136879 research outputs found
Sort by
A Metric for Intrinsic Motivation in Reinforcement Learning Agents
Get PDFClassically, the reward for an agent is given by extrinsic factors which motivate the agent to improve and learn; however, an active area of research within cognitive science and AI is the effect and necessity of intrinsic motivation for an agent. This can manifest itself in many forms from curiosity to reduction of cognitive dissonance to motivation for effectance. Despite the prevalence and perceived importance of intrinsic motivation, there is no metric to measure ���how��� intrinsically motivated an agent is compared to another and is instead, largely empirical. Furthermore, methods that might be stated to be intrinsically motivated can be directly linked to the environment and thus, might be less intrinsically motivated than thought. Thus this thesis presents a general metric for intrinsically motivated agents to suggest that highly intrinsically motivated agents are more robust than less intrinsically motivated agents. First, an overview and review of reinforcement learning and intrinsic motivation is presented. Following this, a general metric is proposed with empirical and mathematical justification to measure the intrinsic motivation of an agent. Lastly, several intrinsic motivation agents are tested to evaluate the metric and compare the relative performance of the agents
Connection Strengthening of Wireless Data Delivery for Metal Pipeline Avoidance in Excavation Work
No full textA real-time pipeline detection system (RPDS) has been proposed to prevent pipeline collisions caused by excavation. Excavators have a considerable potential to cause significant damage to both pipelines and the excavator machinery while performing excavation work on construction projects. There is a need for a device that can provide real-time distance estimation information to excavator operators in order to prevent accidental collisions with ferromagnetic pipelines beneath the Earth���s surface. This device, the RPDS, will utilize magnetic anomaly detection (MAD) to provide distance estimations through a distance estimation algorithm. In order to deliver real-time distance estimation information to the excavator operator, it is crucial that the sensor information delivery throughout the system remains stable. A wired connection between the sensors and the computation unit (e.g., a computer, smartphone, or tablet) would provide the most stable data delivery. However, physical constraints in an actual construction workspace prevent a wired connection. A completely wireless system would address the physical limitations. However, a completely wireless system starting at the sensors would be unstable from interference from the excavator body and the construction site. Therefore, a system that retains a stable flow of real-time wireless data delivery needs to be designed while also considering an actual construction site with a full-sized excavator
Variations in Island Adaptations: Studying Paleodietary and Paleomobility Patterns of Ancient Aboriginals of the Canary Islands via Archaeometry
No full textArchaeologists employ stable isotopes of strontium (^87Sr/^86Sr) and oxygen (��^18O) to trace ancient mobility based on established regional baselines. Additionally, stable isotope analyses (SIA) of carbon (��^13C) and nitrogen (��^15N) from bone and teeth offer insights into human paleodietary patterns by examining stable isotope data from animals and plants exploited by ancient populations. Nevertheless, challenges arise in distinguishing marine from terrestrial diets in regions where both land and marine trophic chains intersect. The study of ��^18O values in mammalian bones is tightly related to hydration patterns and water sources, which can be inaccessible in regions like the Canary Islands, where many natural water sources have vanished over time. Archaeology has yet to fully explain the origin and development of the prevalent cultural, socio-political, and economic variation among archaeological populations of the Canary Islands. This research employs the reticulate model and geographic system analysis to investigate ancient subsistence patterns and paleomobility of pre-Hispanic aboriginals from Tenerife and Gran Canaria via SIA combined with newly generated radiocarbon (^14C) dates. The findings reveal challenges in distinguishing marine and terrestrial paleodiets using human bone collagen ��^13C and the dominant environmental signal expressed in human bone ��^15N values from arid regions of Tenerife. By incorporating previously published stable isotope data from ancient Canary humans, my research explores the potential for detecting inter-insular mobility through ��^18O analysis. Notably, this study represents the first analysis of ^87Sr/^86Sr values derived from ancient Canary human bones. Comparing this data with biologically available island ^87Sr/^86Sr values, the method proves its validity in determining local individuals, conclusively indicating the absence of foreigners among the studied populations. Interestingly, I observed significant differences in ^87Sr/^86Sr values between coastal and inland areas in both archaeological humans and baseline material. My research considers the implications of using inconsistent and problematic ^14C date calibration from human bones in chronologies for island occupation. The results presented here provide a foundation for further sophisticated examination of SIA on human bones, urging archaeologists to prioritize the establishment of reliable paleodietary and paleomobility baselines
