@article{3697, abstract = {The goal of this paper is to evaluate and compare models and methods for learning to recognize basic entities in images in an unsupervised setting. In other words, we want to discover the objects present in the images by analyzing unlabeled data and searching for re-occurring patterns. We experiment with various baseline methods, methods based on latent variable models, as well as spectral clustering methods. The results are presented and compared both on subsets of Caltech256 and MSRC2, data sets that are larger and more challenging and that include more object classes than what has previously been reported in the literature. A rigorous framework for evaluating unsupervised object discovery methods is proposed.}, author = {Tuytelaars,Tinne and Christoph Lampert and Blaschko,Matthew B and Buntine,Wray}, journal = {International Journal of Computer Vision}, number = {2}, pages = {284 -- 302}, publisher = {Springer}, title = {{Unsupervised object discovery: A comparison}}, doi = {10.1007/s11263-009-0271-8}, volume = {88}, year = {2010}, } @inproceedings{3702, abstract = {Hitting and batting tasks, such as tennis forehands, ping-pong strokes, or baseball batting, depend on predictions where the ball can be intercepted and how it can properly be returned to the opponent. These predictions get more accurate over time, hence the behaviors need to be continuously modified. As a result, movement templates with a learned global shape need to be adapted during the execution so that the racket reaches a target position and velocity that will return the ball over to the other side of the net or court. It requires altering learned movements to hit a varying target with the necessary velocity at a specific instant in time. Such a task cannot be incorporated straightforwardly in most movement representations suitable for learning. For example, the standard formulation of the dynamical system based motor primitives (introduced by Ijspeert et al. [1]) does not satisfy this property despite their flexibility which has allowed learning tasks ranging from locomotion to kendama. In order to fulfill this requirement, we reformulate the Ijspeert framework to incorporate the possibility of specifying a desired hitting point and a desired hitting velocity while maintaining all advantages of the original formulation. We show that the proposed movement template formulation works well in two scenarios, i.e., for hitting a ball on a string with a table tennis racket at a specified velocity and for returning balls launched by a ball gun successfully over the net using forehand movements. All experiments were carried out on a Barrett WAM using a four camera vision system.}, author = {Kober,Jens and Mülling,Katharina and Krömer,Oliver and Christoph Lampert and Schölkopf,Bernhard and Peters, Jan}, pages = {853 -- 858}, publisher = {IEEE}, title = {{Movement templates for learning of hitting and batting}}, doi = {10.1109/ROBOT.2010.5509672}, year = {2010}, } @inproceedings{3713, abstract = {We introduce a method to accelerate the evaluation of object detection cascades with the help of a divide-and-conquer procedure in the space of candidate regions. Compared to the exhaustive procedure that thus far is the state-of-the-art for cascade evaluation, the proposed method requires fewer evaluations of the classifier functions, thereby speeding up the search. Furthermore, we show how the recently developed efficient subwindow search (ESS) procedure [11] can be integrated into the last stage of our method. This allows us to use our method to act not only as a faster procedure for cascade evaluation, but also as a tool to perform efficient branch-and-bound object detection with nonlinear quality functions, in particular kernelized support vector machines. Experiments on the PASCAL VOC 2006 dataset show an acceleration of more than 50% by our method compared to standard cascade evaluation.}, author = {Christoph Lampert}, pages = {1022 -- 1029}, publisher = {IEEE}, title = {{An efficient divide-and-conquer cascade for nonlinear object detection}}, doi = {10.1109/CVPR.2010.5540107}, year = {2010}, } @article{3718, abstract = {Long-term depression (LTD) is a form of synaptic plasticity that may contribute to information storage in the central nervous system. Here we report that LTD can be elicited in layer 5 pyramidal neurons of the rat prefrontal cortex by pairing low frequency stimulation with a modest postsynaptic depolarization. The induction of LTD required the activation of both metabotropic glutamate receptors of the mGlu1 subtype and voltage-sensitive Ca(2+) channels (VSCCs) of the T/R, P/Q and N types, leading to the stimulation of intracellular inositol trisphosphate (IP3) receptors by IP3 and Ca(2+). The subsequent release of Ca(2+) from intracellular stores activated the protein phosphatase cascade involving calcineurin and protein phosphatase 1. The activation of purinergic P2Y(1) receptors blocked LTD. This effect was prevented by P2Y(1) receptor antagonists and was absent in mice lacking P2Y(1) but not P2Y(2) receptors. We also found that activation of P2Y(1) receptors inhibits Ca(2+) transients via VSCCs in the apical dendrites and spines of pyramidal neurons. In addition, we show that the release of ATP under hypoxia is able to inhibit LTD by acting on postsynaptic P2Y(1) receptors. In conclusion, these data suggest that the reduction of Ca(2+) influx via VSCCs caused by the activation of P2Y(1) receptors by ATP is the possible mechanism for the inhibition of LTD in prefrontal cortex.