Introduction
1.1 Dark energy problem and Λ-CDM model
Before discussing possible modifications to Einstein’s General Relativity (GR), it is worth stressing which astonishing accomplishment GR is, both **in** terms **of** its profound theoretical foundations and **of** the huge variety **of** phenomena that it can describe. On the first hand, its description **of** a Lorentz invariant space-time couched **in** the language **of** differential geometry it is meaningful and elegant and remains unchanged after more than one century from its first formulation. On the other hand, GR has proven to be spectacularly successful [ 1 ] when tested against experiments and observations, which range from millimeter **scale** laboratory tests to Solar System tests, including also strong regime tests such as binary pulsars dynamics. Within the standard model, GR governs the expansion **of** the Uni- verse, the behavior **of** black holes, the propagation **of** gravitational waves, and **cosmological** **structure** **formation** from planets and stars to the galaxy clusters. Having such an outstanding theory **of** gravity, one may wonder why there is such a huge number **of** alternative theories **in** the literature, and why there are different experimental and observational projects to test GR. Despite its success, there are (at least) two major reasons why it is interesting to study possible modifications **of** GR : the first one is the lack **of** a widely accepted quantum field theory **of** gravity (QFTG). **In** fact, even if different proposals for a QFTG exist, none **of** them has proven to be completely satisfactory **in** reconciling GR to quantum field theory [ 2 ] . Secondly, most **of** our current results **in** cosmology are based on a huge extrapolation **of** our knowledge **of** gravity up to scales where GR has never been tested.

En savoir plus
207 En savoir plus

(Dated: September 22, 2018)
We study **structure** **formation** **in** K-mouflage cosmology whose main feature is the absence **of** screening effect on quasilinear scales. We show that the growth **of** **structure** at the linear level is affected by both a new time dependent Newton constant and a friction term which depend on the background evolution. These combine with the modified background evolution to change the growth rate by up to ten percent since z ∼ 2. At the one loop level, we find that the nonlinearities **of** the K-mouflage models are mostly due to the matter dynamics and that the scalar perturbations can be treated at tree level. We also study the spherical collapse **in** K-mouflage models and show that the critical density contrast deviates from its Λ-CDM value and that, as a result, the halo mass function is modified for **large** masses by an order one factor. Finally we consider the deviation **of** the matter spectrum from Λ-CDM on nonlinear scales where a halo model is utilized. We find that the discrepancy peaks around 1 hMpc −1 with a relative difference which can reach fifty percent. Importantly, these features are still true at larger redshifts, contrary to models **of** the chameleon- f (R) and Galileon types.

En savoir plus
Participating **in** **various** international conferences and summer schools has always been a fruitful and en- joyable experience. I want to thank the organizers **of** the ICTP summer school on cosmology and workshop on **large**-**scale** **structure** (2012), the Varenna summer school (2013), the Les Houches summer school (2013), the Rencontres de Moriond (2014, Cosmology session), the IAU symposia 306 and 308 **in** Lisbon and Tallinn (2014), the CCAPP workshop on cosmic voids (2014), COSMO 2014 **in** Chicago, the MPA-EXC workshop on the dynamic Universe (2014, Garching), the ICTP workshop on **cosmological** structures (2015), the ESO-MPA- EXC **large**-**scale** **structure** conference (2014, Garching), and the Rencontres du Vietnam **in** Quy Nhon (2015, Cosmology session). Best greetings, **in** particular, to the Les Houches’ cosmologists group; thanks also to the organizers **of** the student conferences I attended: Elbereth 2012, 2013, 2014, and the SCGSC 2013. On **various** occasions, I have had the chance to have friendly and interesting discussions (even if sometimes short) within the cosmology community. **In** particular, my work beneﬁted from interactions with Niayesh Afshordi, Raul Angulo, Stephen Bailey, Robert Cahn, Olivier Doré, Torsten Enßlin, Luigi Guzzo, Oliver Hahn, Jean-Christophe Hamil- ton, Alan Heavens, Shirley Ho, Mike Hudson, Eiichiro Komatsu, Ofer Lahav, Mark Neyrinck, Nelson Padilla, Bruce Partridge, Will Percival, David Schlegel, Uroš Seljak, Sergei Shandarin, Ravi Sheth, Svetlin Tassev, and Rien van de Weygaert (among many others). At this point, it also seems needed to acknowledge the decisive contribution **of** a familiar ∼ 20 Mpc/h void (at coordinates x ≈ −100, y ≈ 200 **in** the slice that I usually show), which very nicely makes my point during presentations.

