamics: a Comparison Learning Inverse Dynamics: a Comparison Duy Nguyen-Tuong, Jan Peters, Matthias Seeger, Bernhard Schölkopf Max Planck Institute for Biological Cybernetics Spemannstraße 38, 72076 Tübingen - Germany Abstract.

More information

Robot Mapping. Least Squares Approach to SLAM. Cyrill Stachniss

Robot Mapping. Least Squares Approach to SLAM. Cyrill Stachniss Robot Mapping Least Squares Approach to SLAM Cyrill Stachniss 1 Three Main SLAM Paradigms Kalman filter Particle filter Graphbased least squares approach to SLAM 2 Least Squares in General Approach for

More information

Graphbased. Kalman filter. Particle filter. Three Main SLAM Paradigms. Robot Mapping. Least Squares Approach to SLAM. Least Squares in General

Graphbased. Kalman filter. Particle filter. Three Main SLAM Paradigms. Robot Mapping. Least Squares Approach to SLAM. Least Squares in General Robot Mapping Three Main SLAM Paradigms Least Squares Approach to SLAM Kalman filter Particle filter Graphbased Cyrill Stachniss least squares approach to SLAM 1 2 Least Squares in General! Approach for

More information

Missing Data Analysis for the Employee Dataset

Missing Data Analysis for the Employee Dataset Missing Data Analysis for the Employee Dataset 67% of the observations have missing values! Modeling Setup Random Variables: Y i =(Y i1,...,y ip ) 0 =(Y i,obs, Y i,miss ) 0 R i =(R i1,...,r ip ) 0 ( 1

More information

Virtual Vector Machine for Bayesian Online Classification

Virtual Vector Machine for Bayesian Online Classification Virtual Vector Machine for Bayesian Online Classification Thomas P. Minka, Rongjing Xiang, Yuan (Alan) Qi Appeared in UAI 2009 Presented by Lingbo Li Introduction Online Learning Update model parameters

More information

Humanoid Robotics. Monte Carlo Localization. Maren Bennewitz

Humanoid Robotics. Monte Carlo Localization. Maren Bennewitz Humanoid Robotics Monte Carlo Localization Maren Bennewitz 1 Basis Probability Rules (1) If x and y are independent: Bayes rule: Often written as: The denominator is a normalizing constant that ensures

More information

ME/CS 132: Introduction to Vision-based Robot Navigation! Low-level Image Processing" Larry Matthies"

ME/CS 132: Introduction to Vision-based Robot Navigation! Low-level Image Processing Larry Matthies ME/CS 132: Introduction to Vision-based Robot Navigation! Low-level Image Processing" Larry Matthies" lhm@jpl.nasa.gov, 818-354-3722" Announcements" First homework grading is done! Second homework is due

More information

CS 231. Inverse Kinematics Intro to Motion Capture. 3D characters. Representation. 1) Skeleton Origin (root) Joint centers/ bones lengths

CS 231. Inverse Kinematics Intro to Motion Capture. 3D characters. Representation. 1) Skeleton Origin (root) Joint centers/ bones lengths CS Inverse Kinematics Intro to Motion Capture Representation D characters ) Skeleton Origin (root) Joint centers/ bones lengths ) Keyframes Pos/Rot Root (x) Joint Angles (q) Kinematics study of static

More information

Data Driven Frequency Mapping for Computationally Scalable Object Detection

Data Driven Frequency Mapping for Computationally Scalable Object Detection 8th IEEE International Conference on Advanced Video and Signal-Based Surveillance, 2011 Data Driven Frequency Mapping for Computationally Scalable Object Detection Fatih Porikli Mitsubishi Electric Research

More information

Locally Weighted Learning for Control. Alexander Skoglund Machine Learning Course AASS, June 2005

Locally Weighted Learning for Control. Alexander Skoglund Machine Learning Course AASS, June 2005 Locally Weighted Learning for Control Alexander Skoglund Machine Learning Course AASS, June 2005 Outline Locally Weighted Learning, Christopher G. Atkeson et. al. in Artificial Intelligence Review, 11:11-73,1997

