Présentation PowerPoint - INSA Rouen

Pr Sentation PowerPoint INSA Rouen-ppt Download

  • Date:03 Jul 2020
  • Views:81
  • Downloads:0
  • Size:5.66 MB

Share Presentation : Pr Sentation PowerPoint INSA Rouen

Download and Preview : Pr Sentation PowerPoint INSA Rouen

Report CopyRight/DMCA Form For : Pr Sentation PowerPoint INSA Rouen


Transcription:

S minaire PSIFRE CNRS 254616 Janvier 2003scanu insa rouen frIntelligent sensor.
and learning challengesfor context aware appliances St phane Canuscanu insa rouen frasi insa rouen fr scanu.
INSA Rouen France EULaboratoire PSI S minaire PSIFRE CNRS 25461984 La souris et leMacintoch.
16 Janvier 2003scanu insa rouen fr200X la nouvelle rupture break through S minaire PSIFRE CNRS 2546.
La technologie d aujourd hui16 Janvier 2003scanu insa rouen fr Loi de Moore Communication sans fil .
L re des donn esQuelles applications S minaire PSIFRE CNRS 254616 Janvier 2003.
scanu insa rouen fr S minaire PSIFRE CNRS 254616 Janvier 2003scanu insa rouen fr.
Olympus Optical Co Ltd is pleased to announce its newwearable user interface technologies Employing gestures and other hand movements for input the systemis an ideal match for new wearable PCs S minaire PSI.
FRE CNRS 254616 Janvier 2003scanu insa rouen frhttp www redwoodhouse com wear... http wearables cs bris ac uk p... .
http www ices cmu edu design s... S minaire PSIFRE CNRS 2546Reasearch on wearable16 Janvier 2003.
scanu insa rouen fr S minaire PSIFRE CNRS 254616 Janvier 2003scanu insa rouen fr.
S minaire PSIFRE CNRS 2546context aware appliances16 Janvier 2003scanu insa rouen fr.
The mediacup calm version of the active badge Phone by nighthttp mediacup teco edu overvie... S minaire PSI.
FRE CNRS 2546General Motors and CMU16 Janvier 2003scanu insa rouen fr drives together.
informs you in a parking GM CMU Companion driver interface system S minaire PSIFRE CNRS 2546.
Oops Where is my car 16 Janvier 2003scanu insa rouen fr Old fashion software design process1 Match the sentence.
2 Send the query to the satellite3 Satellite send query to the car on its own frequency4 Car answers Tell the computer what to do where is the switch Distributed software design interaction.
Software agents talk together Future way Programming by Example Show the computer what to do Today s solution Louis my 3 years old sonDisappearing computer.
Your Wish is My Command Programming by Example Henry Lieberman editor Published by Morgan Kaufmann 2001 S minaire PSIFRE CNRS 2546Calm technology16 Janvier 2003.
scanu insa rouen fr Ubiquitous computing One people many computer Technology at our service Reactive to what user do.
Proactive Prepare what to do next Situated sharing context Hans Gellersen Sensing in Ubiquitous Computing Adapted to our needs New functionalities and new behaviors.
New way of communicating Learn to adaptMachines have to know their context M Weiser The Computer for the 21st Century Scientific American September 1991 S minaire PSI.
FRE CNRS 2546What is the context 16 Janvier 2003scanu insa rouen fr user Context input.
activity available meeting Explicit location sensorsContext awareapplication identity profile.
environment monitoring Context output time day night temperature weather Adapted From Henry Lieberman and Ted Selker Out of Context Computer Systems That Adapt To and Learn From Context resources networks services IBM Systems Journal 39 2000 appliance proprioception.
usage functionalities maintenance resources energy history Abstract representation of the situationKnowledge .
How to find it from data S minaire PSIFRE CNRS 254616 Janvier 2003Sensing context from the environment scanu insa rouen fr.
presentation roadmap2 Representation3 Information retrieval4 Context evolution5 User interaction.
Kristof Van Laerhoven Kofi Aidoo Teaching Context to ApplicationsIn Personal and Ubiquitous Computing Volume 5 Issue 1 2001 pp 46 49 S minaire PSIFRE CNRS 2546Context from data.
