Search

Contains fulltext : 253276.pdf (Publisher’s version ) (Open Access), Computational models of visual attention in artificial intelligence and robotics have been inspired by the concept of a saliency map. These models account for the mutual information between the (current) visual information and its estimated causes. However, they fail to consider the circular causality between perception and action. In other words, they do not consider where to sample next, given current beliefs. Here, we reclaim salience as an active inference process that relies on two basic principles: uncertainty minimization and rhythmic scheduling. For this, we make a distinction between attention and salience. Briefly, we associate attention with precision control, i.e., the confidence with which beliefs can be updated given sampled sensory data, and salience with uncertainty minimization that underwrites the selection of future sensory data. Using this, we propose a new account of attention based on rhythmic precision-modulation and discuss its potential in robotics, providing numerical experiments that showcase its advantages for state and noise estimation, system identification and action selection for informative path planning.

Impacts of climate change require that society urgently develops ways to reduce amounts of carbon in the atmosphere. Tropical forests present an important opportunity, as they take up and store large amounts of carbon. It is often suggested that forests with high biodiversity have large stocks and high rates of carbon uptake. Evidence is, however, scattered across geographic areas and scales, and it remains unclear whether biodiversity is just a co‐benefit or also a requirement for the maintenance of carbon stocks and uptake. Here, we perform a quantitative review of empirical studies that analyzed the relationships between plant biodiversity attributes and carbon stocks and carbon uptake in tropical forests. Our results show that biodiversity attributes related to species, traits or structure significantly affect carbon stocks or uptake in 64% of the evaluated relationships. Average vegetation attributes (community‐mean traits and structural attributes) are more important for carbon stocks, whereas variability in vegetation attributes (i.e., taxonomic diversity) is important for both carbon stocks and uptake. Thus, different attributes of biodiversity have complementary effects on carbon stocks and uptake. These biodiversity effects tend to be more often significant in mature forests at broad spatial scales than in disturbed forests at local spatial scales. Biodiversity effects are also more often significant when confounding variables are not included in the analyses, highlighting the importance of performing a comprehensive analysis that adequately accounts for environmental drivers. In summary, biodiversity is not only a co‐benefit, but also a requirement for short‐ and long‐term maintenance of carbon stocks and enhancement of uptake. Climate change policies should therefore include the maintenance of multiple attributes of biodiversity as an essential requirement to achieve long‐term climate change mitigation goals.

This paper considers functional integration in the brain from a computational perspective. We ask what sort of neuronal message passing is mandated by active inference—and what implications this has for context-sensitive connectivity at microscopic and macroscopic levels. In particular, we formulate neuronal processing as belief propagation under deep generative models. Crucially, these models can entertain both discrete and continuous states, leading to distinct schemes for belief updating that play out on the same (neuronal) architecture. Technically, we use Forney (normal) factor graphs to elucidate the requisite message passing in terms of its form and scheduling. To accommodate mixed generative models (of discrete and continuous states), one also has to consider link nodes or factors that enable discrete and continuous representations to talk to each other. When mapping the implicit computational architecture onto neuronal connectivity, several interesting features emerge. For example, Bayesian model averaging and comparison, which link discrete and continuous states, may be implemented in thalamocortical loops. These and other considerations speak to a computational connectome that is inherently state dependent and self-organizing in ways that yield to a principled (variational) account. We conclude with simulations of reading that illustrate the implicit neuronal message passing, with a special focus on how discrete (semantic) representations inform, and are informed by, continuous (visual) sampling of the sensorium.

Impacts of climate change require that society urgently develops ways to reduce amounts of carbon in the atmosphere. Tropical forests present an important opportunity, as they take up and store large amounts of carbon. It is often suggested that forests with high biodiversity have large stocks and high rates of carbon uptake. Evidence is, however, scattered across geographic areas and scales, and it remains unclear whether biodiversity is just a co-benefit or also a requirement for the maintenance of carbon stocks and uptake. Here, we perform a quantitative review of empirical studies that analyzed the relationships between plant biodiversity attributes and carbon stocks and carbon uptake in tropical forests. Our results show that biodiversity attributes related to species, traits or structure significantly affect carbon stocks or uptake in 64% of the evaluated relationships. Average vegetation attributes (community-mean traits and structural attributes) are more important for carbon stocks, whereas variability in vegetation attributes (i.e., taxonomic diversity) is important for both carbon stocks and uptake. Thus, different attributes of biodiversity have complementary effects on carbon stocks and uptake. These biodiversity effects tend to be more often significant in mature forests at broad spatial scales than in disturbed forests at local spatial scales. Biodiversity effects are also more often significant when confounding variables are not included in the analyses, highlighting the importance of performing a comprehensive analysis that adequately accounts for environmental drivers. In summary, biodiversity is not only a co-benefit, but also a requirement for short- and long-term maintenance of carbon stocks and enhancement of uptake. Climate change policies should therefore include the maintenance of multiple attributes of biodiversity as an essential requirement to achieve long-term climate change mitigation goals.