More information

Dipartimento di Elettronica Informazione e Bioingegneria Robotics

Dipartimento di Elettronica Informazione e Bioingegneria Robotics Dipartimento di Elettronica Informazione e Bioingegneria Robotics properties and performance measures @ 25 Redundancy first definition McKerrow When a manipulator can reach a specified position with more

More information

FMA901F: Machine Learning Lecture 3: Linear Models for Regression. Cristian Sminchisescu

FMA901F: Machine Learning Lecture 3: Linear Models for Regression. Cristian Sminchisescu FMA901F: Machine Learning Lecture 3: Linear Models for Regression Cristian Sminchisescu Machine Learning: Frequentist vs. Bayesian In the frequentist setting, we seek a fixed parameter (vector), with value(s)

More information

Autonomous Mobile Robot Design

Autonomous Mobile Robot Design Autonomous Mobile Robot Design Topic: EKF-based SLAM Dr. Kostas Alexis (CSE) These slides have partially relied on the course of C. Stachniss, Robot Mapping - WS 2013/14 Autonomous Robot Challenges Where

More information

( ) =cov X Y = W PRINCIPAL COMPONENT ANALYSIS. Eigenvectors of the covariance matrix are the principal components

( ) =cov X Y = W PRINCIPAL COMPONENT ANALYSIS. Eigenvectors of the covariance matrix are the principal components Review Lecture 14 ! PRINCIPAL COMPONENT ANALYSIS Eigenvectors of the covariance matrix are the principal components 1. =cov X Top K principal components are the eigenvectors with K largest eigenvalues

More information

Computer Vision I - Algorithms and Applications: Multi-View 3D reconstruction

Computer Vision I - Algorithms and Applications: Multi-View 3D reconstruction Computer Vision I - Algorithms and Applications: Multi-View 3D reconstruction Carsten Rother 09/12/2013 Computer Vision I: Multi-View 3D reconstruction Roadmap this lecture Computer Vision I: Multi-View

More information

Locality-Sensitive Codes from Shift-Invariant Kernels Maxim Raginsky (Duke) and Svetlana Lazebnik (UNC)

Locality-Sensitive Codes from Shift-Invariant Kernels Maxim Raginsky (Duke) and Svetlana Lazebnik (UNC) Locality-Sensitive Codes from Shift-Invariant Kernels Maxim Raginsky (Duke) and Svetlana Lazebnik (UNC) Goal We want to design a binary encoding of data such that similar data points (similarity measures

More information

Introduction to Computer Vision

Introduction to Computer Vision Introduction to Computer Vision Michael J. Black Nov 2009 Perspective projection and affine motion Goals Today Perspective projection 3D motion Wed Projects Friday Regularization and robust statistics

More information

Robotics (Kinematics) Winter 1393 Bonab University

Robotics (Kinematics) Winter 1393 Bonab University Robotics () Winter 1393 Bonab University : most basic study of how mechanical systems behave Introduction Need to understand the mechanical behavior for: Design Control Both: Manipulators, Mobile Robots

More information

Unit 2: Locomotion Kinematics of Wheeled Robots: Part 3

Unit 2: Locomotion Kinematics of Wheeled Robots: Part 3 Unit 2: Locomotion Kinematics of Wheeled Robots: Part 3 Computer Science 4766/6778 Department of Computer Science Memorial University of Newfoundland January 28, 2014 COMP 4766/6778 (MUN) Kinematics of

More information

Large-Scale Face Manifold Learning

Large-Scale Face Manifold Learning Large-Scale Face Manifold Learning Sanjiv Kumar Google Research New York, NY * Joint work with A. Talwalkar, H. Rowley and M. Mohri 1 Face Manifold Learning 50 x 50 pixel faces R 2500 50 x 50 pixel random

More information

CS231A Midterm Review. Friday 5/6/2016

CS231A Midterm Review. Friday 5/6/2016 CS231A Midterm Review Friday 5/6/2016 Outline General Logistics Camera Models Non-perspective cameras Calibration Single View Metrology Epipolar Geometry Structure from Motion Active Stereo and Volumetric

More information

Robotic Mapping. Outline. Introduction (Tom)

Robotic Mapping. Outline. Introduction (Tom) Outline Robotic Mapping 6.834 Student Lecture Itamar Kahn, Thomas Lin, Yuval Mazor Introduction (Tom) Kalman Filtering (Itamar) J.J. Leonard and H.J.S. Feder. A computationally efficient method for large-scale