16 Janvier 2003scanu insa rouen frRepresentationInformation retrieval Unbelievable capacity Context evolution.
Moore s law New sensors Artificial nose Bio sensor Personal data.
humor affective computinghttp www stat stanford edu don... S minaire PSIFRE CNRS 2546Biological sensors.
16 Janvier 2003scanu insa rouen frRepresentationInformation retrievalContext evolution.
How are you http www teco edu tea sensors ... S minaire PSIFRE CNRS 2546Expression recognition.
16 Janvier 2003scanu insa rouen frRepresentationInformation retrievalContext evolution.
Machine Perception LabFace Detection and Expression Recognitionhttp markov ucsd edu movellan ... S minaire PSIFRE CNRS 2546.
Too much information16 Janvier 2003scanu insa rouen frkills information DataRepresentation.
Information retrieval We are drowning in information and starving for knowledge Context evolution Rutherford D Roger Critic of the Data Era Data smog.
Non measurable things Ethical consequences the Orwellian futureFilter data S minaire PSI.
FRE CNRS 2546Intelligent sensors16 Janvier 2003scanu insa rouen fr Requirements Representation.
Information retrieval Data Context evolution Accuracy and confidence Self diagnostic Self calibration.
How to do it Uncertainty management Learning ability Network database Adaptation ability.
Fault detection mechanismAssociated software sensorsCanu et al Black box Software Sensor Design for Environmental Monitoring in International Conference onrtificial Neural Networks Skovde Sweden Sep 2 4 1998 and related work on data validation within the EM2S project S minaire PSI.
FRE CNRS 2546Data validation16 Janvier 2003scanu insa rouen fr Mono sensor validation Representation.
Information retrieval Static validation Context evolution Mean variance Dynamic validation Cusum control charts .
Trend analysis Multisensor validation Residual analysis Fusion Joint probability estimation Prior knowledge Balanced relations.
Hierarchical validation Multisensor perceptionInteractive matrix of smart sensors http www accenture com xd xd a... K Van Laerhoven A Schmidt and H W Gellersen Multi Sensor Context Aware Clothing In Proceedings.
of the 6th International Symposium on Wearable Computers 2002 S minaire PSIFRE CNRS 2546Software sensor16 Janvier 2003.
scanu insa rouen frRepresentation Value confidence interval validity domain Information retrieval How to build it Context evolution From a model tracking Kalman filter.
When no model is available learn it Raw data v t Raw data x t environmentRaw data y t .
Raw data z t learning Black box modeling S minaire PSIFRE CNRS 2546Towards proprioceptors.
16 Janvier 2003scanu insa rouen fr Learn RepresentationInformation retrievalContext evolution.
Pr x1 xi xd v How to learn Gaussian mixture EM Include prior Bayesian networks Deal with uncertainty Evidence framework.
Use to Detect non nominal situations Replace missing datad Curse of dimensionality Belman E Petriu et al Sensor based information appliances .
S minaire PSIFRE CNRS 2546What is data 16 Janvier 2003scanu insa rouen fr.
Representation Individuals or measurements Information retrievalContext evolution Associated variables Data set matrix .
line measurements column variable Data point clouds Data exploration recognize patternstoo many data SUMARIZE.
S minaire PSIFRE CNRS 2546Summarize data16 Janvier 2003scanu insa rouen fr.
Representation Non linear components analysis Information retrievalContext evolution Feature space kernel PCA or ICA Local linear.
Quantisation SOM Relevant distance Select features Local adapted representation Feature selection.
Select relevant situations Sparse learning Kernel learningKernel representation J M ntyj rvi J Himberg P Korpip H Mannila Extracting the Context of a Mobile Device User .