En savoir plus
238 En savoir plus

different species X, C[f X ]. For the interaction between photons and leptons, we consider the classical Thomson scattering non-relativistic approach, l ∓ + γ ↔ l ∓ + γ with an interaction rate Γ ' n l σ T , where σ T ' 2 × 10 −3 M eV −2 is the Thomson cross section. For cold dark matter, we consider a collisionless non-relativistic approach, as done **in** **various** famous **structure** **formation** history models. This are the simplest models that agree with observational **large** **scale** **structure** data. For baryons and leptons interactions, we assume a Coulomb Scattering, b ± + l ∓ ↔ b ± + l ∓ **in** the Quantum ElectroDynamic (QED) approach. While for neutrini, we only consider them as a massless relativistic particle fluctuation overdensity and therefore we assume that they do not interact with matter. This is true only **in** the linear regime at **large** scales. Adopting a Fourier transform framework to simplify the equations **in** question, we end up with a set **of** 6 linear differential equations describing the non linear evolution **of** the 3 different species **of** density fluctuations (baryons, photons and neutrinos and Dark Matter) and their corresponding velocities at **large** **scale** as a function **of** conformal time 15 , η, and wavenumber, ~k. However, this system is coupled to the 2 degrees, Φ(η) & Ψ(~k), **of** freedom defined by the perturbations **of** the curved metric. Thus, **in** order to completely specify the system one may solve the time-time component and the spatial trace **of** the Einstein equations using the perturbed metric defined via Eq. 1.22 . Thus we end up with the coupled Boltzmann-Einstein equations that completely specify the system on **large** **scale** structures, i.e. the evolution **of** the density and temperature fluctuations,

En savoir plus
214 En savoir plus

Sloan, 2dF and VIRMOS surveys. This gives the best possible mapping **of** structures traced by galaxies, together with strong constraints on models for **structure** evolution. Unfortunately, it is extremely data-intensive. Moreover, the results depend on both the global **cosmological** parameters and the de- tails **of** galaxy **formation**. Breaking the degeneracy between these two factors is nontrivial. The study **of** **structure** us- ing only clusters **of** galaxies can offer significant advantages both because it is easier to define a complete sample **of** objects over a very **large** volume **of** space and because the objects them- selves are **in** some respects “simpler” to understand (at least **in** terms **of** their **formation** and evolution). Consequent- ly, with currently available observation- al resources, larger volumes **of** the uni- verse can be studied to substantially greater depth, and the interpretation **of** the results is less dependent on models **of** how galaxies form. Such studies can independently check **cosmological** pa- rameter values determined from the CMB and SN studies, can break the de- generacy between the shape **of** the power spectrum and the matter density, and can check other fundamental as- sumptions **of** the standard paradigm (e.g. that the initial fluctuations were gaussian). Unfortunately, clusters **of** galaxies become increasingly difficult to identify optically with increasing dis- tance because their contrast against foreground and background galaxies is strongly reduced. This has greatly ham- pered investigations **of** high-redshift op- tically selected clusters.