More information

CS 231. Inverse Kinematics Intro to Motion Capture

CS 231. Inverse Kinematics Intro to Motion Capture CS 231 Inverse Kinematics Intro to Motion Capture Representation 1) Skeleton Origin (root) Joint centers/ bones lengths 2) Keyframes Pos/Rot Root (x) Joint Angles (q) 3D characters Kinematics study of

More information

Data-driven modeling: A low-rank approximation problem

Data-driven modeling: A low-rank approximation problem 1 / 31 Data-driven modeling: A low-rank approximation problem Ivan Markovsky Vrije Universiteit Brussel 2 / 31 Outline Setup: data-driven modeling Problems: system identification, machine learning,...

More information

From Personal Computers to Personal Robots

From Personal Computers to Personal Robots From Personal Computers to Personal Robots Challenges in Computer Science Education Nikolaus Correll Department of Computer Science University of Colorado at Boulder Mechanism vs. Computer Unimate (1961)

More information

Learning to Track Motion

Learning to Track Motion Learning to Track Motion Maitreyi Nanjanath Amit Bose CSci 8980 Course Project April 25, 2006 Background Vision sensor can provide a great deal of information in a short sequence of images Useful for determining

More information

Computer Vision II Lecture 14

Computer Vision II Lecture 14 Computer Vision II Lecture 14 Articulated Tracking I 08.07.2014 Bastian Leibe RWTH Aachen http://www.vision.rwth-aachen.de leibe@vision.rwth-aachen.de Outline of This Lecture Single-Object Tracking Bayesian

More information

Nonlinear State Estimation for Robotics and Computer Vision Applications: An Overview

Nonlinear State Estimation for Robotics and Computer Vision Applications: An Overview Nonlinear State Estimation for Robotics and Computer Vision Applications: An Overview Arun Das 05/09/2017 Arun Das Waterloo Autonomous Vehicles Lab Introduction What s in a name? Arun Das Waterloo Autonomous

More information

Robot Inverse Kinematics Asanga Ratnaweera Department of Mechanical Engieering

Robot Inverse Kinematics Asanga Ratnaweera Department of Mechanical Engieering PR 5 Robot Dynamics & Control /8/7 PR 5: Robot Dynamics & Control Robot Inverse Kinematics Asanga Ratnaweera Department of Mechanical Engieering The Inverse Kinematics The determination of all possible

More information

Kinect Cursor Control EEE178 Dr. Fethi Belkhouche Christopher Harris Danny Nguyen I. INTRODUCTION

Kinect Cursor Control EEE178 Dr. Fethi Belkhouche Christopher Harris Danny Nguyen I. INTRODUCTION Kinect Cursor Control EEE178 Dr. Fethi Belkhouche Christopher Harris Danny Nguyen Abstract: An XBOX 360 Kinect is used to develop two applications to control the desktop cursor of a Windows computer. Application

More information

RANDOM FEATURES FOR KERNEL DEEP CONVEX NETWORK

RANDOM FEATURES FOR KERNEL DEEP CONVEX NETWORK RANDOM FEATURES FOR KERNEL DEEP CONVEX NETWORK Po-Sen Huang, Li Deng, Mark Hasegawa-Johnson, Xiaodong He Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, USA

More information

CAMERA POSE ESTIMATION OF RGB-D SENSORS USING PARTICLE FILTERING

CAMERA POSE ESTIMATION OF RGB-D SENSORS USING PARTICLE FILTERING CAMERA POSE ESTIMATION OF RGB-D SENSORS USING PARTICLE FILTERING By Michael Lowney Senior Thesis in Electrical Engineering University of Illinois at Urbana-Champaign Advisor: Professor Minh Do May 2015

More information

What is Learning? CS 343: Artificial Intelligence Machine Learning. Raymond J. Mooney. Problem Solving / Planning / Control.

