93887 research outputs found
Sort by
VILP: Imitation Learning with Latent Video Planning
Get PDFIn the era of generative AI, integrating video generation models into robotics opens new possibilities for the general-purpose robot agent. This letter introduces imitation learning with latent video planning (VILP). We propose a latent video diffusion model to generate predictive robot videos that adhere to temporal consistency to a good degree. Our method is able to generate highly time-aligned videos from multiple views, which is crucial for robot policy learning. Our video generation model is highly time-efficient. For example, it can generate videos from two distinct perspectives, each consisting of six frames with a resolution of 96 × 160 pixels, at a rate of 5 Hz. In the experiments, we demonstrate that VILP outperforms the existing video generation robot policy across several metrics: training costs, inference speed, temporal consistency of generated videos, and the performance of the policy. We also compared our method with other imitation learning methods. Our findings indicate that VILP can rely less on extensive high-quality task-specific robot action data while still maintaining robust performance. In addition, VILP possesses robust capabilities in representing multi-modal action distributions. Our paper provides a practical example of how to effectively integrate video generation models into robot policies, potentially offering insights for related fields and directions
The FEW-Nexus: Using soil to grow meaning and relevance in undergraduate general education Earth science courses
Get PDFSoils are central to human survival, through their critical roles in regulating water supplies and providing the medium for most agricultural productivity. Even as fewer and fewer students know where their water and food come from, and where their wastes go, soils tend to be under-emphasized in Earth science general education courses. Combining a place-based focus on soils with the FEW-Nexus model provides an integrating context for concepts and information that students sometimes perceive as disconnected and irrelevant to them. This approach can transform the experiences and learning of undergraduate students in the Earth science general education curriculum
Co-learning about interactions of the WEF Nexus in a Páramo ecosystem: a proposal to bring sustainability education to rural communities in an Andean region (Colombia)
No full textDuring the 2024´s summer, the Colombia-USA Water-Energy-Food Nexus Alliance for the Chingaza Páramo launched its first diploma course, “Caring for the High Mountains: Water-Energy-Food Nexus”. The program aimed to strengthen the exchange of knowledge between academia and rural populations living in the buffer zones of Chingaza National Park to contribute to the conservation of the Páramo from the perspective of the local territory and its inhabitants. 61 residents enrolled from two Páramo buffer zone communities: La Calera and Choachí 20 virtual presentations and 5 in-person workshops were held from June-August, 2024 Participants found that the educational foundations built through the diploma “planted the seeds” and “created linkages” for inter-community collaboration on strengthening local conservation practice
Stress, Coping, and Psychological Growth in Personnel in a High Arctic Weather Station
Get PDFIdeas about positive change following a stressful experience have been of interest to researchers for some time (Antonovsky, 1987; Tedeschi & Calhoun, 1996). A few areas of study have been developed, some with a focus on successful coping such as research on resilience, sense of coherence and hardiness, stress inoculation, posttraumatic growth, and toughening (Tedeschi, Park, and Calhoun, 1998). Suedfeld (2001) has encouraged researchers to focus on the positive effects of work in extreme and unusual environments however, only a few have done so with astronauts, cosmonauts (Ihle et al., 2006: Suedfeld et al., 2012), and high-altitude mountaineers (Smith et al., 2017).To understand the full deployment experience and growth possibilities and to identify the situational impact on the personnel in extreme and unusual environments, a study of stress perception and coping strategies is essential. Work in this domain of stress and coping has been plentiful across many extreme and unusual environments (Leon et al., 2011b; Nicolas et al., 2013; Suedfeld et al., 2009; Suedfeld et al., 2012; Suedfeld, 2015).The current study addresses an additional extreme workplace environment with a combination of stress and outcome variables. Crew members working and living at the Eureka Weather Station (Nunavut, Canada), located on Ellesmere Island at a latitude of 80° North, responded to questionnaires about coping (prior to the mission and postmission), personality (prior to mission), and psychological growth (postmission). Members spent two to four months working in the Canadian High Arctic and faced many environmental and social challenges. Data suggest that the extreme workplace is not as stressful as commonly assumed; further, members use a variety of appropriate coping mechanisms, and postexpedition growth is experienced by all
Principles of Writing Construction Specifications and Special Provisions
Get PDFThis presentation will assist project managers and designers in preparing the necessary information required for Department contracts involving Standard Specifications and Special Provisions. Attendees will be exposed to examples of both the right and wrong way to present a Unique Special Provision, helpful tips on how to navigate the new Sharepoint review process, and questions to ask to adhere to legal, moral, and ethical construction standards
Enhancing Surgical Tool Classification Using a Semi-supervised Learning Autoencoder Approach
Get PDFThe purpose of this study is to develop a robust approach for identifying and classifying surgical instruments while addressing the challenges associated with traditional supervised learning methods, which require large annotated datasets. By comparing the performance of supervised machine learning and deep learning models with a semi-supervised autoencoder approach, this study aims to enhance efficiency and accuracy in automated and robotic-assisted surgical procedures. This will ultimately improve workflow optimization, error minimization, and real-time decision-making in the operating roo
Pulse Flow Reverse Osmosis with Feed Flow Reversal
Get PDFThe purpose of this study is to increase the efficiency and water recovery of household scale reverse osmosis systems. Due to the small membranes and simple design, current household RO systems send much of the feed water to the drainage system. This system will be capable of implementing Feed Flow reversal in order to increase the lifespan of the RO membrane as Pulse Flow RO may lead to uneven loading due to variances in salinity and increased pressure spikes
Mentoring to Retain Engineering Students and Graduates
Get PDFITE has identified work force development as one of three key issues facing engineering firms today; yet, as many as 50% of all graduate engineers may leave the engineering profession within 10 years of graduation. Mentors and mentees will discuss the role that mentorship can play in engineering student and graduate engineer retention in the work force
Nexus Between NEPA and Utilities
Get PDFIntegrating utility coordination within the NEPA framework is crucial for ensuring environmental compliance while also meeting project timelines. This presentation will explore the intersection of these processes, identify key challenges, and outline best practices to streamline coordination. Case studies from recent projects will illustrate effective collaboration and procedural efficiency and offer insights into achieving regulatory compliance and project goals