Search

Two-dimensional heterostructures have been crucial in advancing optoelectronic devices utilizing van der Waals materials. Semiconducting transition metal dichalcogenide monolayers, known for their unique optical properties, offer extensive possibilities for light-emitting devices. Recently, a memory-driven optical device, termed a Mem-emitter, was proposed using these monolayers atop dielectric substrates. The successful realization of such devices heavily depends on selecting the optimal substrate. Here, we report a pronounced memory effect in a MoSe$_2$/clinochlore device, evidenced by electric hysteresis in the intensity and energy of MoSe$_2$ monolayer emissions. This demonstrates both population-driven and transition-rate-driven Mem-emitter abilities. Our theoretical approach correlates these memory effects with internal state variables of the substrate, emphasizing that clinochlore layered structure is crucial for a robust and rich memory response. This work introduces a novel two-dimensional device with promising applications in memory functionalities, highlighting the importance of alternative insulators in fabricating van der Waals heterostructures.

We use James Webb Space Telescope (JWST) Mid-Infrared Instrument (MIRI) medium-resolution spectrometer (MRS) observations of UGC 8782, CGCG 012-070 and NGC 3884 to investigate the origin of the H$_2$ emission. These three nearby AGN hosts are known to present H$_2$ emission excess relative to star-forming galaxies, as traced by the H$_2$ S(3)/PAH$_{\rm 11.3\mu m}$ line ratio. For every spaxel, we also measure the velocity width ($W_{\rm 80}$) of H$_2$ S(3). We find that the distribution of H$_2$ in the space of these parameters - H$_2$/PAH and its velocity width - is bimodal: one cluster of points shows approximately constant values of $W_{\rm 80}$, while the other cluster exhibits a strong correlation between $W_{\rm 80}$ and H$_2$/PAH. We estimated the temperature of the H$_2$ gas assuming a power-law distribution and find flatter distributions in regions where a correlation between $W_{\rm 80}$ and H$_2$ S(3)/PAH$_{\rm 11.3\mu m}$ is observed. Additionally, we observed a correlation between the [Fe II]$_{\rm 5.34 \mu m}$ (a known shock tracer) and the H$_2$ emission. This indicates that the excess H$_2$ emission excess is associated to shock heating of the gas, generated by outflows or by the interaction of the radio jet with the ambient gas., Comment: 12 pages, 5 figures, submitted to ApJ. Comments are welcome

We use the James Webb Space Telescope (JWST) Mid-Infrared Instrument (MIRI) medium-resolution spectrometer (MRS) observations of the radio loud AGN host UGC 8782 to map the warm molecular and ionized gas kinematics. The data reveal spatially resolved outflows in the inner 2 kpc, seen in low ionization (traced by the [Ar ii] 6.99um emission) and in warm molecular gas (traced by the H2 rotational transitions). We find a maximum mass-outflow rate of 4.90 (2.04) M/yr at ~900 pc from the nucleus for the warm outflow (198 K< T< 1000 K) and estimate and outflow rate of up to 1.22 (0.51) M/yr for the hotter gas phase (T > 1000 K). These outflows can clear the entire nuclear reservoir of warm molecular gas in about 1 Myr. The derived kinetic power of the molecular outflows lead to coupling efficiencies of 2-5 percent of the AGN luminosity, way above the minimum expected to the AGN feedback be effective quenching the star formation., Comment: Accepted for publication at APJ

Despite the remarkable progress in emerging solar cell technologies such as hybrid organic-inorganic perovskites, there are still significant limitations related to the stability of the devices and their non-ideal electrical behavior under certain external stimuli. We present a conceptual framework for characterizing photovoltaic devices by integrating cyclic voltammetry (CV) and impedance spectroscopy (IS). This framework is constructed from a microscopic, multi-mode perspective that explicitly accounts for drift, diffusion, displacement, and memory contributions. We derive comprehensive analytical expressions for current-voltage relationships and complex admittance. Our model reveals the inseparable connection between hysteresis behaviors in current-voltage characteristics observed in CV and the apparent capacitive and inductive behaviors seen in IS spectral analysis. We demonstrate how CV and IS naturally complement each other, providing a deeper microscopic understanding of device performance and limitations. Additionally, we establish the relationship between intrinsic material parameters and experimentally accessible extrinsic parameters such as light intensity, temperature, DC bias, voltage amplitude, and frequency. This framework enables unprecedented optimization of solar cell performance, marking a significant advancement towards sustainability., Comment: 6 figures