}, author = {Guzmán, José and Schmidt, Hartmut and Franke, Heike and Krügel, Ute and Eilers, Jens and Illes, Peter and Gerevich, Zoltan}, journal = {Neuropharmacology}, number = {6}, pages = {406 -- 415}, publisher = {Elsevier}, title = {{P2Y1 receptors inhibit long-term depression in the prefrontal cortex.}}, doi = {10.1016/j.neuropharm.2010.05.013}, volume = {59}, year = {2010}, } @inproceedings{3719, abstract = {The induction of a signaling pathway is characterized by transient complex formation and mutual posttranslational modification of proteins. To faithfully capture this combinatorial process in a math- ematical model is an important challenge in systems biology. Exploiting the limited context on which most binding and modification events are conditioned, attempts have been made to reduce the com- binatorial complexity by quotienting the reachable set of molecular species, into species aggregates while preserving the deterministic semantics of the thermodynamic limit. Recently we proposed a quotienting that also preserves the stochastic semantics and that is complete in the sense that the semantics of individual species can be recovered from the aggregate semantics. In this paper we prove that this quotienting yields a sufficient condition for weak lumpability and that it gives rise to a backward Markov bisimulation between the original and aggregated transition system. We illustrate the framework on a case study of the EGF/insulin receptor crosstalk.}, author = {Feret, Jérôme and Henzinger, Thomas A and Koeppl, Heinz and Petrov, Tatjana}, location = {Jena, Germany}, pages = {142--161}, publisher = {Open Publishing Association}, title = {{Lumpability abstractions of rule-based systems}}, volume = {40}, year = {2010}, } @article{3735, abstract = {The visual system is challenged with extracting and representing behaviorally relevant information contained in natural inputs of great complexity and detail. This task begins in the sensory periphery: retinal receptive fields and circuits are matched to the first and second-order statistical structure of natural inputs. This matching enables the retina to remove stimulus components that are predictable (and therefore uninformative), and primarily transmit what is unpredictable (and therefore informative). Here we show that this design principle applies to more complex aspects of natural scenes, and to central visual processing. We do this by classifying high-order statistics of natural scenes according to whether they are uninformative vs. informative. We find that the uninformative ones are perceptually nonsalient, while the informative ones are highly salient, and correspond to previously identified perceptual mechanisms whose neural basis is likely central. Our results suggest that the principle of efficient coding not only accounts for filtering operations in the sensory periphery, but also shapes subsequent stages of sensory processing that are sensitive to high-order image statistics.}, author = {Gasper Tkacik and Prentice, Jason S and Victor,Jonathan D and Balasubramanian, Vijay}, journal = {PNAS}, number = {42}, pages = {18149 -- 18154}, publisher = {National Academy of Sciences}, title = {{Local statistics in natural scenes predict the saliency of synthetic textures}}, doi = {10.1073/pnas.0914916107}, volume = {107}, year = {2010}, } @article{3736, abstract = {In retina and in cortical slice the collective response of spiking neural populations is well described by "maximum-entropy" models in which only pairs of neurons interact. We asked, how should such interactions be organized to maximize the amount of information represented in population responses? To this end, we extended the linear-nonlinear-Poisson model of single neural response to include pairwise interactions, yielding a stimulus-dependent, pairwise maximum-entropy model. We found that as we varied the noise level in single neurons and the distribution of network inputs, the optimal pairwise interactions smoothly interpolated to achieve network functions that are usually regarded as discrete–stimulus decorrelation, error correction, and independent encoding. These functions reflected a trade-off between efficient consumption of finite neural bandwidth and the use of redundancy to mitigate noise. Spontaneous activity in the optimal network reflected stimulus-induced activity patterns, and single-neuron response variability overestimated network noise. Our analysis suggests that rather than having a single coding principle hardwired in their architecture, networks in the brain should adapt their function to changing noise and stimulus correlations.