REVIEW 7 cited by
QPLEX: Duplex Dueling Multi-Agent Q-Learning
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
We explore value-based multi-agent reinforcement learning (MARL) in the popular paradigm of centralized training with decentralized execution (CTDE). CTDE has an important concept, Individual-Global-Max (IGM) principle, which requires the consistency between joint and local action selections to support efficient local decision-making. However, in order to achieve scalability, existing MARL methods either limit representation expressiveness of their value function classes or relax the IGM consistency, which may suffer from instability risk or may not perform well in complex domains. This paper presents a novel MARL approach, called duPLEX dueling multi-agent Q-learning (QPLEX), which takes a duplex dueling network architecture to factorize the joint value function. This duplex dueling structure encodes the IGM principle into the neural network architecture and thus enables efficient value function learning. Theoretical analysis shows that QPLEX achieves a complete IGM function class. Empirical experiments on StarCraft II micromanagement tasks demonstrate that QPLEX significantly outperforms state-of-the-art baselines in both online and offline data collection settings, and also reveal that QPLEX achieves high sample efficiency and can benefit from offline datasets without additional online exploration.
Forward citations
Cited by 7 Pith papers
-
Zero-Incentive Dynamics: a look at reward sparsity through the lens of unrewarded subgoals
Unrewarded bottleneck transitions, called zero-incentive dynamics, cause state-of-the-art subgoal-based RL methods to fail, and learning quality degrades sharply with delay between subgoal completion and reward.
-
An Extended Benchmarking of Multi-Agent Reinforcement Learning Algorithms in Complex Fully Cooperative Tasks
Algorithms that set state of the art on SMAC and GRF often underperform standard baselines on fully cooperative benchmarks, including image-based tasks.
-
Active Domain Knowledge Acquisition with 100-Dollar Budget: Enhancing LLMs via Cost-Efficient, Expert-Involved Interaction in Sensitive Domains
A budget-aware framework (PU-ADKA) selects which domain expert an LLM should query under a fixed $100 budget, improving specialized-domain answers at low cost.
-
ToMacVF : Temporal Macro-action Value Factorization for Asynchronous Multi-Agent Reinforcement Learning
A temporal macro-action value factorization method with a segmented replay buffer improves asynchronous multi-agent RL performance, but the claimed proof that it generalizes standard IGM is invalid.
-
Speaking the Language of Teamwork: LLM-Guided Credit Assignment in Multi-Agent Reinforcement Learning
LLM-generated, agent-specific potential-based rewards accelerate sparse-reward MARL training in grid world and pistonball benchmarks.
-
Double Distillation Network for Multi-Agent Reinforcement Learning
DDN, a double distillation network for cooperative MARL, claims to eliminate CTDE inherent error via global-to-local knowledge distillation and an RND-style intrinsic reward, with better SMAC and Predator-Prey win rates.
-
Novel Multi-Agent Action Masked Deep Reinforcement Learning for General Industrial Assembly Lines Balancing Problems
A multi-agent, action-masked deep RL framework for general assembly line balancing that claims faster convergence to optimal schedules and polynomial action-space growth on small benchmark instances.
Discussion (0). Continue with ORCID to comment.