8th Symposium on Human Machine Systems HMS Kassel Germany 2001 S minaire PSIFRE CNRS 2546Kernel representation16 Janvier 2003.
scanu insa rouen frDistance mapsExample in 2 dimension of the influence map of the black dist i j I i j exp circle Red color denotes a high influence while the low.
influence zones are in blue Analyze data proximity through the kernel map B Scholkopf and A Smola Leaning with Kernels MIT Press 2001 S minaire PSIFRE CNRS 2546.
Example of kernel map16 Janvier 2003scanu insa rouen frData clouds in two dimensions Associated kernel mapEven in d dimensions you can visualize.
S Canu and al Functionnal learning through kernels invited lecture at the NATO institute in Leuven 2002 S minaire PSIFRE CNRS 2546Looking for hiden shapes16 Janvier 2003.
scanu insa rouen fr RepresentationInformation retrieval Data point information noise Context evolution Principal curve.
Non linear PCA Independent curve Non linear ICAKernel representation linear analysis Bal zs K gl.
http www iro umontreal ca kegl... S minaire PSIFRE CNRS 254616 Janvier 2003scanu insa rouen fr.
in high dimensional space RepresentationInformation retrievalContext evolution J B Tenenbaum V de Silva and J C Langford.
http isomap stanford edu handf... S minaire PSIFRE CNRS 2546Information retrieval16 Janvier 2003.
scanu insa rouen frRepresentation What for Information retrievalContext evolution User profiling.
User identification Battery discharge rate Sequence induction Classification problem Decision theory.
Example based programming Learning machineSelect relevant cases S minaire PSIFRE CNRS 2546.
A brief historical16 Janvier 2003scanu insa rouen frperspective DataRepresentation.
of machine learning Information retrieval Before machines Context evolution Statistics PCA DA regression CART kNN 70 s Learning is logic Grammatical inference in expert systems.
80 s Learning is human Neural networks backprop 90 s Learning is a problem COLT Kernel machines SVM Mixture of experts adaboost.
What is the learning problem T Hastie R Tibshirani and J Friedman The elements of statistical learning Springer 2001 S minaire PSIFRE CNRS 2546What is learning .
16 Janvier 2003scanu insa rouen fr Training set x 1 y1 x i yi x n yn Test point xn 1 looking for f such that y n 1 f xn 1 Learning is balancing.
Fit Summarize1 Hypothesis set Neural networks Kernels 2 Fitting criterion least square absolute value 3 Compression criterion penalization Margin 4 Balancing mechanism cross validation generalization .
Learning is summarizing S minaire PSIFRE CNRS 254616 Janvier 2003scanu insa rouen fr.
discrimination DataRepresentationseparable case Information retrievalContext evolutionall points .
w b Use hyperplane S minaire PSIFRE CNRS 254616 Janvier 2003.
scanu insa rouen frdiscrimination DataRepresentationseparable case Information retrievalContext evolution.
all points S minaire PSIFRE CNRS 2546The classifier16 Janvier 2003.
scanu insa rouen frMargin Representation Information retrievalContext evolutionMargin classify.
all points S minaire PSIFRE CNRS 2546Maximize the16 Janvier 2003.
scanu insa rouen frmargin DataRepresentationBe sparse Information retrievalContext evolution.
wx b 1 correctlyMargin classifyall points wx b 1 .
Support Vector Machines SVM V N Vapnik The nature of statistical learning theory Springer Verlag 1995 S minaire PSIFRE CNRS 2546What is learning .
16 Janvier 2003scanu insa rouen fr Training set x 1 y1 x i yi x n yn Test point xn 1 looking for f such that y n 1 f xn 1 Learning is balancing.
Fit SummarizeContext-aware systems are typically integratedin wearable computers and mobile computing systems(see for example [10]).2. The new context can be created at runtime by the users.The users define the new contextual objectby con-figurationwhich specifies how context-dependent ob-jects are influenced by the newly created contextualobjects.

Related Presentations