}, author = {Gasper Tkacik and Prentice, Jason S and Balasubramanian, Vijay and Schneidman, Elad}, journal = {PNAS}, number = {32}, pages = {14419 -- 14424}, publisher = {National Academy of Sciences}, title = {{Optimal population coding by noisy spiking neurons}}, doi = {10.1073/pnas.1004906107}, volume = {107}, year = {2010}, } @article{3738, abstract = {Central to the functioning of a living cell is its ability to control the readout or expression of information encoded in the genome. In many cases, a single transcription factor protein activates or represses the expression of many genes. As the concentration of the transcription factor varies, the target genes thus undergo correlated changes, and this redundancy limits the ability of the cell to transmit information about input signals. We explore how interactions among the target genes can reduce this redundancy and optimize information transmission. Our discussion builds on recent work [Tkacik, Phys. Rev. E 80, 031920 (2009)], and there are connections to much earlier work on the role of lateral inhibition in enhancing the efficiency of information transmission in neural circuits; for simplicity we consider here the case where the interactions have a feed forward structure, with no loops. Even with this limitation, the networks that optimize information transmission have a structure reminiscent of the networks found in real biological systems.}, author = {Walczak, Aleksandra M and Gasper Tkacik and Bialek, William S}, journal = {Physical Review E Statistical Nonlinear and Soft Matter Physics}, number = {4}, publisher = {American Institute of Physics}, title = {{Optimizing information flow in small genetic networks. II. Feed-forward interactions}}, doi = {10.1103/PhysRevE.81.041905}, volume = {81}, year = {2010}, } @unpublished{3743, abstract = {These are notes for a set of 7 two-hour lectures given at the 2010 Summer School on Quantitative Evolutionary and Comparative Genomics at OIST, Okinawa, Japan. The emphasis is on understanding how biological systems process information. We take a physicist's approach of looking for simple phenomenological descriptions that can address the questions of biological function without necessarily modeling all (mostly unknown) microscopic details; the example that is developed throughout the notes is transcriptional regulation in genetic regulatory networks. We present tools from information theory and statistical physics that can be used to analyze noisy nonlinear biological networks, and build generative and predictive models of regulatory processes.}, author = {Gasper Tkacik}, booktitle = {ArXiv}, pages = {1 -- 52}, publisher = {ArXiv}, title = {{From statistical mechanics to information theory: understanding biophysical information-processing systems}}, volume = {q-bio.MN}, year = {2010}, } @article{3748, abstract = {The chemotaxis signalling network in Escherichia coli that controls the locomotion of bacteria is a classic model system for signal transduction1, 2. This pathway modulates the behaviour of flagellar motors to propel bacteria towards sources of chemical attractants. Although this system relaxes to a steady state in response to environmental changes, the signalling events within the chemotaxis network are noisy and cause large temporal variations of the motor behaviour even in the absence of stimulus3. That the same signalling network governs both behavioural variability and cellular response raises the question of whether these two traits are independent. Here, we experimentally establish a fluctuation–response relationship in the chemotaxis system of living bacteria. Using this relationship, we demonstrate the possibility of inferring the cellular response from the behavioural variability measured before stimulus. In monitoring the pre- and post-stimulus switching behaviour of individual bacterial motors, we found that variability scales linearly with the response time for different functioning states of the cell. This study highlights that the fundamental relationship between fluctuation and response is not constrained to physical systems at thermodynamic equilibrium4 but is extensible to living cells5. Such a relationship not only implies that behavioural variability and cellular response can be coupled traits, but it also provides a general framework within which we can examine how the selection of a network design shapes this interdependence}, author = {Park, Heungwon and Pontius, William and Calin Guet