University of Grenoble

About: University of Grenoble is a education organization based out in Saint-Martin-d'Hères, France. It is known for research contribution in the topics: Population & Large Hadron Collider. The organization has 25658 authors who have published 45143 publications receiving 909760 citations.

The HARPS search for southern extra-solar planets: XVIII. An Earth-mass planet in the GJ 581 planetary system

Abstract: The GJ 581 planetary system was already known to harbour three planets, including two presumably rocky planets which straddle its habitable zone. We report here the detection of an additional planet ‐ GJ 581e ‐ with a minimum mass of 1.9 M" . With a period of 3.15 days, it is the innermost planet of the system and has a # 5% transit probability. We also correct our previous confusion of the orbital period of GJ 581d (the outermost planet) with a one-year alias, thanks to an extended time span and many more measurements. The revised period is 66.8 days, and locates the semi-major axis inside the habitable zone of the low mass star. The dynamical stability of the 4-planet system imposes an upper bound on the orbital plane inclination. The planets cannot be more massive than approximately 1.6 times their minimum mass.

Highly efficient and tunable spin-to-charge conversion through Rashba coupling at oxide interfaces

Abstract: The spin–orbit interaction couples the electrons’ motion to their spin. As a result, a charge current running through a material with strong spin–orbit coupling generates a transverse spin current (spin Hall effect, SHE) and vice versa (inverse spin Hall effect, ISHE). The emergence of SHE and ISHE as charge-to-spin interconversion mechanisms offers a variety of novel spintronic functionalities and devices, some of which do not require any ferromagnetic material. However, the interconversion efficiency of SHE and ISHE (spin Hall angle) is a bulk property that rarely exceeds ten percent, and does not take advantage of interfacial and low-dimensional effects otherwise ubiquitous in spintronic hetero- and mesostructures. Here, we make use of an interface-driven spin–orbit coupling mechanism—the Rashba effect—in the oxide two-dimensional electron system (2DES) LaAlO3/SrTiO3 to achieve spin-to-charge conversion with unprecedented efficiency. Through spin pumping, we inject a spin current from a NiFe film into the oxide 2DES and detect the resulting charge current, which can be strongly modulated by a gate voltage. We discuss the amplitude of the effect and its gate dependence on the basis of the electronic structure of the 2DES and highlight the importance of a long scattering time to achieve efficient spin-to-charge interconversion. The Rashba effect at the LaAlO3/SrTiO3 interface is shown to enable large and gate-tunable spin-to-charge conversion through the inverse Rashba–Edelstein effect.The spin current is injected, through spin pumping, from a NiFe film.

What do we gain from simplicity versus complexity in species distribution models

Abstract: Species distribution models (SDMs) are widely used to explain and predict species ranges and environmental niches. They are most commonly constructed by inferring species' occurrence-environment relationships using statistical and machine-learning methods. The variety of methods that can be used to construct SDMs (e.g. generalized linear/additive models, tree-based models, maximum entropy, etc.), and the variety of ways that such models can be implemented, permits substantial flexibility in SDM complexity. Building models with an appropriate amount of complexity for the study objectives is critical for robust inference. We characterize complexity as the shape of the inferred occurrence-environment relationships and the number of parameters used to describe them, and search for insights into whether additional complexity is informative or superfluous. By building 'under fit' models, having insufficient flexibility to describe observed occurrence-environment relationships, we risk misunderstanding the factors shaping species distributions. By building 'over fit' models, with excessive flexibility, we risk inadvertently ascribing pattern to noise or building opaque models. However, model selection can be challenging, especially when comparing models constructed under different modeling approaches. Here we argue for a more pragmatic approach: researchers should constrain the complexity of their models based on study objective, attributes of the data, and an understanding of how these interact with the underlying biological processes. We discuss guidelines for balancing under fitting with over fitting and consequently how complexity affects decisions made during model building. Although some generalities are possible, our discussion reflects differences in opinions that favor simpler versus more complex models. We conclude that combining insights from both simple and complex SDM building approaches best advances our knowledge of current and future species ranges.

The Replication Recipe: What Makes for a Convincing Replication?

Abstract: Psychological scientists have recently started to reconsider the importance of close replications in building a cumulative knowledge base; however, there is not a consensus about what constitutes a convincing replication study. To facilitate close and convincing replication attempts we have developed a Replication Recipe, outlining standard criteria for a convincing close replication. This includes faithfully recreating the original study while keeping track of differences, achieving high statistical power, checking the study’s assumptions in new contexts, and pre-registering the study. We also discuss methods for evaluating and reporting replications. Identifying differences between replication and original (sample, culture, lab context, etc.) allows researchers to identify where their replication is on the continuum from “close” to “conceptual”. Our replication recipe can be used by established researchers, teachers, and students to conduct meaningful replication studies and integrate replications into their scholarly habits.

Noisy Networks For Exploration

Abstract: We introduce NoisyNet, a deep reinforcement learning agent with parametric noise added to its weights, and show that the induced stochasticity of the agent's policy can be used to aid efficient exploration. The parameters of the noise are learned with gradient descent along with the remaining network weights. NoisyNet is straightforward to implement and adds little computational overhead. We find that replacing the conventional exploration heuristics for A3C, DQN and dueling agents (entropy reward and $\epsilon$-greedy respectively) with NoisyNet yields substantially higher scores for a wide range of Atari games, in some cases advancing the agent from sub to super-human performance.