Revisiting Populations in multi-agent Communication
Authors: Paul Michel, Mathieu Rita, Kory Wallace Mathewson, Olivier Tieleman, Angeliki Lazaridou
ICLR 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this paper we reassess the validity of the standard training protocol and illustrate its limitations. Specifically, we analyze population-level communication at the equilibrium in sender-receiver Lewis games. We find that receivers co-adapt to senders they are interacting with, which limits the effect of the population. Informed by this analysis, we propose an alternative training protocol based on partitioning agents. Partitioning isolates sender-receiver pairs, limits co-adaptation, and results in a new global optimization objective where agents maximize (1) their respective internal communication accuracy and (2) their alignment with other agents. In experiments, we find that agents trained in partitioned populations are able to communicate successfully with new agents which they have never interacted with and tend to develop a shared language. Moreover, we observe that larger populations develop languages that are more compositional. Our findings suggest that scaling up to populations in multi-agent communication can be beneficial, but that it matters how we scale up. |
| Researcher Affiliation | Collaboration | Paul Michel ENS-PSL Mathieu Rita INRIA, Paris Kory Mathewson Deep Mind Olivier Tieleman Deep Mind Angeliki Lazaridou Deep Mind |
| Pseudocode | No | The paper describes procedures and algorithms in text, but it does not include a formally labeled pseudocode block or algorithm figure. |
| Open Source Code | Yes | Code to reproduce our experiments can be found at https://github.com/pmichel31415/EGG |
| Open Datasets | Yes | Image Net (Deng et al., 2009) dataset of natural images. |
| Dataset Splits | Yes | In each setting we hold out 1, 000 combinations to be used as a validation, and 1, 000 more for use as a test set. |
| Hardware Specification | No | The paper does not explicitly describe the specific hardware used to run its experiments, such as GPU models, CPU models, or detailed memory specifications. |
| Software Dependencies | No | The paper mentions "Our implementation, based on the EGG toolkit (Kharitonov et al., 2021)" but does not provide specific version numbers for EGG or any other software dependencies (e.g., Python, PyTorch/TensorFlow versions). |
| Experiment Setup | Yes | In all experiments we train with a batch size of 1024 with the Adam optimizer (Kingma & Ba, 2014) using a learning rate of 0.001 for the attribute/value dataset and 0.0001 for Imagenet. The other parameters are set to β1 = 0.9, β2 = 0.999 and ε = 10 8. We apply ℓ2 regularization with a coefficient of 10 5. |