and Marko, John F and Emonet,Thierry and Cluzel,Philippe}, journal = {Nature}, pages = {819 -- 823}, publisher = {Nature Publishing Group}, title = {{Interdependence of behavioural variability and response to small stimuli in bacteria}}, doi = {10.1038/nature09551}, volume = {468}, year = {2010}, } @article{3749, abstract = {In E. coli, chemotactic behavior exhibits perfect adaptation that is robust to changes in the intracellular concentration of the chemotactic proteins, such as CheR and CheB. However, the robustness of the perfect adaptation does not explicitly imply a robust chemotactic response. Previous studies on the robustness of the chemotactic response relied on swarming assays, which can be confounded by processes besides chemotaxis, such as cellular growth and depletion of nutrients. Here, using a high-throughput capillary assay that eliminates the effects of growth, we experimentally studied how the chemotactic response depends on the relative concentration of the chemotactic proteins. We simultaneously measured both the chemotactic response of E. coli cells to L: -aspartate and the concentrations of YFP-CheR and CheB-CFP fusion proteins. We found that the chemotactic response is fine-tuned to a specific ratio of [CheR]/[CheB] with a maximum response comparable to the chemotactic response of wild-type behavior. In contrast to adaptation in chemotaxis, that is robust and exact, capillary assays revealed that the chemotactic response in swimming bacteria is fined-tuned to wild-type level of the [CheR]/[CheB] ratio.}, author = {Park, Heungwon and Calin Guet and Emonet,Thierry and Cluzel,Philippe}, journal = {Current Microbiology}, number = {3}, pages = {764 -- 769}, publisher = {Springer}, title = {{Fine-tuning of chemotactic response in E. coli determined by high-throughput capillary assay}}, doi = {10.1007/s00284-010-9778-z}, volume = {62}, year = {2010}, } @article{3759, abstract = {We propose a mesh-based surface tracking method for fluid animation that both preserves fine surface details and robustly adjusts the topology of the surface in the presence of arbitrarily thin features like sheets and strands. We replace traditional re-sampling methods with a convex hull method for connecting surface features during topological changes. This technique permits arbitrarily thin fluid features with minimal re-sampling errors by reusing points from the original surface. We further reduce re-sampling artifacts with a subdivision-based mesh-stitching algorithm, and we use a higher order interpolating subdivision scheme to determine the location of any newly-created vertices. The resulting algorithm efficiently produces detailed fluid surfaces with arbitrarily thin features while maintaining a consistent topology with the underlying fluid simulation.}, author = {Wojtan, Christopher J and Thürey, Nils and Gross, Markus and Turk, Greg}, journal = {ACM Transactions on Graphics}, number = {4}, publisher = {ACM}, title = {{Physics-inspired topology changes for thin fluid features}}, doi = {10.1145/1778765.1778787}, volume = {29}, year = {2010}, } @article{3761, abstract = {We present an algorithm for creating realistic animations of characters that are swimming through fluids. Our approach combines dynamic simulation with data-driven kinematic motions (motion capture data) to produce realistic animation in a fluid. The interaction of the articulated body with the fluid is performed by incorporating joint constraints with rigid animation and by extending a solid/fluid coupling method to handle articulated chains. Our solver takes as input the current state of the simulation and calculates the angular and linear accelerations of the connected bodies needed to match a particular motion sequence for the articulated body. These accelerations are used to estimate the forces and torques that are then applied to each joint. Based on this approach, we demonstrate simulated swimming results for a variety of different strokes, including crawl, backstroke, breaststroke, and butterfly. The ability to have articulated bodies interact with fluids also allows us to generate simulations of simple water creatures that are driven by simple controllers.