MuDE: Multi-agent decomposed reward-based exploration.

In cooperative multi-agent reinforcement learning, agents jointly optimize a centralized value function based on the rewards shared by all agents and learn decentralized policies through value function decomposition. Although such a learning framework is considered effective, estimating individual contribution from the rewards, which is essential for learning highly cooperative behaviors, is difficult. In addition, it becomes more challenging when reinforcement and punishment, help in increasing or decreasing the specific behaviors of agents, coexist because the processes of maximizing reinforcement and minimizing punishment can often conflict in practice. This study proposes a novel exploration scheme called multi-agent decomposed reward-based exploration (MuDE), which preferably explores the action spaces associated with positive sub-rewards based on a modified reward decomposition scheme, thus effectively exploring action spaces not reachable by existing exploration schemes. We evaluate MuDE with a challenging set of StarCraft II micromanagement and modified predator-prey tasks extended to include reinforcement and punishment. The results show that MuDE accurately estimates sub-rewards and outperforms state-of-the-art approaches in both convergence speed and win rates.

Biofilm-forming ability of Staphylococcus aureus strains isolated from human skin.

BACKGROUND: Staphylococcus aureus produces various toxins and enzymes, and its presence can exacerbate skin conditions. Previous studies have shown that S. aureus is involved in skin deterioration, even in normal tissue. Biofilm strains show much greater resistance to antimicrobial agents and therefore require a much higher concentration of biocide than planktonic counterparts. OBJECTIVE: As such, alternative strategies and more effective therapeutic agents against biofilm-producing S. aureus in skin are of great interest. Therefore, we turned our attention to differences in 50 clinical biofilm strains isolated from human facial skin. METHODS: Based on S. aureus density on facial skin, we divided donors into two groups: relatively low density (LSG) and high density (HSG). In general, strong biofilm-forming strains were detected in the HSG donors. Two strains from each of the groups were submitted to gene microarray analysis to investigate expression differences and confirmed by RT-PCR. RESULTS: In total, 111 of 7775 genes were differentially expressed between low (SA2 and SA7) vs. high (SA10 and SA33) biofilm-forming clinical strains. These genes include already well-known as biofilm formation related genes like icaABCD and lrgAB, and newly identified genes (sdrC, sspBCP) by RT-PCR. Comparison of gene expression differences between the two groups available at NCBI Gene Expression Omnibus accession number GSE44268. CONCLUSION: Our results suggest that S. aureus density in the skin is closely related to biofilm-forming ability, and we have identified several potential target genes that may be involved in regulating biofilm formation in situ.