Skip to content

Latest commit

 

History

History
43 lines (41 loc) · 3.57 KB

doc_experiment_class.md

File metadata and controls

43 lines (41 loc) · 3.57 KB

Experiment class

Each experiment is described via an Experiment object, which possesses the following attributes:

  • exec Prefix of the executable. It is set to aurora for all experiments.
  • algo Name of the algorithm to consider. The main values for that variable are:
    • For AURORA:
      • aurora_uniform - AURORA with a uniform selector
      • aurora_novelty - AURORA with a novelty selector
      • aurora_surprise - AURORA with a surprise selector (the surprise score corresponds to the error of reconstruction by the encoder)
      • aurora_nov_sur - AURORA mixing a novelty (50%) and surprise (50%) selectors
      • aurora_curiosity - AURORA with a curiosity selector
    • For TAXONS:
      • taxons - TAXONS (using a selection procedure based on Novelty and Surprise) as described in Unsupervised Learning and Exploration of Reachable Outcome Space (Paolo et al., 2020)
      • taxo_n - TAXO_N (using a selection only based on Novelty) as described in (Paolo et al., 2019)
      • taxo_s - TAXO_S (using a selection only based on Surprise) as described in (Paolo et al., 2019)
    • For the Hand-coded baselines:
      • hand_coded_qd - Hand-Coded QD algorithm using an unstructured archive with 2-dimensional hand-coded behavioural descriptors, and a uniform selector
      • hand_coded_qd_no_sel - Hand-Coded QD algorithm without any selector (new individuals are generated randomly). This is equivalent to random search.
      • hand_coded_taxons - Equivalent to Novelty Search, as described in Novelty Search makes Evolvability Inevitable (Doncieux et al., 2020)
  • env Name of the environment/task to consider. The main values for that variable are:
    • hard_maze - Maze task
    • hexa_cam_vertical - Hexapod task
    • air_hockey - Air-Hockey task
  • latent_space - Number of dimensions of the Behavioural Descriptor. In the case of AURORA and TAXONS, it corresponds to the number of dimensions of the latent space.
  • fixed_l - If that value is not None, then the distance threshold
  • encoder_type - Type of encoder used to encode the data. It is mostly equal to one of the following elements:
    • EncoderType.cnn_ae for the maze and hexapod tasks
    • EncoderType.mlp_ae for the air-hockey task
    • EncoderType.none for the hand-coded variants
  • lp_norm - decides which L^p norm to use to compute distances (set to 2 in all our experiments)
  • has_fit - If false, the fitness of individuals is not taken into account: the algorithm then becomes a pure divergent search procedure. It is set to True for all experiments (except for TAXONS and Novelty Search, which do have any mechanism to consider the fitness).
  • use_volume_adaptive_threshold (set to False by default)
    • If False, the AURORA experiments use the Container Size Control technique (CSC).
    • If True, the AURORA experiments use the Volume Adaptive Threshold technique (VAT).
  • taxons_elitism - If true, TAXONS and Novelty Search select the best individuals from the set {parents + offspring}. If false, the parents are always discarded.
  • update_container_period - Value of the container update period T_{\mathcal{C}} (10 by default)
  • coefficient_proportional_control_l - Value of K_{CSC} (5e-6 by default)

Other possible values for those variables can be found in the collections_experiments folder or in the compilation_variable file.