Despite attaining high empirical generalization, the sharpness of models trained with sharpness-aware minimization (SAM) do not always correlate with generalization error. Instead of viewing SAM as minimizing sharpness to improve generalization, our paper considers a new perspective based on SAM's training dynamics. We propose that perturbations in SAM perform perturbed forgetting, where they discard undesirable model biases to exhibit learning signals that generalize better. We relate our notion of forgetting to the information bottleneck principle, use it to explain observations like the better generalization of smaller perturbation batches, and show that perturbed forgetting can exhibit a stronger correlation with generalization than flatness. While standard SAM targets model biases exposed by the steepest ascent directions, we propose a new perturbation that targets biases exposed through the model's outputs. Our output bias forgetting perturbations outperform standard SAM, GSAM, and ASAM on ImageNet, robustness benchmarks, and transfer to CIFAR-{10,100}, while sometimes converging to sharper regions. Our results suggest that the benefits of SAM can be explained by alternative mechanistic principles that do not require flatness of the loss surface., Comment: Published as a conference paper at ICML 2024. 9 pages main, 15 pages total including references and appendix

The modulation efficiency of a tubular-lattice hollow-core fiber (HCF) by means of flexural acoustic waves is investigated in detail for the first time. The main acousto-optic properties of the HCF are evaluated employing 2D and 3D models based on the finite element method. The induced coupling of the fundamental and first higher-order modes is simulated in the wavelength range from 743 to 1355 nm. Significant acoustic (amplitude, period, strain, energy) and optical parameters (effective index, beat length, birefringence, coupling coefficient) are analyzed. The simulations are compared to experimental results, indicating higher modulation performance in HCFs compared to standard optical fibers. In addition, useful insights into the design and fabrication of all-fiber acousto-optic devices based on HCFs are provided, enabling potential application in tunable spectral filters and mode-locked fiber lasers.

Purpose: This study assesses the effectiveness of Deep Learning (DL) for creating synthetic CT (sCT) images in MR-guided adaptive radiation therapy (MRgART). Methods: A Cycle-GAN model was trained with MRI and CT scan slices from MR-LINAC treatments, generating sCT volumes. The analysis involved retrospective treatment plan data from patients with various tumors. sCT images were compared with standard CT scans using mean absolute error in Hounsfield Units (HU) and image similarity metrics (SSIM, PSNR, NCC). sCT volumes were integrated into a clinical treatment system for dosimetric re-evaluation. Results: The model, trained on 8405 frames from 57 patients and tested on 357 sCT frames from 17 patients, showed sCTs comparable to dCTs in electron density and structural similarity with MRI scans. The MAE between sCT and dCT was 49.2 +/- 13.2 HU, with sCT NCC exceeding dCT by 0.06, and SSIM and PSNR at 0.97 +/- 0.01 and 19.9 +/- 1.6 respectively. Dosimetric evaluations indicated minimal differences between sCTs and dCTs, with sCTs showing better air-bubble reconstruction. Conclusions: DL-based sCT generation on MR-Linacs is accurate for dose calculation and optimization in MRgART. This could facilitate MR-only treatment planning, enhancing simulation and adaptive planning efficiency on MR-Linacs.

