Find A Publication

Latest Publications

  • Semantic segmentation has been a long standing challenging task in computer vision. It aims at assigning a label to each image pixel and needs significant number of pixellevel annotated data, which is often unavailable. To address this lack, in this paper, we leverage, on one hand, massive amount of available unlabeled or weakly labeled data, and on the other hand, non-real images created through Generative Adversarial Networks. In particular, we propose a semi-supervised framework ,based on Generative Adversarial Networks (GANs), which consists of a generator network to provide extra training examples to a multi-class classifier, acting as discriminator in the GAN framework, that assigns sample a label y from the K possible classes or marks it as a fake sample (extra class). The underlying idea is that adding large fake visual data forces real samples to be close in the feature space, enabling a bottom-up clustering process, which, in turn, improves multiclass pixel classification. To ensure higher quality of generated images for GANs with consequent improved pixel classification, we extend the above framework by adding weakly annotated data, i.e., we provide class level information to the generator. We tested our approaches on several challenging benchmarking visual datasets, i.e. PASCAL, SiftFLow, Stanford and CamVid, achieving competitive performance also compared to state-of-the-art semantic segmentation method Read More
  • A strained graphene monolayer is shown to operate as a highly efficient quantum heat engine delivering maximum power. The efficiency and power of the proposed device exceeds that of recent proposals. The reason for these excellent characteristics is that strain enables complete valley separation in transmittance through the device, implying that increasing strain leads to very high Seeback coefficient as well as lower conductance. In addition, since time-reversal symmetry is unbroken in our system, the proposed strained graphene quantum heat engine can also act as a high performance refrigerator. Read More
  • Many link prediction algorithms require the computation of a similarity metric on each vertex pair, which is quadratic in the number of vertices and infeasible for large networks. We develop a class of link prediction algorithms based on a spectral embedding and the k closest pairs algorithm that are scalable to very large networks. We compare the prediction accuracy and runtime of these methods to existing algorithms on several large link prediction tasks. Our methods achieve comparable accuracy to standard algorithms but are significantly faster. Read More
  • We present pySecDec, a new version of the program SecDec, which performs the factorisation of dimensionally regulated poles in parametric integrals, and the subsequent numerical evaluation of the finite coefficients. The algebraic part of the program is now written in the form of python modules, which allow a very flexible usage. The optimization of the C++ code, generated using FORM, is improved, leading to a faster numerical convergence. The new version also creates a library of the integrand functions, such that it can be linked to user-specific codes for the evaluation of matrix elements in a way similar to analytic integral libraries. Read More
  • The aim of this work is to quantify the spatio-temporal dynamics of flow-driven amoeboid locomotion in small ($\sim$100 $\mu$m) fragments of the true slime mold \phys {\it polycephalum}. In this model organism, cellular contraction drives intracellular flows, and these flows transport the chemical signals that regulate contraction in the first place. As a consequence of these non-linear interactions, a diversity of migratory behaviors can be observed in migrating \phys fragments. To study these dynamics, we measure the spatio-temporal distributions of the velocities of the endoplasm and ectoplasm of each migrating fragment, the traction stresses it generates on the substratum, and the concentration of free intracellular calcium. Using these unprecedented experimental data, we classify migrating \phys fragments according to their dynamics, finding that they often exhibit spontaneously coordinated waves of flow, contractility and chemical signaling. We show that \phys fragments exhibiting symmetric spatio-temporal patterns of endoplasmic flow migrate significantly slower than fragments with asymmetric patterns. In addition, our joint measurements of ectoplasm velocity and traction stress at the substratum suggest that forward motion of the ectoplasm is enabled by a succession of stick-slip transitions, which we conjecture are also organized in the form of waves. Combining our experiments with a simplified convection-diffusion model, we show that the convective transport of calcium ions may be key for establishing and maintaining the spatio-temporal patterns of calcium concentration that regulate the generation of contractile forces. Read More
  • Sparse coding (SC) is an automatic feature extraction and selection technique that is widely used in unsupervised learning. However, conventional SC vectorizes the input images, which breaks apart the local proximity of pixels and destructs the elementary object structures of images. In this paper, we propose a novel two-dimensional sparse coding (2DSC) scheme that represents the input images as the tensor-linear combinations under a novel algebraic framework. 2DSC learns much more concise dictionaries because it uses the circular convolution operator, since the shifted versions of atoms learned by conventional SC are treated as the same ones. We apply 2DSC to natural images and demonstrate that 2DSC returns meaningful dictionaries for large patches. Moreover, for mutli-spectral images denoising, the proposed 2DSC reduces computational costs with competitive performance in comparison with the state-of-the-art algorithms. Read More