What is Learning? CS 343: Artificial Intelligence Machine Learning. Raymond J. Mooney. Problem Solving / Planning / Control. What is Learning? CS 343: Artificial Intelligence Machine Learning Herbert Simon: Learning is any process by which a system improves performance from experience. What is the task? Classification Problem

More information

In Homework 1, you determined the inverse dynamics model of the spinbot robot to be

In Homework 1, you determined the inverse dynamics model of the spinbot robot to be Robot Learning Winter Semester 22/3, Homework 2 Prof. Dr. J. Peters, M.Eng. O. Kroemer, M. Sc. H. van Hoof Due date: Wed 6 Jan. 23 Note: Please fill in the solution on this sheet but add sheets for the

More information

CSE 417T: Introduction to Machine Learning. Lecture 22: The Kernel Trick. Henry Chai 11/15/18

CSE 417T: Introduction to Machine Learning. Lecture 22: The Kernel Trick. Henry Chai 11/15/18 CSE 417T: Introduction to Machine Learning Lecture 22: The Kernel Trick Henry Chai 11/15/18 Linearly Inseparable Data What can we do if the data is not linearly separable? Accept some non-zero in-sample

More information

Humanoid Robotics. Least Squares. Maren Bennewitz

Humanoid Robotics. Least Squares. Maren Bennewitz Humanoid Robotics Least Squares Maren Bennewitz Goal of This Lecture Introduction into least squares Use it yourself for odometry calibration, later in the lecture: camera and whole-body self-calibration

More information

Jane Li. Assistant Professor Mechanical Engineering Department, Robotic Engineering Program Worcester Polytechnic Institute

Jane Li. Assistant Professor Mechanical Engineering Department, Robotic Engineering Program Worcester Polytechnic Institute Jane Li Assistant Professor Mechanical Engineering Department, Robotic Engineering Program Worcester Polytechnic Institute (3 pts) Compare the testing methods for testing path segment and finding first

More information

Basis Functions. Volker Tresp Summer 2016

Basis Functions. Volker Tresp Summer 2016 Basis Functions Volker Tresp Summer 2016 1 I am an AI optimist. We ve got a lot of work in machine learning, which is sort of the polite term for AI nowadays because it got so broad that it s not that

More information

A novel approach to motion tracking with wearable sensors based on Probabilistic Graphical Models

A novel approach to motion tracking with wearable sensors based on Probabilistic Graphical Models A novel approach to motion tracking with wearable sensors based on Probabilistic Graphical Models Emanuele Ruffaldi Lorenzo Peppoloni Alessandro Filippeschi Carlo Alberto Avizzano 2014 IEEE International

More information

Final Exam Study Guide

Final Exam Study Guide Final Exam Study Guide Exam Window: 28th April, 12:00am EST to 30th April, 11:59pm EST Description As indicated in class the goal of the exam is to encourage you to review the material from the course.

More information

Motion Models (cont) 1 3/15/2018

Motion Models (cont) 1 3/15/2018 Motion Models (cont) 1 3/15/018 Computing the Density to compute,, and use the appropriate probability density function; i.e., for zeromean Gaussian noise: 3/15/018 Sampling from the Velocity Motion Model

More information

Grasping Known Objects with Aldebaran Nao

Grasping Known Objects with Aldebaran Nao CS365 Project Report Grasping Known Objects with Aldebaran Nao By: Ashu Gupta( ashug@iitk.ac.in) Mohd. Dawood( mdawood@iitk.ac.in) Department of Computer Science and Engineering IIT Kanpur Mentor: Prof.

More information

Articulated Pose Estimation with Flexible Mixtures-of-Parts

Articulated Pose Estimation with Flexible Mixtures-of-Parts Articulated Pose Estimation with Flexible Mixtures-of-Parts PRESENTATION: JESSE DAVIS CS 3710 VISUAL RECOGNITION Outline Modeling Special Cases Inferences Learning Experiments Problem and Relevance Problem:

More information

Filtering, scale, orientation, localization, and texture. Nuno Vasconcelos ECE Department, UCSD (with thanks to David Forsyth)

Filtering, scale, orientation, localization, and texture. Nuno Vasconcelos ECE Department, UCSD (with thanks to David Forsyth) Filtering, scale, orientation, localization, and texture Nuno Vasconcelos ECE Department, UCSD (with thanks to David Forsyth) Beyond edges we have talked a lot about edges while they are important, it

More information

Edge and corner detection

Edge and corner detection Edge and corner detection Prof. Stricker Doz. G. Bleser Computer Vision: Object and People Tracking Goals Where is the information in an image? How is an object characterized? How can I find measurements