There are many declarative frameworks that allow us to implement code formatters relatively easily for any specific language, but constructing them is cumbersome. The first problem is that "everybody" wants to format their code differently, leading to either many formatter variants or a ridiculous number of configuration options. Second, the size of each implementation scales with a language's grammar size, leading to hundreds of rules. In this paper, we solve the formatter construction problem using a novel approach, one that automatically derives formatters for any given language without intervention from a language expert.We introduce a code formatter called CODEBUFF that uses machine learning to abstract formatting rules from a representative corpus, using a carefully designed feature set. Our experiments on Java, SQL, and ANTLR grammars show that CODEBUFF is efficient, has excellent accuracy, and is grammar invariant for a given language. It also generalizes to a 4th language tested during manuscript preparation.

Temperature exerts a fundamental control on ecosystem function, species’ distributions and ecological processes across a range of spatial scales. At the landscape scale, near-surface air temperature may vary spatially over short distances, particularly inmountainous regions. Both the magnitude and spatial pattern of surface temperature may vary on diurnal, seasonal and inter-annual timescales. Furthermore, temperatures measured at the surface of vegetation, influenced by the energy balance of the surface, can differ considerably from air temperature. In order to explore spatial patterns in temperature across the MoorHouse sector of the MoorHouse—Upper Teesdale National Nature Reserve (NNR), Northern Pennines, UK,we derived anempirical linear regression model to predict airtemperature at 1 mheight as a function of landscape metrics derived from a digital elevation model (DTM), and coupled this to an existing physical land-surface model (JULES) in order to predict and map thermal climate at the vegetation surface across the study area. Spatial patterns in temperature associated with altitudinal lapse rate, katabatic flow and a local fohn effect were incorporated into the regression model. JULES was driven using spatially distributed air temperatures from the empirical model, along with distributed solar and long-wave radiation flux estimates adjusted for surface slope and aspect, and sky-view in order to model the surface energy balance and predict thermal climate at the vegetation surface (skin temperature). Aggregate properties such as annual degree days above 5 degC (GDD5), number of ‘‘frost days’’ when the temperature fell below 0 degC (FD0) and number of ‘‘severe frost days’’ when the minimum temperature fell below 5degC (FD5) were mapped across the reserve for the years 1994–2006. Spatial mapping of surface temperature revealed differences in the 12-year average spatial pattern between GDD5, FD0 and FD5, and differences in the spatial patterns of FD0 and FD5 bet

A closed-loop liquid-fueld active control technique was applied in a dump combustor to enhance its combustion performance. Practical issues involving liquid fuel scaling effects and affordable control methods were studied experimentally. The results shed new light on the requirement of critical fuel flux effects of fuel droplet size on control and novel controller concepts that would help the future development work. The critical fuel flux was found to be dependent on the fuel droplet size and initial magnitude of the instabilities. When the fuel droplet size Do was reduced in the controlled injection the control efficiency for heat flux actuation increased significantly. The analysis yielded an exponential dependency on the droplet size. For a moderate droplet Reynolds number considered in this study the amplitude of controlled heat release for a given fuel amount was inversely proportional to the droplet size by a factor of Do(-1.4). Also two novel controller concepts which incorporated practical designs were tested and were shown to work effectively compared to the baseline case. One of the concepts was based on injecting fuel pulses at sub-harmonic frequencies of the instability thus addressing the limited actuator frequency response. The other concept utilized both open-loop and closed-loop control schemes to obtain enhanced performance including extension of the stable combustion zone. These results open up the possibility of applying the active combustion control technology to advanced propulsion devices., Pres: Symposium RTO Applied Vehicle Tech Panel (AVT), Braunschweig, Germany, p36-1/36-18,8-11 May 2000. This article is from ADA395700 Active Control Technology for Enhanced Performance Operational Capabilities of Military Aircraft, Land Vehicles and Sea Vehicles (Technologies des systemes a commandes actives pour l'amelioration des performances operationnelles des aeronefs militaires, des vehicules terrestres et des vehicules maritimes)