SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2201.08484 · ICLR · 2022

Iterated Reasoning with Mutual Information in Cooperative and Byzantine Decentralized Teaming

Sachin Konan, Esmaeil Seraj, Matthew Gombolay

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 10 functions out of this paper's own repositories and ran 7 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
deepmind/pysc2 canonical 6 of 7
core-robotics-lab/infopg canonical 1 of 3
FunctionStatusWhere it lives
MOA_PolicyHelper Ran core-robotics-lab/infopg/pistonball/policy_piston.py
pointer only (licence: GPL-3.0) · get_code("593ef84a0fec9ab2")
memoize Ran deepmind/pysc2/pysc2/lib/memoize.py
pointer only (licence: Apache-2.0) · get_code("1314ce1206a217e8")
numpy_to_python Ran deepmind/pysc2/pysc2/lib/actions.py
pointer only (licence: Apache-2.0) · get_code("795b48aa13444f75")
piece_wise_linear Ran deepmind/pysc2/pysc2/lib/colors.py
pointer only (licence: Apache-2.0) · get_code("1902983401727afd")
shuffled_hue Ran deepmind/pysc2/pysc2/lib/colors.py
pointer only (licence: Apache-2.0) · get_code("094587ca8ce60a1a")
smooth_hue_palette Ran deepmind/pysc2/pysc2/lib/colors.py
pointer only (licence: Apache-2.0) · get_code("ee4a38c9d2b4ba59")
summarize_array_diffs Ran deepmind/pysc2/pysc2/lib/np_util.py
pointer only (licence: Apache-2.0) · get_code("b9be4b0c3dee19d6")
spatial Not yet run deepmind/pysc2/pysc2/lib/actions.py
pointer only (licence: Apache-2.0) · get_code("93e90c9405886816")
train_adv_infopg_5 Not yet run core-robotics-lab/infopg/pistonball/test_piston_ball.py
pointer only (licence: GPL-3.0) · get_code("d04972da01192ed6")
train_infopg_5 Not yet run core-robotics-lab/infopg/pistonball/test_piston_ball.py
pointer only (licence: GPL-3.0) · get_code("8673d75351fb3782")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Information sharing is key in building team cognition and enables coordination and cooperation. High-performing human teams also benefit from acting strategically with hierarchical levels of iterated communication and rationalizability, meaning a human agent can reason about the actions of their teammates in their decisionmaking. Yet, the majority of prior work in Multi-Agent Reinforcement Learning (MARL) does not support iterated rationalizability and only encourage inter-agent communication, resulting in a suboptimal equilibrium cooperation strategy. In this work, we show that reformulating an agent's policy to be conditional on the policies of its neighboring teammates inherently maximizes Mutual Information (MI) lower-bound when optimizing under Policy Gradient (PG). Building on the idea of decision-making under bounded rationality and cognitive hierarchy theory, we show that our modified PG approach not only maximizes local agent rewards but also implicitly reasons about MI between agents without the need for any explicit ad-hoc regularization terms. Our approach, InfoPG, outperforms baselines in learning emergent collaborative behaviors and sets the state-of-the-art in decentralized cooperative MARL tasks. Our experiments validate the utility of InfoPG by achieving higher sample efficiency and significantly larger cumulative reward in several complex cooperative multi-agent domains.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2201.08484")
get_code_for_paper("2201.08484")
have("2201.08484")

Connect an agent — have() is free.