En savoir plus
The experiments on neutrino flavor oscillations demonstrating that neutrinos are indeed massive are thus **of** crucial importance and it is necessary to examine minutely the impact **of** those masses on **various** **cosmological** observables. Understandably, such a discovery has triggered a considerable e↵ort **in** theoretical, numerical and observational cosmology to infer the consequences on the cosmic **structure** growth. The first study **in** which massive neutrinos are properly treated **in** the linear theory **of** gravitational perturbations dates back from ref. [ 3 ] (see also its companion paper ref. [ 4 ]). The consequences **of** these results are thoroughly presented **in** ref. [ 5 ], where the connection between neutrino masses and cosmology - **in** the standard case **of** three neutrino species - is investigated **in** full detail. It is shown that CMB anisotropies are indirectly sensitive to massive neutrinos whereas the late-time **large**-**scale** **structure** growth rate, via its time and **scale** dependences, o↵ers a much more direct probe **of** the neutrino mass spectrum. To a **large** extent current and future cosmology projects aim at exploiting these dependences to put constraints on the neutrino masses. Indeed, the impact **of** massive neutrinos on the **structure** growth has proved to be significative enough to make such constraints possible, as shown for instance **in** [ 6 – 11 ]. These physical interpretations are based on numerical experiments, the early incarnations dating back from the work **of** ref. [ 12 ], which have witnessed a renewed interest **in** the last years [ 13 – 16 ], and also on theoretical investigations such as [ 17 – 20 ], where the e↵ect **of** massive neutrinos **in** the non- linear regime is investigated with the help **of** Perturbation Theory. An important point is that it is potentially possible to get better constraints than what the predictions **of** linear theory o↵er. Observations **of** the **large**-**scale** **structure** within the local universe are indeed sensitive to the non-linear growth **of** **structure** and thus also to the impact **of** mode-coupling e↵ects on this growth. Such a coupling is expected to strengthen the role played by the matter and energy content **of** the universe on **cosmological** perturbation properties. This is true for instance for the dark energy equation **of** state [ 21 ] or for the masses, even if small, **of** the neutrino species, as shown **in** numerical experiments [ 14 ].

En savoir plus
240 En savoir plus

the one **of** Λ-CDM. At the perturbative level, and first **in** the linear theories, deviations from GR occur on scales lower than the Compton wavelength **of** the scalar field [29]. As Solar System tests and the screening **of** the Milky Way imply that the **cosmological** range **of** the scalar must be less than 1 Mpc [30], the effects **of** these models on linear scales are suppressed and only **in** the quasilinear to mildly nonlinear regimes one can expect to see signif- icant deviations. Symmetrons and dilatons screen grav- ity **in** a stronger way **in** the local environment imply- ing that constraints on these models are less severe than on chameleon-f (R) theories. This implies that the ef- fects **of** the symmetron and to a lesser extent **of** the dila- ton on **large**-**scale** structures are enhanced compared to chameleon-f (R) models. Typically, one expects to see a peak **in** the deviations from GR on the scales correspond- ing to the range **of** the scalar field, especially **in** the power spectrum **of** density fluctuations [9]. On small and **large** scales, the models converge towards GR. On small scales, this is due to the screening effect and on **large** scales this is also the screening property outside the Compton ra- dius.

En savoir plus
pulicani@lirmm.fr
Abstract. Genome architecture can be drastically modified through a succession **of** **large**-**scale** rearrangements. **In** the quest to infer these re- arrangement **scenarios**, it is often the case that the parsimony principal alone does not impose enough constraints. **In** this paper we make an initial effort towards computing **scenarios** that respect chromosome con- **formation**, by using Hi-C data to guide our computations. We confirm the validity **of** a model – along with optimization problems Minimum Local Scenario and Minimum Local Parsimonious Scenario – de- veloped **in** previous work that is based on a partition into equivalence classes **of** the adjacencies between syntenic blocks. To accomplish this we show that the quality **of** a clustering **of** the adjacencies based on Hi-C data is directly correlated to the quality **of** a rearrangement sce- nario that we compute between Drosophila melanogaster and Drosophila yakuba. We evaluate a simple greedy strategy to choose the next rear- rangement based on Hi-C, and motivate the study **of** the solution space **of** Minimum Local Parsimonious Scenario.