Quantum systems of infinite dimension, such as bosonic oscillators, provide vast resources for quantum sensing. Yet, a general theory on how to manipulate such bosonic modes for sensing beyond parameter estimation is unknown. We present a general algorithmic framework, quantum signal processing interferometry (QSPI), for quantum sensing at the fundamental limits of quantum mechanics by generalizing Ramsey-type interferometry. Our QSPI sensing protocol relies on performing nonlinear polynomial transformations on the oscillator's quadrature operators by generalizing quantum signal processing (QSP) from qubits to hybrid qubit-oscillator systems. We use our QSPI sensing framework to make efficient binary decisions on a displacement channel in the single-shot limit. Theoretical analysis suggests the sensing accuracy, given a single-shot qubit measurement, scales inversely with the sensing time or circuit depth of the algorithm. We further concatenate a series of such binary decisions to perform parameter estimation in a bit-by-bit fashion. Numerical simulations are performed to support these statements. Our QSPI protocol offers a unified framework for quantum sensing using continuous-variable bosonic systems beyond parameter estimation and establishes a promising avenue toward efficient and scalable quantum control and quantum sensing schemes beyond the NISQ era., Comment: 26 pages, 8 figures, 1 table

We demonstrate a broadband acousto-optic notch filter based on a tubular-lattice hollow-core fiber for the first time. The guided optical modes are modulated by acoustically induced dynamic long-period gratings along the fiber. The device is fabricated employing a short interaction length (7.7 cm) and low drive voltages (10 V). Modulated spectral bands with 20 nm half-width and maximum depths greater than 60 % are achieved. The resonant notch wavelength is tuned from 743 to 1355 nm (612 nm span) by changing the frequency of the electrical signal. The results indicate a broader tuning range compared to previous studies using standard and hollow-core fibers. It further reveals unique properties for reconfigurable spectral filters and fiber lasers, pointing to the fast switching and highly efficient modulation of all-fiber photonic devices.

Although neural networks are conventionally optimized towards zero training loss, it has been recently learned that targeting a non-zero training loss threshold, referred to as a flood level, often enables better test time generalization. Current approaches, however, apply the same constant flood level to all training samples, which inherently assumes all the samples have the same difficulty. We present AdaFlood, a novel flood regularization method that adapts the flood level of each training sample according to the difficulty of the sample. Intuitively, since training samples are not equal in difficulty, the target training loss should be conditioned on the instance. Experiments on datasets covering four diverse input modalities - text, images, asynchronous event sequences, and tabular - demonstrate the versatility of AdaFlood across data domains and noise levels.

This research studies two systems composed by the Timoshenko beam model for double wall carbon nanotubes, coupled with the heat equation governed by Fourier's law. For the first system, the coupling is given by the speed the rotation of the vertical filament in the beam $\beta\psi_t$ from the first beam of Tymoshenko and the Laplacian of temperature $\delta\theta_{xx}$, where we also consider the damping terms fractionals $\gamma_1(-\partial_{xx})^{\tau_1}\phi_t$, $\gamma_2(-\partial_{xx})^{\tau_2} y_t$ and $\gamma_3(-\partial_{xx})^{\tau_3} z_t$, where $(\tau_1, \tau_2, \tau_3) \in [0,1]^3$. For this first system we proved that the semigroup $S_1(t)$ associated to system decays exponentially for all $(\tau_1 , \tau_2 , \tau_3 ) \in [0,1]^3$. The second system also has three fractional damping $\gamma_1(-\partial_{xx})^{\beta_1}\phi_t$, $\gamma_2(-\partial_{xx})^{\beta_2} y_t$ and $\gamma_3(-\partial_{xx})^{\beta_3} z_t$, with $(\beta_1, \beta_2, \beta_3) \in [0,1]^3$. Furthermore, the couplings between the heat equation and the Timoshenko beams of the double wall carbon nanotubes for the second system is given by the Laplacian of the rotation speed of the vertical filament in the beam $\beta\psi_{xxt}$ of the first beam of Timoshenko and the Lapacian of the temperature $\delta\theta_{xx}$. For the second system, we prove the exponential decay of $S_2(t)$ for $(\beta_1, \beta_2, \beta_3) \in [0,1]^3$ and also show that $S_2(t)$ admits Gevrey classes $s>(\phi+1)/(2\phi)$ for $\phi=\min\{\beta_1,\beta_2,\beta_3\}, \forall (\beta_1,\beta_2,\beta_3)\in (0,1)^3$, and proving that $S_2(t)$ is analytic when the parameters $(\beta_1, \beta_2, \beta_3) \in [1/2,1]^3$. One of the motivations for this research was the work; Ramos et al. \cite{Ramos2023CNTs}, whose partial results are part of our results obtained for the first system for $(\tau_1, \tau_2, \tau_3) = (0, 0, 0)$., Comment: 27 pages, 4 figures. arXiv admin note: text overlap with arXiv:2210.12570