More information

Geometric camera models and calibration

Geometric camera models and calibration Geometric camera models and calibration http://graphics.cs.cmu.edu/courses/15-463 15-463, 15-663, 15-862 Computational Photography Fall 2018, Lecture 13 Course announcements Homework 3 is out. - Due October

More information

COMP30019 Graphics and Interaction Kinematics

COMP30019 Graphics and Interaction Kinematics COMP30019 Graphics and Interaction Kinematics Department of Computing and Information Systems The Lecture outline Introduction Forward kinematics Inverse kinematics Kinematics I am robot (am I?) Forward

More information

Announcements. CS 188: Artificial Intelligence Spring Advanced Applications. Robot folds towels. Robotic Control Tasks

Announcements. CS 188: Artificial Intelligence Spring Advanced Applications. Robot folds towels. Robotic Control Tasks CS 188: Artificial Intelligence Spring 2011 Advanced Applications: Robotics Announcements Practice Final Out (optional) Similar extra credit system as practice midterm Contest (optional): Tomorrow night

More information

CS 188: Artificial Intelligence Spring Announcements

CS 188: Artificial Intelligence Spring Announcements CS 188: Artificial Intelligence Spring 2011 Advanced Applications: Robotics Pieter Abbeel UC Berkeley A few slides from Sebastian Thrun, Dan Klein 1 Announcements Practice Final Out (optional) Similar

More information

3D Human Motion Analysis and Manifolds

3D Human Motion Analysis and Manifolds D E P A R T M E N T O F C O M P U T E R S C I E N C E U N I V E R S I T Y O F C O P E N H A G E N 3D Human Motion Analysis and Manifolds Kim Steenstrup Pedersen DIKU Image group and E-Science center Motivation

More information

Knowledge Discovery and Data Mining

Knowledge Discovery and Data Mining Knowledge Discovery and Data Mining Basis Functions Tom Kelsey School of Computer Science University of St Andrews http://www.cs.st-andrews.ac.uk/~tom/ tom@cs.st-andrews.ac.uk Tom Kelsey ID5059-02-BF 2015-02-04

More information

Stereo and Epipolar geometry

Stereo and Epipolar geometry Previously Image Primitives (feature points, lines, contours) Today: Stereo and Epipolar geometry How to match primitives between two (multiple) views) Goals: 3D reconstruction, recognition Jana Kosecka

More information

Announcements. CS 188: Artificial Intelligence Fall Robot motion planning! Today. Robotics Tasks. Mobile Robots

Announcements. CS 188: Artificial Intelligence Fall Robot motion planning! Today. Robotics Tasks. Mobile Robots CS 188: Artificial Intelligence Fall 2007 Lecture 6: Robot Motion Planning 9/13/2007 Announcements Project 1 due (yesterday)! Project 2 (Pacman with ghosts) up in a few days Reminder: you are allowed to

More information

CS 188: Artificial Intelligence Fall Announcements

CS 188: Artificial Intelligence Fall Announcements CS 188: Artificial Intelligence Fall 2007 Lecture 6: Robot Motion Planning 9/13/2007 Dan Klein UC Berkeley Many slides over the course adapted from either Stuart Russell or Andrew Moore Announcements Project

More information

Statistics of Natural Image Categories

Statistics of Natural Image Categories Statistics of Natural Image Categories Authors: Antonio Torralba and Aude Oliva Presented by: Sebastian Scherer Experiment Please estimate the average depth from the camera viewpoint to all locations(pixels)

More information

Visual Pathways to the Brain

Visual Pathways to the Brain Visual Pathways to the Brain 1 Left half of visual field which is imaged on the right half of each retina is transmitted to right half of brain. Vice versa for right half of visual field. From each eye

More information

The Efficient Extension of Globally Consistent Scan Matching to 6 DoF

The Efficient Extension of Globally Consistent Scan Matching to 6 DoF The Efficient Extension of Globally Consistent Scan Matching to 6 DoF Dorit Borrmann, Jan Elseberg, Kai Lingemann, Andreas Nüchter, Joachim Hertzberg 1 / 20 Outline 1 Introduction 2 Algorithm 3 Performance