En savoir plus
The issue **of** “landscape legacies” (Wallin et al. 1994) is highlighted by our findings. The cumulative impacts **of** past disturbance and management (logging, fire suppression, etc.) have resulted **in** the present age-class **structure**. This legacy may pose challenges and (or) opportunities for management objectives (Östlund et al. 1997). **In** all **of** the AC **scenarios**, there is a time lag **of** over a century before the targeted AC **structure** is reached, as there is a **large** discrepancy between the initial age-class **structure** and any **of** the target distribu- tions. This long time lag required to shape the age-class **structure** implies a need for proactive management, since it has significant consequences for the stand ages and spatial pattern **of** harvest and may lead to a potential conflict be- tween the harvest flow and target AC **structure** objectives. Given the uncertainty **of** changes **in** climate, economies, and social values over such a long time frame (Kaufmann et al. 1994; Chapin and Whiteman 1998), the focus should be on the transition period and how the current forest state can be shaped into a desired condition. If short-term costs (eco- nomic or ecological) are too significant, a plan is unlikely to be acceptable regardless **of** the long-term benefits. **In** the study area, the period 50 years **in** the future is most critical to conservation objectives, since all **scenarios** that do not specify hard AC targets pass through a phase **of** very little old forest on the landscape as the current old forest is de- pleted before the young crop ages into older classes.

En savoir plus
where star **formation** is most efficient. For higher and lower halo masses, the star **formation** rates are reduced due to feedback processes. Modern **large** volume simulations reproduce the stellar to halo mass relationship at low and high redshifts reasonably well 124 , 309 .
Gas around galaxies: One **of** the key advantages **of** hydrodynamical simulations compared to semi- analytic models (see Box 1) is their ability to make detailed predictions for the distribution and properties **of** gas around galaxies including the circumgalactic medium, the intracluster medium, and the intergalac- tic medium. The circumgalactic and intergalactic media are quite diffuse (n ∼ 10 −3 − 10 −7 cm −3 ) and cool (T ∼ 10 4−6 K) and observations **in** emission, like Lyman-α and metal lines, are therefore rather challenging. However, absorption line observations from background quasars can probe the distribution, enrichment, and ionization state **of** this gas. One **of** the first successes **of** hydrodynamical simulations has been the reproduction **of** the declining trend **of** the number **of** absorbing clouds per unit redshift and linear interval **of** H I column density with column density **in** the Lyman-α forest 288 . Reproducing properties **of** the circumgalactic medium, however, is significantly more challenging. Observations **of** this gas indicate that it features a rich multi-phase **structure** where individual lines **of** sight simultaneously contain highly ionized, warm, and cool atomic species 310 , 311 . The coolest and densest parts **of** this gas have spatial scales **of** 10 − 100 pc 312 , although the coherence **scale** can reach up to ∼ 1 kpc 313 . These spatial scales are below the typical circumgalactic gas resolution limits **of** galaxy **formation** simulations. More recently, **cosmological** simulations with special circumgalactic gas refinement schemes have been employed to overcome some **of** the resolution limitations. Such simulations increase the numerical resolution **in** the circumgalactic gas reaching smaller spatial scales 314 – 317 . At z = 2 such simulations can reach a spatial resolution below ∼ 100 pc 316 , and at z = 0 below ∼ 1 kpc within the circumgalactic medium 315 . **In** addition to resolution

En savoir plus
~0.5 º C to its freezing point.
• No other external seeding was required to initiate the generation **of** frazil ice
• When supercooling started, frazil production increased rapidly and **large** amounts **of** crystals were produced over a short period **of** time

= γ(σ 0 ). (A.6)
The formulation (A.4)–(A.6) is however merely formal, as the optimization problem (A.4) is **in** this case ill-defined. The trouble comes from the poloidal degrees **of** freedom being **in** a sense not constrained enough by the macro state constraints **of** Equation (A.2). The problem is apparent **in** the definition **of** the partition function (A.5) : the integral R ξ there involved does not **in** general converge. We think that this behavior is an avatar **of** the UV catastrophe encountered **in** the statistical theories **of** ideal 3D flows. A phenomenological taming **of** the problem can be achieved by further constraining the set **of** macro state fields over which the optimization problem (A.4) is solved. This requires the use **of** additional ansatz, some **of** which we below discuss. **In** order to carry out some explicit calculations and retrieve the equations **of** Table (2), we will consider two simplified sets **of** macro-constraints (A.2). (i) **In** the Two-Level modeling, we prescribe the toroidal field to be a two-level, symmetric distribution, viz., α(σ) = pδ(σ − 1) + (1 − p)δ(σ + 1). Only five constraints then remain from the set **of** constraints (A.2) : the energy, two toroidal areas A ± , and two partial circulations Γ ± .