Transit riders' feedback provided in ridership surveys, customer relationship management (CRM) channels, and in more recent times, through social media is key for transit agencies to better gauge the efficacy of their services and initiatives. Getting a holistic understanding of riders' experience through the feedback shared in those instruments is often challenging, mostly due to the open-ended, unstructured nature of text feedback. In this paper, we propose leveraging traditional transit CRM feedback to develop and deploy a transit-topic-aware large language model (LLM) capable of classifying open-ended text feedback to relevant transit-specific topics. First, we utilize semi-supervised learning to engineer a training dataset of 11 broad transit topics detected in a corpus of 6 years of customer feedback provided to the Washington Metropolitan Area Transit Authority (WMATA). We then use this dataset to train and thoroughly evaluate a language model based on the RoBERTa architecture. We compare our LLM, MetRoBERTa, to classical machine learning approaches utilizing keyword-based and lexicon representations. Our model outperforms those methods across all evaluation metrics, providing an average topic classification accuracy of 90%. Finally, we provide a value proposition of this work demonstrating how the language model, alongside additional text processing tools, can be applied to add structure to open-ended text sources of feedback like Twitter. The framework and results we present provide a pathway for an automated, generalizable approach for ingesting, visualizing, and reporting transit riders' feedback at scale, enabling agencies to better understand and improve customer experience.

We report on the creation of an array of spin-squeezed ensembles of cesium atoms via Rydberg dressing, a technique that offers optical control over local interactions between neutral atoms. We optimize the coherence of the interactions by a stroboscopic dressing sequence that suppresses super-Poissonian loss. We thereby prepare squeezed states of $N=200$ atoms with a metrological squeezing parameter $\xi^2 = 0.77(9)$ quantifying the reduction in phase variance below the standard quantum limit. We realize metrological gain across three spatially separated ensembles in parallel, with the strength of squeezing controlled by the local intensity of the dressing light. Our method can be applied to enhance the precision of tests of fundamental physics based on arrays of atomic clocks and to enable quantum-enhanced imaging of electromagnetic fields., Comment: 18 pages, 11 figures, typos corrected, edits for clarity

A temporal point process (TPP) is a stochastic process where its realization is a sequence of discrete events in time. Recent work in TPPs model the process using a neural network in a supervised learning framework, where a training set is a collection of all the sequences. In this work, we propose to train TPPs in a meta learning framework, where each sequence is treated as a different task, via a novel framing of TPPs as neural processes (NPs). We introduce context sets to model TPPs as an instantiation of NPs. Motivated by attentive NP, we also introduce local history matching to help learn more informative features. We demonstrate the potential of the proposed method on popular public benchmark datasets and tasks, and compare with state-of-the-art TPP methods., Comment: Accepted to ICLR2023

