Interactive Learning of the Acoustic Properties of Objects by a Robot

Slides:



Advertisements
Similar presentations
Pseudo-Relevance Feedback For Multimedia Retrieval By Rong Yan, Alexander G. and Rong Jin Mwangi S. Kariuki
Advertisements

ECG Signal processing (2)
Image classification Given the bag-of-features representations of images from different classes, how do we learn a model for distinguishing them?
Data Mining Classification: Alternative Techniques
Pattern Recognition and Machine Learning
An Introduction of Support Vector Machine
Classification using intersection kernel SVMs is efficient Joint work with Subhransu Maji and Alex Berg Jitendra Malik UC Berkeley.
Support Vector Machines
SVM—Support Vector Machines
Machine learning continued Image source:
An Overview of Machine Learning
Patch to the Future: Unsupervised Visual Prediction
Studying Relationships between Human Posture and Health Risk Factors during Sedentary Activities Tejas Srinivasan Mentors: Vladimir Pavlovic Saehoon Yi.
Classification and Decision Boundaries
Discriminative and generative methods for bags of features
Pattern Recognition and Machine Learning
Image classification Given the bag-of-features representations of images from different classes, how do we learn a model for distinguishing them?
Support Vector Machines (SVMs) Chapter 5 (Duda et al.)
1 Pattern Recognition Pattern recognition is: 1. The name of the journal of the Pattern Recognition Society. 2. A research area in which patterns in data.
1 Classification: Definition Given a collection of records (training set ) Each record contains a set of attributes, one of the attributes is the class.
1 Pattern Recognition Pattern recognition is: 1. The name of the journal of the Pattern Recognition Society. 2. A research area in which patterns in data.
Prénom Nom Document Analysis: Data Analysis and Clustering Prof. Rolf Ingold, University of Fribourg Master course, spring semester 2008.
Support Vector Machines
Modeling Gene Interactions in Disease CS 686 Bioinformatics.
Pattern Recognition. Introduction. Definitions.. Recognition process. Recognition process relates input signal to the stored concepts about the object.
Face Processing System Presented by: Harvest Jang Group meeting Fall 2002.
05/06/2005CSIS © M. Gibbons On Evaluating Open Biometric Identification Systems Spring 2005 Michael Gibbons School of Computer Science & Information Systems.
Review Rong Jin. Comparison of Different Classification Models  The goal of all classifiers Predicating class label y for an input x Estimate p(y|x)
Discriminative and generative methods for bags of features
Extracting Places and Activities from GPS Traces Using Hierarchical Conditional Random Fields Yong-Joong Kim Dept. of Computer Science Yonsei.
Step 3: Classification Learn a decision rule (classifier) assigning bag-of-features representations of images to different classes Decision boundary Zebra.
This week: overview on pattern recognition (related to machine learning)
Classification 2: discriminative models
DATA MINING LECTURE 10 Classification k-nearest neighbor classifier Naïve Bayes Logistic Regression Support Vector Machines.
CSE 473/573 Computer Vision and Image Processing (CVIP) Ifeoma Nwogu Lecture 24 – Classifiers 1.
1 SUPPORT VECTOR MACHINES İsmail GÜNEŞ. 2 What is SVM? A new generation learning system. A new generation learning system. Based on recent advances in.
Multimodal Information Analysis for Emotion Recognition
Kernel Methods A B M Shawkat Ali 1 2 Data Mining ¤ DM or KDD (Knowledge Discovery in Databases) Extracting previously unknown, valid, and actionable.
Classifiers Given a feature representation for images, how do we learn a model for distinguishing features from different classes? Zebra Non-zebra Decision.
1 Pattern Recognition Pattern recognition is: 1. A research area in which patterns in data are found, recognized, discovered, …whatever. 2. A catchall.
Pattern Recognition 1 Pattern recognition is: 1. The name of the journal of the Pattern Recognition Society. 2. A research area in which patterns in data.
A Passive Approach to Sensor Network Localization Rahul Biswas and Sebastian Thrun International Conference on Intelligent Robots and Systems 2004 Presented.
Handwritten digit recognition
Christopher M. Bishop, Pattern Recognition and Machine Learning.
Face Detection Using Large Margin Classifiers Ming-Hsuan Yang Dan Roth Narendra Ahuja Presented by Kiang “Sean” Zhou Beckman Institute University of Illinois.
Visual Categorization With Bags of Keypoints Original Authors: G. Csurka, C.R. Dance, L. Fan, J. Willamowski, C. Bray ECCV Workshop on Statistical Learning.
Gang WangDerek HoiemDavid Forsyth. INTRODUCTION APROACH (implement detail) EXPERIMENTS CONCLUSION.
1Ellen L. Walker Category Recognition Associating information extracted from images with categories (classes) of objects Requires prior knowledge about.
CS 1699: Intro to Computer Vision Support Vector Machines Prof. Adriana Kovashka University of Pittsburgh October 29, 2015.
Final Exam Review CS479/679 Pattern Recognition Dr. George Bebis 1.
CS Statistical Machine learning Lecture 12 Yuan (Alan) Qi Purdue CS Oct
DATA MINING LECTURE 10b Classification k-nearest neighbor classifier
A Kernel Approach for Learning From Almost Orthogonal Pattern * CIS 525 Class Presentation Professor: Slobodan Vucetic Presenter: Yilian Qin * B. Scholkopf.
SVMs in a Nutshell.
Jivko Sinapov, Kaijen Hsiao and Radu Bogdan Rusu Proprioceptive Perception for Object Weight Classification.
Finding Clusters within a Class to Improve Classification Accuracy Literature Survey Yong Jae Lee 3/6/08.
SUPERVISED AND UNSUPERVISED LEARNING Presentation by Ege Saygıner CENG 784.
Shape2Pose: Human Centric Shape Analysis CMPT888 Vladimir G. Kim Siddhartha Chaudhuri Leonidas Guibas Thomas Funkhouser Stanford University Princeton University.
V k equals the vector difference between the object and the block across the first and last frames in the image sequence or more formally: Toward Learning.
Non-separable SVM's, and non-linear classification using kernels Jakob Verbeek December 16, 2011 Course website:
San Diego May 22, 2013 Giovanni Saponaro Giampiero Salvi
Introduction to Machine Learning
School of Computer Science & Engineering
Reflectance Function Approximation
Recognizing Deformable Shapes
Pattern Recognition CS479/679 Pattern Recognition Dr. George Bebis
Interactive Object Recognition Using Proprioceptive Feedback
Outline Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proceedings of the IEEE, vol. 86, no.
Machine Learning with Clinical Data
Presentation transcript:

