No abstract

For the past years, the analysts have been predicting a tremendous and continuous increase in mobile traffic, causing much of industry and academia to seek out any and all methods to increase wireless network capacity. In this paper, we investigate one such method, cellular data offloading onto direct connections between proximate user devices, which has been shown to provide significant wireless capacity gains. To do so, we formulate a new system model that couples a cellular network in licensed bands and a device-to-device (D2D) network in unlicensed bands. We propose that devices be continually associated with the cellular base station and use this connectivity to help manage their direct connections in unlicensed spectrum. In particular, we demonstrate that assisted offloading of cellular user sessions onto the D2D links improves the degree of spatial reuse and reduces the impact of interference. In this study, a session is a real-time flow of data from one user to another, which adheres to a Poisson point process (PPP). By contrast to a throughput-or capacity-centric system view, the application of PPP enables formulations where entire user sessions, rather than singular data packets, are arriving at random and leaving the system after being served. The proposed methodology is flexible enough to accommodate practical offloading scenarios, network selection algorithms, quality of service measures, and advanced wireless technologies. In this study, we are primarily interested in evaluating the data session blocking probability in dynamically loaded cellular and D2D networks, but given the importance of energy efficiency for mobile devices, we are also interested in characterizing the energy expenditure of a typical data session in these different networks. First with our advanced analytical methodology and then with our detailed system-level simulator, we evaluate the performance of network-assisted data session offloading from cellular to D2D connections under a variety of conditions. This analysis represents a useful tool in the development of practical offloading schemes and ongoing standardization efforts.

Chronic infection perturbs immune homeostasis. While prior studies have reported dysregulation of effector and memory cells, little is known about the effects on naïve T cell populations. We performed a cross-sectional study of chronic hepatitis C (cHCV) patients using tetramer-associated magnetic enrichment to study antigen-specific inexperienced CD8+ T cells (i.e., tumor or unrelated virus-specific populations in tumor-free and sero-negative individuals). cHCV showed normal precursor frequencies, but increased proportions of memory-phenotype inexperienced cells, as compared to healthy donors or cured HCV patients. These observations could be explained by low surface expression of CD5, a negative regulator of TCR signaling. Accordingly, we demonstrated TCR hyperactivation and generation of potent CD8+ T cell responses from the altered T cell repertoire of cHCV patients. In sum, we provide the first evidence that naïve CD8+ T cells are dysregulated during cHCV infection, and establish a new mechanism of immune perturbation secondary to chronic infection.DOI: http://dx.doi.org/10.7554/eLife.07916.001

No abstract

No abstract

12 1536-1284/15/25.00 © 2015 IEEE Alexander Pyattaev, Sergey Andreev, and Yevgeni Koucheryavy are with Tampere University of Technology. Kerstin Johnsson is with Intel Corporation.

No abstract

In exploratory high-dimensional data analysis, local intrinsic dimension estimation can sometimes be used in order to discriminate between data sets sampled from different low-dimensional structures. Global intrinsic dimension estimators can in many cases be adapted to local estimation, but this leads to problems with high negative bias or high variance. We introduce a method that exploits the curse/blessing of dimensionality and produces local intrinsic dimension estimators that have very low bias, even in cases where the intrinsic dimension is higher than the number of data points, in combination with relatively low variance. We show that our estimators have a very good ability to classify local data sets by their dimension compared to other local intrinsic dimension estimators; furthermore we provide examples showing the usefulness of local intrinsic dimension estimation in general and our method in particular for stratification of real data sets.