Publications

We introduce Ego4D, a massive-scale egocentric video dataset and benchmark suite. It offers 3,670 hours of daily-life activity video spanning hundreds of scenarios (household, outdoor, workplace, leisure, etc.) captured by 931 unique camera wearers from 74 worldwide locations and 9 different countries. The approach to collection is designed to uphold rigorous privacy and ethics standards, with consenting participants and robust de-identification procedures where relevant. Ego4D dramatically expands the volume of diverse egocentric video footage publicly available to the research community. Portions of the video are accompanied by audio, 3D meshes of the environment, eye gaze, stereo, and/or synchronized videos from multiple egocentric cameras at the same event. Furthermore, we present a host of new benchmark challenges centered around understanding the first-person visual experience in the past (querying an episodic memory), present (analyzing hand-object manipulation, audio-visual conversation, and social interactions), and future (forecasting activities). By publicly sharing this massive annotated dataset and benchmark suite, we aim to push the frontier of first-person perception.

Summary This paper summarizes and analyses available data on the surface energy balance of Arctic tundra and boreal forest. The complex interactions between ecosystems and their surface energy balance are also examined, including climatically induced shifts in ecosystem type that might amplify or reduce the effects of potential climatic change. High latitudes are characterized by large annual changes in solar input. Albedo decreases strongly from winter, when the surface is snow‐covered, to summer, especially in nonforested regions such as Arctic tundra and boreal wetlands. Evapotranspiration ( Q E ) of high‐latitude ecosystems is less than from a freely evaporating surface and decreases late in the season, when soil moisture declines, indicating stomatal control over Q E , particularly in evergreen forests. Evergreen conifer forests have a canopy conductance half that of deciduous forests and consequently lower Q E and higher sensible heat flux ( Q H ). There is a broad overlap in energy partitioning between Arctic and boreal ecosystems, although Arctic ecosystems and light taiga generally have higher ground heat flux because there is less leaf and stem area to shade the ground surface, and the thermal gradient from the surface to permafrost is steeper. Permafrost creates a strong heat sink in summer that reduces surface temperature and therefore heat flux to the atmosphere. Loss of permafrost would therefore amplify climatic warming. If warming caused an increase in productivity and leaf area, or fire caused a shift from evergreen to deciduous forest, this would increase Q E and reduce Q H . Potential future shifts in vegetation would have varying climate feedbacks, with largest effects caused by shifts from boreal conifer to shrubland or deciduous forest (or vice versa) and from Arctic coastal to wet tundra. An increase of logging activity in the boreal forests appears to reduce Q E by roughly 50% with little change in Q H , while the ground heat flux is strongly enhanced.

The production rate and fractional catabolic rate of apoB and apoAI have been measured using stable isotope infusions of 13C or 2H leucine for many years and has led to our improved understanding of the pathophysiology of dyslipidemia and atherosclerosis. Modern tandem mass spectrometry methods have demonstrated the highly complex complement of proteins that associate with HDL and LDL and alterations in these proteins may impact lipoprotein function and cardiovascular risk. Very little is known about the turnover of the proteins that comprise the apolipoproteome. To study the fesability of measuring the turnover of multiple apolipoproteins associated with HDL and LDL we performed 20 hour IV infusions of U-13C labeled leucine in 20 healthy subjects. VLDL, LDL and HDL were isolated by ultracentrifugation and digested with trypsin. We analyzed peptides containing one leucine with analysis of multiple transitions using an Agilent triple quad. The enrichment of HDL and LDL peptides from multiple apolipoproteins were determined. The fractional synthesis of each protein was calculated by comparison of the average peptide enrichment for each protein to the average enrichment in VLDL apoB after 20 hours, at which time the VLDL apoB is fully turned over and its enrichment reflects the hepatic leucine precursor pool enrichment. Mean and SD values for HDL associated proteins are presented in table 1. The range of protein turnover rates on HDL confirms the dynamic state of the particle components. Fractional synthesis was similar for the majority of proteins on HDL and LDL with the notable exception of apoL1 and apoM. Using these methods, the regulation of the apolipoprotein composition of HDL and LDL can now be investigated.

<p>H & E staining of 22Rv1-DR xenograft tumor tissues from mice treated with Vehicle control, docetaxel alone, DSF/CuCl2, and docetaxel plus DSF/CuCl2.</p>

Deformation-induced nanoscale twinning is one of the mechanisms responsible for the excellent combination of strength and fracture toughness of the single-phase, face-centered cubic CrMnFeCoNi (Cantor) alloy, especially at cryogenic temperatures. Here, we use a novel, modified dogbone geometry that permits the sampling of varying stress and strain regions within a single tensile specimen to characterize the onset of twinning in CrMnFeCoNi at 293 K, 198 K and 77 K. Electron backscatter diffraction (EBSD) and backscattered electron (BSE) imaging revealed the presence of deformation nano-twins in regions of the samples that had experienced plastic strains of ∼25% at 293 K, ∼16% at 198 K, and ∼8% at 77 K, which are similar to the threshold strains described by Laplanche et al. (Acta Mater. 118, 2016, 152–163). From these strains we estimate that the critical tensile stress for the onset of twinning in this alloy is on the order of 750 MPa.

The sharing of scarce resources among multiple rational agents is one of the classical problems in economics. In exchange economies, which are used to model such situations, agents begin with an initial endowment of resources and exchange them in a way that is mutually beneficial until they reach a competitive equilibrium (CE). CE allocations are Pareto efficient and fair. Consequently, they are used widely in designing mechanisms for fair division. However, computing CEs requires the knowledge of agent preferences which are unknown in several applications of interest. In this work, we explore a new online learning mechanism, which, on each round, allocates resources to the agents and collects stochastic feedback on their experience in using that allocation. Its goal is to learn the agent utilities via this feedback and imitate the allocations at a CE in the long run. We quantify CE behavior via two losses and propose a randomized algorithm which achieves $\bigOtilde(\sqrt{T})$ loss after $T$ rounds under both criteria. Empirically, we demonstrate the effectiveness of this mechanism through numerical simulations.