Interactive Learning of the Acoustic Properties of Objects by a Robot Jivko Sinapov Mark Wiemer Alexander Stoytchev {jsinapov|banff|alexs}@iastate.edu Iowa State University

Motivation: why study sound? Sound Producing Event 1. Conscious - mention Gaver: “In particular, such an examination suggests that a given sound provides information about an interaction of materials at a location in an environment.” Human beings and many animals have the remarkable ability to infer and detect events in the world using acoustic information. This example, adopted from Gaver, illustrates how sound waves allow us to experience events that are beyond the reach of range of the rest of our senses. Based on our experience in life, we can easily tell that a car is approaching just from the sound we hear. Gaver in particular argues that when we hear a non-speech sound, our brain tries to figure out the physical event that caused that generated that sound. [Gaver, 1993]

Motivation (2) Why should a robot use acoustic information? Human environments are cluttered with objects that generate sounds Help robot perceive events and objects outside of field of view Help robot perceive material properties of objects

Related Work Krotkov et al. (1996) and Klatzky et al. (2000): Perception of material using contact sounds. Learned sound models for tapping aluminum, brass, glass, wood, and plastic (one object per material) Richmond and Pai (2000) Robotic platform for measuring contact sounds between robot’s end effector and object surfaces Models the contact sounds from different materials using spectrogram averaging [Richmond and Pai, 200]