More information

Machine Learning Feature Creation and Selection

Machine Learning Feature Creation and Selection Machine Learning Feature Creation and Selection Jeff Howbert Introduction to Machine Learning Winter 2012 1 Feature creation Well-conceived new features can sometimes capture the important information

More information

10-701/15-781, Fall 2006, Final

10-701/15-781, Fall 2006, Final -7/-78, Fall 6, Final Dec, :pm-8:pm There are 9 questions in this exam ( pages including this cover sheet). If you need more room to work out your answer to a question, use the back of the page and clearly

More information

Filtering Images. Contents

Filtering Images. Contents Image Processing and Data Visualization with MATLAB Filtering Images Hansrudi Noser June 8-9, 010 UZH, Multimedia and Robotics Summer School Noise Smoothing Filters Sigmoid Filters Gradient Filters Contents

More information

A Parallel, In-Place, Rectangular Matrix Transpose Algorithm

A Parallel, In-Place, Rectangular Matrix Transpose Algorithm Stefan Amberger ICA & RISC amberger.stefan@gmail.com A Parallel, In-Place, Rectangular Matrix Transpose Algorithm Description of Algorithm and Correctness Proof Table of Contents 1. Introduction 2. Description

More information

Transductive and Inductive Methods for Approximate Gaussian Process Regression

Transductive and Inductive Methods for Approximate Gaussian Process Regression Transductive and Inductive Methods for Approximate Gaussian Process Regression Anton Schwaighofer 1,2 1 TU Graz, Institute for Theoretical Computer Science Inffeldgasse 16b, 8010 Graz, Austria http://www.igi.tugraz.at/aschwaig

More information

Geometric Transformations and Image Warping

Geometric Transformations and Image Warping Geometric Transformations and Image Warping Ross Whitaker SCI Institute, School of Computing University of Utah Univ of Utah, CS6640 2009 1 Geometric Transformations Greyscale transformations -> operate

More information

A New Algorithm for Measuring and Optimizing the Manipulability Index

A New Algorithm for Measuring and Optimizing the Manipulability Index DOI 10.1007/s10846-009-9388-9 A New Algorithm for Measuring and Optimizing the Manipulability Index Ayssam Yehia Elkady Mohammed Mohammed Tarek Sobh Received: 16 September 2009 / Accepted: 27 October 2009

More information

Advanced Topics in Digital Communications Spezielle Methoden der digitalen Datenübertragung

Advanced Topics in Digital Communications Spezielle Methoden der digitalen Datenübertragung Advanced Topics in Digital Communications Spezielle Methoden der digitalen Datenübertragung Dr.-Ing. Carsten Bockelmann Institute for Telecommunications and High-Frequency Techniques Department of Communications

More information

ME 597/747 Autonomous Mobile Robots. Mid Term Exam. Duration: 2 hour Total Marks: 100

ME 597/747 Autonomous Mobile Robots. Mid Term Exam. Duration: 2 hour Total Marks: 100 ME 597/747 Autonomous Mobile Robots Mid Term Exam Duration: 2 hour Total Marks: 100 Instructions: Read the exam carefully before starting. Equations are at the back, but they are NOT necessarily valid

More information

Direct Methods in Visual Odometry

Direct Methods in Visual Odometry Direct Methods in Visual Odometry July 24, 2017 Direct Methods in Visual Odometry July 24, 2017 1 / 47 Motivation for using Visual Odometry Wheel odometry is affected by wheel slip More accurate compared

More information

Local Search Methods. CS 188: Artificial Intelligence Fall Announcements. Hill Climbing. Hill Climbing Diagram. Today

Local Search Methods. CS 188: Artificial Intelligence Fall Announcements. Hill Climbing. Hill Climbing Diagram. Today CS 188: Artificial Intelligence Fall 2006 Lecture 5: Robot Motion Planning 9/14/2006 Local Search Methods Queue-based algorithms keep fallback options (backtracking) Local search: improve what you have

More information

Multiple View Geometry in Computer Vision

Multiple View Geometry in Computer Vision Multiple View Geometry in Computer Vision Prasanna Sahoo Department of Mathematics University of Louisville 1 Structure Computation Lecture 18 March 22, 2005 2 3D Reconstruction The goal of 3D reconstruction