}, author = {Kwatra, Nipun and Wojtan, Christopher J and Carlson, Mark and Essa, Irfan and Mucha, Peter and Turk, Greg}, journal = {IEEE Transactions on Visualization and Computer Graphics}, number = {1}, pages = {70 -- 80}, publisher = {IEEE}, title = {{Fluid simulation with articulated bodies}}, doi = {10.1109/TVCG.2009.66}, volume = {16}, year = {2010}, } @article{3766, abstract = {We present an approach to simulate flows driven by surface tension based on triangle meshes. Our method consists of two simulation layers: the first layer is an Eulerian method for simulating surface tension forces that is free from typical strict time step constraints. The second simulation layer is a Lagrangian finite element method that simulates sub-grid scale wave details on the fluid surface. The surface wave simulation employs an unconditionally stable, symplectic time integration method that allows for a high propagation speed due to strong surface tension. Our approach can naturally separate the grid-and sub-grid scales based on a volumepreserving mean curvature flow. As our model for the sub-grid dynamics enforces a local conservation of mass, it leads to realistic pinch off and merging effects. In addition to this method for simulating dynamic surface tension effects, we also present an efficient non-oscillatory approximation for capturing damped surface tension behavior. These approaches allow us to efficiently simulate complex phenomena associated with strong surface tension, such as Rayleigh-Plateau instabilities and crown splashes, in a short amount of time.}, author = {Thürey, Nils and Wojtan, Christopher J and Gross, Markus and Turk, Greg}, journal = {ACM Transactions on Graphics}, number = {4}, publisher = {ACM}, title = {{A multiscale approach to mesh-based surface tension flows}}, doi = {10.1145/1778765.1778785}, volume = {29}, year = {2010}, } @article{3772, author = {Barton, Nicholas H}, journal = {PLoS Genetics}, number = {6}, publisher = {Public Library of Science}, title = {{Understanding adaptation in large populations}}, doi = {10.1371/journal.pgen.1000987}, volume = {6}, year = {2010}, } @article{3773, abstract = {If distinct biological species are to coexist in sympatry, they must be reproductively isolated and must exploit different limiting resources. A two-niche Levene model is analysed, in which habitat preference and survival depend on underlying additive traits. The population genetics of preference and viability are equivalent. However, there is a linear trade-off between the chances of settling in either niche, whereas viabilities may be constrained arbitrarily. With a convex trade-off, a sexual population evolves a single generalist genotype, whereas with a concave trade-off, disruptive selection favours maximal variance. A pure habitat preference evolves to global linkage equilibrium if mating occurs in a single pool, but remarkably, evolves to pairwise linkage equilibrium within niches if mating is within those niches--independent of the genetics. With a concave trade-off, the population shifts sharply between a unimodal distribution with high gene flow and a bimodal distribution with strong isolation, as the underlying genetic variance increases. However, these alternative states are only simultaneously stable for a narrow parameter range. A sharp threshold is only seen if survival in the 'wrong' niche is low; otherwise, strong isolation is impossible. Gene flow from divergent demes makes speciation much easier in parapatry than in sympatry.}, author = {Barton, Nicholas H}, journal = {Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences}, number = {1547}, pages = {1825 -- 1840}, publisher = {Royal Society}, title = {{What role does natural selection play in speciation?}}, doi = {10.1098/rstb.2010.0001}, volume = {365}, year = {2010}, } @article{3774, abstract = {1. Hybridisation with an invasive species has the potential to alter the phenotype and hence the ecology of a native counterpart. 2. Here data from populations of native red deer Cervus elaphus and invasive sika deer Cervus nippon in Scotland is used to assess the extent to which hybridisation between them is causing phenotypic change. This is done by regression of phenotypic traits against genetic hybrid scores. 3. Hybridisation is causing increases in the body weight of sika-like deer and decreases in the body weight of red-like females. Hybridisation is causing increases in jaw length and increases in incisor arcade breadth in sika-like females. Hybridisation is also causing decreases in incisor arcade breadth in red-like females. 4. There is currently no evidence that hybridisation is causing changes in the kidney fat weight or pregnancy rates of either population. 5. Increased phenotypic similarity between the two species is likely to lead to further hybridisation. The ecological consequences of this are difficult to predict.