Related Work (2) Torres-Jara, Natale and Fitzpatrick (2005) Robot taps objects and records spectrogram of sound Recognize objects using spectrogram matching Recognized 4 test objects used during training. Tapping objects Spectrogram of tapping

Our Study Demonstrate object recognition using acoustic features from interaction 18 Different Objects 3 Different behaviors: push, grasp, drop Evaluate different machine learning algorithms

Robot and Objects 7-DOF Barret WAM arm with Barret Hand 18 Different objects: Get rid of figure text, paste list of objects in ppt

Robot Behaviors Three behaviors: grasp, push, drop Grasping: Talk to kevin Segment it better, skip first few seconds

Robot Behaviors Three behaviors: grasp, push, drop Pushing:

Robot Behaviors Three behaviors: grasp, push, drop Dropping:

Sound Feature Representation Step 1: segment sound wave during interaction: Step 2: Compute Discrete Fourier transform (DFT) of sound wave: Step 3: Compute 2-D histogram of DFT matrix using block averaging: Put labels for time and frequency on dft picture 5 frequency bins Frequency Time 10 temporal bins

Object Recognition using Acoustic Properties of Objects Problem: given robot’s behavior and detected sound features from interaction, predict the object. Example: Behavior: Sound Features: Object Class: grasp

Problem Formulation Let be the set of exploratory behaviors Let be the set of objects, Let be a data point such that: , , and For each behavior learn a model that can estimate

Learning Algorithms K-NN Support Vector Machine (SVM) Bayesian Network Simple instance-based algorithm Uses Euclidean distance function Support Vector Machine (SVM) Discriminative approach, uses Kernel trick Bayesian Network Probabilistic graphical model Sound Features are discretized into bins Add picture for each algorithm

Learning Algorithms: k-NN, SVM, and Bayesian Network k-NN: memory-based learning algorithm With k = 3: 2 neighbors 1 neighbors Test point ? Therefore, Pr(red) = 0.66 Pr(blue) = 0.33

Learning Algorithms: k-NN, SVM, and Bayesian Network Support Vector Machine: discriminative learning algorithm Finds maximum margin hyperplane that separates two classes Uses Kernel trick to map data points into a feature space in which such a hyperplane exists [http://www.imtech.res.in/raghava/rbpred/svm.jpg]

Learning Algorithms: k-NN, SVM, and Bayesian Network Bayesian Network: a probabilistic graphical model Full power of statistical modeling and inference Learning: learns both the structure of the network and the parameters (conditional probability tables) Numerical features are discretized into bins A B C D E

Using Multiple Behaviors Given trained models , , Given novel sounds , , from behaviors performed on the same object Assign prediction to object class that maximizes: Fix cropping from paper

Evaluation 6 trials recorded with each of the 18 objects with each of the 3 behaviors Leave-one-out cross-validation Compared performance of learning algorithms as well as behaviors Performance Measure:

Results Chance accuracy = 1/18 = 5.6667%

Confusion Matrix for model Mpush using Bayesian Network Predicted → 4 - 2 5 1 6 3 Perfect classification and no false positives for: Add pictures of rest of the objects

Confusion Matrix for model Mcombined using Bayesian Network Predicted → 6 - 1 5 Conclusion: The errors made by models Mgrasp, Mpush and Mdrop are uncorrelated.

Learning rate of algorithms Compare performance of the model Mgrasp as a function of dataset size for: k-NN Support Vector Machine Bayesian Network Get rid of figure caption, Add next figure after end of talk in case

Learning Rate per Behavior with Bayesian Network

Summary and Conclusions Accurate acoustic-based object recognition with 18 objects and 3 behaviors Using multiple behaviors improves recognition regardless of learning algorithm Bayesian network performed best with given feature representation Grasping and Pushing interaction produces sound features that are more informative of the object than Dropping “improves recognition regardless of the type of learning algorithm being used”

Future Work Scaling up: Increase number of objects Vary object and robot pose Autonomous interaction Use unsupervised learning to form object sound categories More powerful feature representations Temporal features (i.e. periodicity) of sounds Use models to detect events in the world performed by others (humans or other robots) Change to arial