Adaptation in InfoSpiders

• Unsupervised individual adaptation

  • Q-learning: adjust neural net weights to predict relevance locally - the measured relevance is used as reinforcement signal

    E(Da) + m * max l ll - is used as teaching input
    

• Unsupervised population evolution

  • Select agents to match resource bias
  • Mutate internal queries: selective query expansion
  • Mutate weights and other parts of genotype

Previous slide

Next slide

Back to first slide

View graphic version