More information

CLASSIFICATION AND CHANGE DETECTION

CLASSIFICATION AND CHANGE DETECTION IMAGE ANALYSIS, CLASSIFICATION AND CHANGE DETECTION IN REMOTE SENSING With Algorithms for ENVI/IDL and Python THIRD EDITION Morton J. Canty CRC Press Taylor & Francis Group Boca Raton London NewYork CRC

More information

CS545 Contents IX. Inverse Kinematics. Reading Assignment for Next Class. Analytical Methods Iterative (Differential) Methods

CS545 Contents IX. Inverse Kinematics. Reading Assignment for Next Class. Analytical Methods Iterative (Differential) Methods CS545 Contents IX Inverse Kinematics Analytical Methods Iterative (Differential) Methods Geometric and Analytical Jacobian Jacobian Transpose Method Pseudo-Inverse Pseudo-Inverse with Optimization Extended

More information

Autonomous Navigation in Complex Indoor and Outdoor Environments with Micro Aerial Vehicles

Autonomous Navigation in Complex Indoor and Outdoor Environments with Micro Aerial Vehicles Autonomous Navigation in Complex Indoor and Outdoor Environments with Micro Aerial Vehicles Shaojie Shen Dept. of Electrical and Systems Engineering & GRASP Lab, University of Pennsylvania Committee: Daniel

More information

Data-driven modeling: A low-rank approximation problem

Data-driven modeling: A low-rank approximation problem 1 / 34 Data-driven modeling: A low-rank approximation problem Ivan Markovsky Vrije Universiteit Brussel 2 / 34 Outline Setup: data-driven modeling Problems: system identification, machine learning,...

More information

Structure from Motion

Structure from Motion 11/18/11 Structure from Motion Computer Vision CS 143, Brown James Hays Many slides adapted from Derek Hoiem, Lana Lazebnik, Silvio Saverese, Steve Seitz, and Martial Hebert This class: structure from

More information

Image Analysis, Classification and Change Detection in Remote Sensing

Image Analysis, Classification and Change Detection in Remote Sensing Image Analysis, Classification and Change Detection in Remote Sensing WITH ALGORITHMS FOR ENVI/IDL Morton J. Canty Taylor &. Francis Taylor & Francis Group Boca Raton London New York CRC is an imprint

More information

LS-SVM Functional Network for Time Series Prediction

LS-SVM Functional Network for Time Series Prediction LS-SVM Functional Network for Time Series Prediction Tuomas Kärnä 1, Fabrice Rossi 2 and Amaury Lendasse 1 Helsinki University of Technology - Neural Networks Research Center P.O. Box 5400, FI-02015 -

More information

Support Vector Machines.

Support Vector Machines. Support Vector Machines srihari@buffalo.edu SVM Discussion Overview 1. Overview of SVMs 2. Margin Geometry 3. SVM Optimization 4. Overlapping Distributions 5. Relationship to Logistic Regression 6. Dealing

More information

Basic Elements. Geometry is the study of the relationships among objects in an n-dimensional space

Basic Elements. Geometry is the study of the relationships among objects in an n-dimensional space Basic Elements Geometry is the study of the relationships among objects in an n-dimensional space In computer graphics, we are interested in objects that exist in three dimensions We want a minimum set

More information

Schedule for Rest of Semester

Schedule for Rest of Semester Schedule for Rest of Semester Date Lecture Topic 11/20 24 Texture 11/27 25 Review of Statistics & Linear Algebra, Eigenvectors 11/29 26 Eigenvector expansions, Pattern Recognition 12/4 27 Cameras & calibration

More information

Gas Distribution Modeling Using Sparse Gaussian Process Mixture Models

Gas Distribution Modeling Using Sparse Gaussian Process Mixture Models Gas Distribution Modeling Using Sparse Gaussian Process Mixture Models Cyrill Stachniss, Christian Plagemann, Achim Lilienthal, Wolfram Burgard University of Freiburg, Germany & Örebro University, Sweden

More information

Statistical image models

Statistical image models Chapter 4 Statistical image models 4. Introduction 4.. Visual worlds Figure 4. shows images that belong to different visual worlds. The first world (fig. 4..a) is the world of white noise. It is the world

More information