En savoir plus
Other applications are **of** paramount importance **in** our society. The healthcare industry has strongly benefited from statistical algorithms. The diagnosis **of** **various** diseases can be auto- mated or assisted. The review **in** Kononenko ( 2001 ) mentions several statistical methods such as naive and semi-naive Bayesian classifiers, k-nearest neighbors, neural networks or decision trees. The article Dreiseitl et al. ( 2001 ) compares algorithms such as logistic regression, arti- ficial neural networks, decision trees, and support vector machines on the task **of** diagnosing pigmented skin lesions, **in** order to distinguish common nevi from dysplastic nevi or melanoma. The authors **of** Shipp et al. ( 2002 ) describe a method to ease the detection **of** blood cancer. The discovery **of** new drugs **in** the pharmaceutical industry is a growing challenge. The more active compounds are discovered, the less likely it is to discover a new drug with positive impact. **In** order to keep innovating, the pharmaceutical industry needs to increase the capacity **of** screening active compounds. Standard high-throughput screening methods become more and more costly as the number **of** active compounds already tested increases. As **in** ranking

En savoir plus
179 En savoir plus

3
1. Introduction
Numerous trace metals are released to rivers, coastal areas, and ultimately the oceans by human activities, and their behaviors **in** many marine taxa such as bivalves, cephalopods, crustaceans, and fish have been thoroughly investigated **in** past decades (e.g. Bustamante et al., 2003; Eisler, 2010; Metian et al., 2013). However, lithium (Li) concentrations **in** marine organisms have received little attention, despite the exponentially increasing use **of** this element **in** high- tech industries due to its unique physicochemical properties. Li **in** the oceans is dominantly derived from two natural sources, i.e. high-temperature hydrothermal fluxes at mid-ocean ridges and river inputs. It exits the ocean mainly via the **formation** **of** marine authigenic aluminosilicate clays on the seafloor (Chan et al., 2006). Due to its long oceanic residence time (~1.2 million years) and its weak capacity to adsorb onto marine particles (Decarreau et al., 2012), Li is homogeneously distributed throughout the water column (Misra and Froelich, 2012). Thus, the oceanic concentration **of** dissolved Li is constant at 0.183 ± 0.003 µg/mL, irrespective **of** latitude and depth (Aral and Vecchio-Sadus, 2011; Misra and Froelich, 2012; Riley and Tengudai, 1964).

En savoir plus
Quantum Monte Carlo (QMC) methods are known to be powerful stochastic approaches for solving the Schr¨odinger equation.[1] Although they have been widely used **in** computational physics during the last twenty years, they are still **of** marginal use **in** computational chemistry.[2] Two major reasons can be **in**- voked for that: i) the N -body problem encountered **in** chemistry is particularly challenging (a set **of** strongly interacting electrons **in** the field **of** highly-attractive nuclei) ii.) the level **of** numerical accuracy required is very high (the so-called “chemical accuracy”). **In** computational chemistry, the two standard approaches used presently are the Density Functional Theory (DFT) approaches and the **various** post-Hartree-Fock wavefunction-based methods (Configuration Interac- tion, Coupled Cluster, etc.) **In** practice, DFT methods are the most popular approaches, essentially because they combine both a reasonable accuracy and a favorable scaling **of** the computational effort as a function **of** the number **of** electrons. On the other hand, post-HF methods are also employed since they lead to a greater and much controlled accuracy than DFT. Unfortunately, the price to pay for such an accuracy is too high to be **of** practical use for **large** molecular systems.

