Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
MALinZero: Efficient Low-Dimensional Search for Mastering Complex Multi-Agent Planning
Authors: Sizhe Tang, Jiayu Chen, Tian Lan
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | MALin Zero demonstrates state-of-the-art performance on multi-agent benchmarks such as matrix games, SMAC, and SMACv2, outperforming both model-based and model-free multi-agent reinforcement learning baselines with faster learning speed and better performance. |
| Researcher Affiliation | Academia | Sizhe Tang The George Washington University EMAIL Jiayu Chen Carnegie Mellon University EMAIL Tian Lan The George Washington University EMAIL |
| Pseudocode | Yes | Algorithm 1 MALin Zero 1: procedure DYNAMIC NODE GENERATION 2: a arg maxa A a θ + c(s)P(s, a)trace(V ) a V 1a 3: return (s,a) 4: end procedure 1: procedure EXPANSION |
| Open Source Code | Yes | We provide all the information needed to reproduce the results presented in this paper together with the source code. |
| Open Datasets | Yes | We evaluate MALin Zero on three reinforcement learning benchmarks: Mat Game, Star Craft Multi Agent Challenge (SMAC)[22] and SMACv2 [23]. |
| Dataset Splits | No | The paper refers to environments like Mat Game, SMAC, and SMACv2, and mentions 'Training Steps' and 'Evaluation episodes' but does not specify explicit training/test/validation dataset splits with percentages or sample counts. |
| Hardware Specification | Yes | All experiments are conducted using NVIDIA RTX A6000 GPUs and NVIDIA A100 GPUs. |
| Software Dependencies | No | The paper mentions optimizers like 'Adam' and 'RMSprop epsilon' and refers to frameworks like 'Efficient Zero [40]' and 'MAZero [10]', but it does not provide specific version numbers for general software dependencies or programming languages used. |
| Experiment Setup | Yes | Table 2: Hyper-parameters for MALin Zero in Mat Game, SMAC and SMACv2 environments |