10,000 publications from this institution
Humanoid robots can, in principle, use their legs to go almost anywhere. Developing controllers capable of traversing diverse terrains, however, remains a considerable challenge. Classical controllers are hard to generalize broadly while the learning-based methods have primarily focused on gentle terrains. Here, we present a learning-based approach for blind humanoid locomotion capable of traversing challenging natural and man-made terrain. Our method uses a transformer model to predict the next action based on the history of proprioceptive observations and actions. The model is first pre-trained on a dataset of flat-ground trajectories with sequence modeling, and then fine-tuned on uneven terrain using reinforcement learning. We evaluate our model on a real humanoid robot across a variety of terrains, including rough, deformable, and sloped surfaces. The model demonstrates robust performance, in-context adaptation, and emergent terrain representations. In real-world case studies, our humanoid robot successfully traversed over 4 miles of hiking trails in Berkeley and climbed some of the steepest streets in San Francisco.
In vivo molecular imaging holds promise for understanding the underlying mechanisms of health, injury, aging, and disease, as it can detect distinct biochemical processes such as enzymatic activity, reactive small-molecule fluxes, or post-translational modifications. Current imaging techniques often detect only a single biochemical process, but, within whole organisms, multiple types of biochemical events contribute to physiological and pathological phenotypes. In this report, we present a general strategy for dual-analyte detection in living animals that employs in situ formation of firefly luciferin from two complementary caged precursors that can be unmasked by different biochemical processes. To establish this approach, we have developed Peroxy Caged Luciferin-2 (PCL-2), a H(2)O(2)-responsive boronic acid probe that releases 6-hydroxy-2-cyanobenzothiazole (HCBT) upon reacting with this reactive oxygen species, as well as a peptide-based probe, z-Ile-Glu-ThrAsp-D-Cys (IETDC), which releases D-cysteine in the presence of active caspase 8. Once released, HCBT and D-cysteine form firefly luciferin in situ, giving rise to a bioluminescent signal if and only if both chemical triggers proceed. This system thus constitutes an AND-type molecular logic gate that reports on the simultaneous presence of H(2)O(2) and caspase 8 activity. Using these probes, chemoselective imaging of either H(2)O(2) or caspase 8 activity was performed in vitro and in vivo. Moreover, concomitant use of PCL-2 and IETDC in vivo establishes a concurrent increase in both H(2)O(2) and caspase 8 activity during acute inflammation in living mice. Taken together, this method offers a potentially powerful new chemical tool for studying simultaneous oxidative stress and inflammation processes in living animals during injury, aging, and disease, as well as a versatile approach for concurrent monitoring of multiple analytes using luciferin-based bioluminescence imaging technologies.
Cluster computing applications, whether frameworks like MapReduce and Dryad, or customized applications like search platforms and social networks, have application-level requirements and higher-level abstractions to express them. Networking, however, still remains at the level of forwarding packets and balancing flows, and there exists no networking abstraction that can take advantage of the rich semantics readily available from these data parallel applications. The result is a plethora of seemingly disjoint, yet somehow connected, pieces of work to address networking challenges in these applications. We propose an application layer, data plane abstraction, coflow, that can express the requirements of (data) parallel programming models used in clusters today and makes it easier to express, reason about, and act upon these requirements.
FLUXNET and Remote Sensing Open Workshop: Towards Upscaling Flux Information From Towers to the Globe; Berkeley, California, 7–9 June 2011 About 140 scientists from 15 countries, representing the remote sensing and land‐atmosphere flux measurement and modeling communities, attended a workshop to address the technical, methodological, and scientific challenges to integrating flux and optical measurements across a spectrum of spatial and temporal scales.
An entry from the Cambridge Structural Database, the world’s repository for small molecule crystal structures. The entry contains experimental data from a crystal diffraction study. The deposited dataset for this entry is freely available from the CCDC and typically includes 3D coordinates, cell parameters, space group, experimental conditions and quality measures.
The field of dry deposition has had periods of ups and downs in activity and research.Unfortunately algorithms in important models have been fossilized to consider the Wesely model of 1989.While that was a very good and appropriate algorithm 30 years ago, we know more about land surface fluxes, how to model stomatal conductance and have been datasets and parameterization information in 2020.So, I was excited to see C1
Abstract Introduction Organizations and systems that deliver health care may better adapt to rapid change in their environments by acting as learning organizations and learning health systems (LHSs). Despite widespread recognition that multilevel forces shape capacity for learning within care delivery organizations, there is no agreed‐on, comprehensive, multilevel framework to inform LHS research and practice. Methods We develop such a framework, which can enhance both research on LHSs and practical steps toward their development. We draw on existing frameworks and research within organization and implementation science and synthesize contributions from three influential frameworks: the Consolidated Framework for Implementation Research, the social‐ecological framework, and the organizational change framework. These frameworks come, respectively, from the fields of implementation science, public health, and organization science. Results Our proposed integrative framework includes both intraorganizational levels (individual, team, mid‐management, organization) and the operating and general environments in which delivery organizations operate. We stress the importance of examining interactions among influential factors both within and across system levels and focus on the effects of leadership, incentives, and culture. Additionally, we indicate that organizational learning depends substantially on internal and cross‐level alignment of these factors. We illustrate the contribution of our multilevel perspective by applying it to the analysis of three diverse implementation initiatives that aimed at specific care improvements and enduring system learning. Conclusions The framework and perspective developed here can help investigators and practitioners broadly scan and then investigate forces influencing improvement and learning and may point to otherwise unnoticed interactions among influential factors. The framework can also be used as a planning tool by managers and practitioners.
Drawing on archival materials and personal testimonies, I reconstruct the conditions under which Bourdieu came to receive the Gold Medal of the National Center for Scientific Research, France’s highest science prize, in 1993 as a signal case study of the existential predicament and institutional trappings of scholarly consecration. Bourdieu’s award speech and the ceremony at which he read it present a triple interest for the history and sociology of sociology. They illustrate how a shaping figure in the discipline personally experienced, reflexively viewed, and practically navigated the nexus of science, authority, and power. They mark 1993 as a pivot-year in Bourdieu’s intellectual evolution, leading to a new agenda foregrounding the state as paramount symbolic power, the alchemy of group formation, and the unfinished promise of democratic politics; and they help explain why he ventured more forthrightly into civic debate in the 1990s. Bourdieu’s ambivalent acceptance of the prize also illustrates his co...
We investigated the influence of chloride on the secondary passive film (SPF) on laser directed energy deposition Alloy 718 during electrochemical machining. The results show that SPF formed in chloride-containing electrolyte is more defective than that formed in chloride-free solution, due to the stepped-up cation ejection by chloride. Chloride accelerates the SPF failure via enhanced cation vacancy condensation, SPF dissolution and possibly via surface vacancy pairs’ coalescence, restraining the formation of CrO 3 and inducing a better surface quality than does the chloride-free electrolyte. Based on the Point Defect Model, a mechanism describing the influence of chloride on SPF was developed.
