• Like
The COCH project
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

The COCH project

  • 277 views
Published

 

Published in Technology , Education
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
277
On SlideShare
0
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
4
Comments
0
Likes
0

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. M. Barreto, D. Jimenez, H. Satizabal, Andrés Pérez-Uribe, Eduardo Sanchez REDS Institute (http://reds.eivd.ch) University of Applied Sciences of Western-Switzerland - HEIG-VD The COCH project 27.02.06
  • 2. Modeling for the agro
    • Bio-inspired & statistical computational techniques capable of producing complex models to predict/describe the site-specific behavior of given crops
  • 3. Bio-inspired systems: family album
    • Artificial intelligence
    • Expert systems
    • Computational intelligence (IEEE)
      • artificial neural networks
      • evolutionary computation
      • fuzzy systems
    • Machine learning
    • Modern heuristics
    • Statistical learning
    • Data mining
    By studying, understanding and exploiting nature’s “tricks” and “artifices”, engineers can provide innovative solutions to engineering problems
  • 4. Features of bio-inspired techniques
    • Bio-inspired techniques have proven to be powerful tools for modeling and prediction using numerical noisy data and non numerical information.
    • Bio-inspired techniques appear to perform quite well without strong assumptions on the data.
    • Certain techniques provide innovative ways to process and visualize highly-dimensional information.
    • The Fuzzy logic formalism enables the integration of imprecise information (e.g., expert knowledge) and the generation of human understandable outputs.
  • 5. Modeling process numerical data noisy data non-numerical data training test design loop model validation loop model inputs prediction explanation/ visualization
  • 6. hypothesis modeling simulation prediction explanation
  • 7. Prediction-oriented experiments using the CENICA ÑA database highly dimensional numerical data (temp,HR, radiacion, préc., edad cosecha, num. cortes) non-numerical data (zona agroecologica) supervised training simulation prediction
  • 8. Characterization-oriented experiments using the CENICA ÑA database highly dimensional numerical data (temp,HR, radiacion, préc., edad cosecha, corte ) non-numerical data (zona agroecologica) self-organization simulation test/prediction explanation & visualization
  • 9. Research issues I
    • Incremental modeling:
    • Construction of models using growing data bases is a challenging issue. In our case, the information of fruit crops will be continuously collected along the modeling process, for this reason, the model must be able to adapt its parameters according to the changes of incoming information. This process is closely related to continuous online learning systems in which the model structure has to be plastic but stable enough to be able to learn new characteristics of data while retaining previous information
  • 10. Research issues II
    • Integration of heterogeneous information:
    • The modeling methodology has to allow for the possibility to include information gathered from multiple sources. We are interested to include expert knowledge, traditional knowledge, and information obtained by agronomical and climate analysis. In order to deal with multiple sources of diverse nature of data, we propose to develop a so-called “mixture of experts” approach.
  • 11. Research issues III
    • Intelligent visualization:
    • Building a model is rarely an end in itself; instead, the goal of most analysis is to make a decision. To assist in this analysis, we propose the development of intelligent interfaces that allow for visual decision support based in bio-inspired techniques and data mining.
  • 12. COCH 3i (“triple I”) research I ncremental modeling I ntegration of heterogeneous data I ntelligent visualization 4th dimension: model validation (usefulness & biological response) model exploitation