}, author = {Senn, Helen and Swanson, Graeme and Goodman, Simon and Barton, Nicholas H and Pemberton, Josephine}, journal = {Journal of Animal Ecology}, number = {2}, pages = {414 -- 425}, publisher = {Wiley-Blackwell}, title = {{Phenotypic correlates of hybridisation between red and sika deer (genus Cervus)}}, doi = {10.1111/j.1365-2656.2009.01633.x}, volume = {79}, year = {2010}, } @article{3776, abstract = {The prevalence of recombination in eukaryotes poses one of the most puzzling questions in biology. The most compelling general explanation is that recombination facilitates selection by breaking down the negative associations generated by random drift (i.e. Hill-Robertson interference, HRI). I classify the effects of HRI owing to: deleterious mutation, balancing selection and selective sweeps on: neutral diversity, rates of adaptation and the mutation load. These effects are mediated primarily by the density of deleterious mutations and of selective sweeps. Sequence polymorphism and divergence suggest that these rates may be high enough to cause significant interference even in genomic regions of high recombination. However, neither seems able to generate enough variance in fitness to select strongly for high rates of recombination. It is plausible that spatial and temporal fluctuations in selection generate much more fitness variance, and hence selection for recombination, than can be explained by uniformly deleterious mutations or species-wide selective sweeps.}, author = {Barton, Nicholas H}, journal = {Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences}, number = {1552}, pages = {2559 -- 2569}, publisher = {Royal Society}, title = {{Genetic linkage and natural selection}}, doi = {10.1098/rstb.2010.0106}, volume = {365}, year = {2010}, } @article{3777, abstract = {Under the classical view, selection depends more or less directly on mutation: standing genetic variance is maintained by a balance between selection and mutation, and adaptation is fuelled by new favourable mutations. Recombination is favoured if it breaks negative associations among selected alleles, which interfere with adaptation. Such associations may be generated by negative epistasis, or by random drift (leading to the Hill-Robertson effect). Both deterministic and stochastic explanations depend primarily on the genomic mutation rate, U. This may be large enough to explain high recombination rates in some organisms, but seems unlikely to be so in general. Random drift is a more general source of negative linkage disequilibria, and can cause selection for recombination even in large populations, through the chance loss of new favourable mutations. The rate of species-wide substitutions is much too low to drive this mechanism, but local fluctuations in selection, combined with gene flow, may suffice. These arguments are illustrated by comparing the interaction between good and bad mutations at unlinked loci under the infinitesimal model.}, author = {Barton, Nicholas H}, journal = {Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences}, number = {1544}, pages = {1281 -- 1294}, publisher = {Royal Society}, title = {{Mutation and the evolution of recombination}}, doi = {10.1098/rstb.2009.0320}, volume = {365}, year = {2010}, } @article{3779, abstract = {Crosses between closely related species give two contrasting results. One result is that species hybrids may be inferior to their parents, for example, being less fertile [1]. The other is that F1 hybrids may display superior performance (heterosis), for example with increased vigour [2]. Although various hypotheses have been proposed to account for these two aspects of hybridisation, their biological basis is still poorly understood [3]. To gain further insights into this issue, we analysed the role that variation in gene expression may play. We took a conserved trait, flower asymmetry in Antirrhinum, and determined the extent to which the underlying regulatory genes varied in expression among closely related species. We show that expression of both genes analysed, CYC and RAD, varies significantly between species because of cis-acting differences. By making a quantitative genotype-phenotype map, using a range of mutant alleles, we demonstrate that the species lie on a plateau in gene expression-morphology space, so that the variation has no detectable phenotypic effect. However, phenotypic differences can be revealed by shifting genotypes off the plateau through genetic crosses. Our results can be readily explained if genomes are free to evolve within an effectively neutral zone in gene expression space. The consequences of this drift will be negligible for individual loci, but when multiple loci across the genome are considered, we show that the variation may have significant effects on phenotype and fitness, causing a significant drift load. By considering these consequences for various gene-expression-fitness landscapes, we conclude that F1 hybrids might be expected to show increased performance with regard to conserved traits, such as basic physiology, but reduced performance with regard to others. Thus, our study provides a new way of explaining how various aspects of hybrid performance may arise through natural variation in gene activity.}, author = {Rosas, Ulises and Barton, Nicholas H and Copsey, Lucy and Barbier De Reuille, Pierre and Coen, Enrico}, journal = {PLoS Biology}, number = {7}, publisher = {Public Library of Science}, title = {{Cryptic variation between species and the basis of hybrid performance}}, doi = {10.1371/journal.pbio.1000429}, volume = {8}, year = {2010}, }