Compositional languages emerge in a neural iterated learning model

Authors: Yi Ren, Shangmin Guo, Matthieu Labeau, Shay B. Cohen, Simon Kirby

ICLR 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments confirm our analysis, and also demonstrate that the emerged languages largely improve the generalizing power of the neural agent communication.
Researcher Affiliation Academia 1 University of Edinburgh, United Kingdom, 2 University of Cambridge, United Kingdom 3 LTCI, T el ecom Paris, Institut Polytechnique de Paris, France
Pseudocode Yes Algorithm 1: The NIL algorithm.
Open Source Code Yes The code is available at https://github.com/Joshua-Ren/Neural_Iterated_Learning.
Open Datasets No The paper describes generating its own object space based on attributes (Na) and values (Nv), rather than using a publicly available dataset.
Dataset Splits Yes We measure this ability by looking at their validation game performance: we restrict the training examples to a limited numbers of objects (i.e., the training set), and look at how good are the agents at playing the game on the others (i.e., the validation set)." and "Valid set size 0 8 16 32
Hardware Specification No The paper does not provide specific hardware details such as GPU/CPU models, processor types, or memory amounts used for experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., Python, PyTorch, CUDA versions).
Experiment Setup Yes Unless specifically stated, the experiments mentioned in this paper use the hyper-parameters given in Table 3.