En savoir plus
telescope mount. However, instead **of** a telescope there is a special projector, which sends out a narrow, well-defined light beam. The radio
telescope is equipped with detectors to sense this light beam, and to follow it. It does not matter if there are irregularities **in** the gears; the telescope just follows the light beam. **In** the days when computers were expensive, this system worked well without needing one. However, there was one area where a computer would have been helpful; helping the telescope to find the light beam. Maintenance and other needs often result **in** the Master Equatorial pointing **in** one direction and the telescope pointing **in** another. To start making observations the telescope has to be slaved to the light beam. This was accomplished using a device known as the Coordinate Converter, or Co-Co. This unit, a strange mixture **of** motors, synchros, gears and other devices, converts the antenna position information into the same units as those used by the Master Equatorial. The Telescope Operator, sitting **in** the Control Room, can see displays **of** the telescope and Master Equatorial coordinates, and then drive the telescope until it “sees the light” and locks **in**.

En savoir plus
Ideally, the relative di fference between observation and simulation dR=0, though **in** practice we expect some scatter, caused by inaccuracies **of** the RTM input. Biases are 20
either caused by the calibration **of** the measurement, or by systematic deviations **of** the input data. The probability **of** the latter has been minimised by the selection **of** the spectral range **of** the study, where the sensitivity to most input parameters is small and by a very strict cloud mask. The only ingredient **of** the simulation that is di fficult to filter is the ozone profile. A deviation **in** the ozone profile, however, gives a clear spectral 25

En savoir plus
2
Universit´ e de Nice Sophia Antipolis, CNRS, LJAD, UMR 7351, 06100 Nice, France (Dated: September 14, 2015)
We study the motion **of** small inertial particles **in** stratified turbulence. We derive a simplified model, valid within the Boussinesq approximation, for the dynamics **of** small particles **in** presence **of** a mean linear density profile. By means **of** extensive direct numerical simulations, we investigate the statistical distribution **of** particles as a function **of** the two dimensionless parameters **of** the problem. We find that vertical confinement **of** particles is mainly ruled by the degree **of** stratification, with a weak dependency on the particle properties. Conversely, small **scale** fractal clustering, typical **of** inertial particles **in** turbulence, depends on the particle relaxation time and is almost independent on the flow stratification. The implications **of** our findings for the **formation** **of** thin phytoplankton layers are discussed.

En savoir plus
3.1.5 Preferences systems
Actions cannot be compared one by one because **of** their generic definition. To accomplish this comparison, decision-
makers, or the analyst judging by their names, must develop a relational preference system. This system reflects diverse views that can be opposed, or even contradictory. Thus, the system must tolerate ambiguity, contradiction, and learning wherever possible (Roy 1985 ). Preference systems are also called ‘‘approach and the dominant culture’’ (Merad 2010 ). They are set **of** beliefs, attitudes and assumptions shared by a group as a result **of** past experiences (Merad 2010 ). We have determined the preference system for decision-makers **in** Table 2 .

En savoir plus
Flexibility
Flexibility has not been studied to the same extent as complexity, and has not developed as varied a set **of** definitions. For us a system is flexible if one can implement classes **of** changes **in** its specifications with relative ease. Note that we do not claim that even small changes **in** specifications always result **in** ease **of** implementation, just that the system is designed to handle certain classes **of** changes **in** an easy manner. How does this translate to system implementation? Again we do not yet know the general answer. However, if we consider the paths **of** interconnections **in** a system, a system with many alternate paths from the inputs to the outputs will usually be able to handle certain changes **in** specifications relatively well. Moreover, a system with many alternate paths can increase the number **of** paths a great deal with a small increase **in** the number **of** interconnections and additions **of** new nodes or modifications **of** existing ones. We shall therefore define the flexibility **of** a system as the number **of** paths **in** it, counting loops just once. **In** some architectures, notably the network **structure** and the layered hierarchy, the number **of** potential paths that keeps the generic structural form unchanged is very high. This, however, cannot be said **of** the tree structured hierarchy.

En savoir plus