https://www.nature.com/articles/s41598-025-18449-1 Problem기존 deep reinforcement learning traffic signal conrol은 보통 intersection마다 하나의 RL agent를 둔다.문제는, 각 intersection이 local objective만 최적화하면, 전체 road network의 traffic efficiency는 오히려 최적이 아닐 수 있을 수 있다는 점따라서 "How should multiple intersection coordinate so that local decisions also improve regional traffic" 이라는 질문을 던지는 SHLight (Sample selection-ba..

출처: https://dl.acm.org/doi/10.1145/3219819.3220096 IntelliLight | Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data MiningEmploying an optimal traffic light control policy has the potential of having a positive impact, both economic and environmental, on urban mobility. Reinforcement learning techniques have shown promising results in optimizing control po..

https://elifesciences.org/reviewed-preprints/87517 Statistical analysis supports pervasive RNA subcellular localization and alternative 3’ UTR regulationeLife assessment This paper describes an important, well-organized study into an under-exploited area of spatial transcriptomics. The limitations of the approach are generally made clear, but there is insufficient orthogonal validation to demons..

출처: https://www.nature.com/articles/s41467-024-49457-w PROBLEM STATEMENT"colocalization" 즉, 두 gene의 RNA들이 세포 안에서 서로 가까이 위치하는 경향이 있는가?1. PP testProximal Pair test를 통해 한 cell 안에서 gene A와 B의 transcript pair가 거리 d 이내에 있는 경우가 우연히 기대되는 것보다 많은지 확인한다. 단, 단순히 거리 d 이내에 있는 경우의 수를 세는 것이 아니라, 그 cell 자체의 spatial density를 null background에 반영해 비교한다. 이는 cell마다 크기도 다르고, RNA density도 다르고, transcript 수 자체도 다르기 때문이..

출처: https://www.biorxiv.org/content/10.64898/2026.01.20.700581v2 MANTIS: Analytics toolkit for spatial metabolomics with matching spatial transcriptomics dataMotivation Joint Spatial Metabolomics (SM) and Spatial Transcriptomics (ST) profiling is a powerful approach to fine-mapping of metabolic states associated with tissue function. Current computational tools for analysis of “SM+ST” data focus..

AbstractAutomated Traffic Signal Performance Measures (ATSPMs) have been widely adopted by transportation agencies for real‐time monitoring and optimization of signal operations, yet their application to intersection safety remains limited and mainly based on small sample sizes. To address this gap, an eighteen-month dataset of cycle-level ATSPM metrics was integrated with geolocated crash recor..