Inspired by animal navigation strategies, we introduce a novel computational model to navigate and map a space rooted in biologically inspired principles. Animals exhibit extraordinary navigation prowess, harnessing memory, imagination, and strategic decision-making to traverse complex and aliased environments adeptly. Our model aims to replicate these capabilities by incorporating a dynamically expanding cognitive map over predicted poses within an active inference framework, enhancing our agent's generative model plasticity to novelty and environmental changes. Through structure learning and active inference navigation, our model demonstrates efficient exploration and exploitation, dynamically expanding its model capacity in response to anticipated novel un-visited locations and updating the map given new evidence contradicting previous beliefs. Comparative analyses in mini-grid environments with the clone-structured cognitive graph model (CSCG), which shares similar objectives, highlight our model's ability to rapidly learn environmental structures within a single episode, with minimal navigation overlap. Our model achieves this without prior knowledge of observation and world dimensions, underscoring its robustness and efficacy in navigating intricate environments.

Daria de Tinguy

Tim Verbelen

Bart Dhoedt

Frontiers in Computational Neuroscience

FRONTIERS IN COMPUTATIONAL NEUROSCIENCE 是一本一流的电子期刊，致力于推广大脑功能的理论建模，促进理论和实验神经科学之间的跨学科互动。对大脑惊人能力的理解进展仍然有限，我们相信只有深入的理论思考和不同学科和方法之间的相互促进合作才能实现。因此，我们邀请对广泛主题的原创贡献，以展示这种合作的成果，或为未来的联盟提供刺激。我们的目标是为神经系统的前沿理论研究提供一个互动论坛，并向更广泛的神经科学界传播最好的理论研究。从生物物理驱动的神经元和突触的逼真模拟到推理和决策的高级抽象模型，各种风格和所有级别的模型都受到欢迎。虽然该杂志主要侧重于基于理论和驱动的研究，但我们欢迎验证和测试理论结论的实验研究。另外：COMP NEURO

Frontiers in Computational Neuroscience is a leading electronic journal dedicated to advancing theoretical modeling of brain functions and fostering interdisciplinary interactions between theoretical and experimental neuroscience. Our understanding of the brain's remarkable capabilities remains limited, and we believe that only through deep theoretical reflection and collaborative efforts across various disciplines and methods can we achieve breakthroughs. Therefore, we invite original contributions on a wide range of topics to showcase the results of such collaboration or to inspire future alliances. Our goal is to provide an interactive forum for cutting-edge theoretical research on the nervous system and to disseminate the best theoretical studies to the broader neuroscience community. From biophysically driven realistic simulations of neurons and synapses to advanced abstract models of reasoning and decision-making, a variety of styles and all levels of models are welcome. While the journal primarily focuses on theory-driven research, we also welcome experimental studies that validate and test theoretical conclusions.

《Frontiers in Computational Neuroscience》是一本顶尖的电子期刊，致力于推动大脑功能的理论建模，促进理论与实验神经科学之间的跨学科互动。尽管我们对大脑的惊人能力的理解仍然有限，但我们相信，只有通过深入的理论思考和不同学科及方法之间的相互促进合作，才能取得突破。因此，我们邀请对广泛主题的原创贡献，以展示这种合作的成果，或为未来的联盟提供启发。我们的目标是为神经系统的前沿理论研究提供一个互动论坛，并向更广泛的神经科学界传播最优秀的理论研究。从生物物理驱动的神经元和突触的逼真模拟，到推理和决策的高级抽象模型，各种风格和所有级别的模型都受到欢迎。虽然该期刊主要侧重于基于理论和驱动的研究，但我们也欢迎验证和测试理论结论的实验研究。

Category	Quartile
医学	4
医学, 数学与计算生物学	3
医学, 神经科学	4

Learning dynamic cognitive map with autonomous navigation

ivySCI AI Smartly Parses PDF, Answers Researchers' Questions, and Helps You Understand Papers in Seconds

Journal Info