In 2021, Duarte, Oliveira, and Souza [MFCS 2021] showed some problems that are FPT when parameterized by the treewidth of the complement graph (called co-treewidth). Since the degeneracy of a graph is at most its treewidth, they also introduced the study of co-degeneracy (the degeneracy of the complement graph) as a parameter. In 1976, Bondy and Chv\'{a}tal [DM 1976] introduced the notion of closure of a graph: let $\ell$ be an integer; the $(n+\ell)$-closure, $\operatorname{cl}_{n+\ell}(G)$, of a graph $G$ with $n$ vertices is obtained from $G$ by recursively adding an edge between pairs of nonadjacent vertices whose degree sum is at least $n+\ell$ until no such pair remains. A graph property $\Upsilon$ defined on all graphs of order $n$ is said to be $(n+\ell)$-stable if for any graph $G$ of order $n$ that does not satisfy $\Upsilon$, the fact that $uv$ is not an edge of $G$ and that $G+uv$ satisfies $\Upsilon$ implies $d(u)+d(v)< n+\ell$. Duarte et al. [MFCS 2021] developed an algorithmic framework for co-degeneracy parameterization based on the notion of closures for solving problems that are $(n+\ell)$-stable for some $\ell$ bounded by a function of the co-degeneracy. In this paper, we first determine the stability of the property of having a bounded cycle cover. After that, combining the framework of Duarte et al. [MFCS 2021] with some results of Jansen, Kozma, and Nederlof [WG 2019], we obtain a $2^{\mathcal{O}(k)}\cdot n^{\mathcal{O}(1)}$-time algorithm for Minimum Cycle Cover on graphs with co-degeneracy at most $k$, which generalizes Duarte et al. [MFCS 2021] and Jansen et al. [WG 2019] results concerning the Hamiltonian Cycle problem.

As cities struggle to adapt to more ``people-centered'' urbanism, transportation planning and engineering must innovate to expand the street network strategically in order to ensure efficiency but also to deter sprawl. Here, we conducted a study of over 200 cities around the world to understand the impact that the patterns of deceleration points in streets due to traffic signs has in trajectories done from motorized vehicles. We demonstrate that there is a ubiquitous nonlinear relationship between time and distance in the optimal trajectories within each city. More precisely, given a specific period of time $\tau$, without any traffic, one can move on average up to the distance $\left \langle D \right \rangle \sim\tau^\beta$. We found a super-linear relationship for almost all cities in which $\beta>1.0$. This points to an efficiency of scale when traveling large distances, meaning the average speed will be higher for longer trips when compared to shorter trips. We demonstrate that this efficiency is a consequence of the spatial distribution of large segments of streets without deceleration points, favoring access to routes in which a vehicle can cross large distances without stops. These findings show that cities must consider how their street morphology can affect travel speed., Comment: 10 pages, 7 figures

Predicting multiple heterogeneous biological and medical targets is a challenge for traditional deep learning models. In contrast to single-task learning, in which a separate model is trained for each target, multi-task learning (MTL) optimizes a single model to predict multiple related targets simultaneously. To address this challenge, we propose the Multi-gate Mixture-of-Experts with Exclusivity (MMoEEx). Our work aims to tackle the heterogeneous MTL setting, in which the same model optimizes multiple tasks with different characteristics. Such a scenario can overwhelm current MTL approaches due to the challenges in balancing shared and task-specific representations and the need to optimize tasks with competing optimization paths. Our method makes two key contributions: first, we introduce an approach to induce more diversity among experts, thus creating representations more suitable for highly imbalanced and heterogenous MTL learning; second, we adopt a two-step optimization [6, 11] approach to balancing the tasks at the gradient level. We validate our method on three MTL benchmark datasets, including Medical Information Mart for Intensive Care (MIMIC-III) and PubChem BioAssay (PCBA)., Comment: 10 pages, 7 figures, BIOKDD, IEEE/ACM

Single encoder-decoder methodologies for semantic segmentation are reaching their peak in terms of segmentation quality and efficiency per number of layers. To address these limitations, we propose a new architecture based on a decoder which uses a set of shallow networks for capturing more information content. The new decoder has a new topology of skip connections, namely backward and stacked residual connections. In order to further improve the architecture we introduce a weight function which aims to re-balance classes to increase the attention of the networks to under-represented objects. We carried out an extensive set of experiments that yielded state-of-the-art results for the CamVid, Gatech and Freiburg Forest datasets. Moreover, to further prove the effectiveness of our decoder, we conducted a set of experiments studying the impact of our decoder to state-of-the-art segmentation techniques. Additionally, we present a set of experiments augmenting semantic segmentation with optical flow information, showing that motion clues can boost pure image based semantic segmentation approaches.

The cut-set $\partial(S)$ of a graph $G=(V,E)$ is the set of edges that have one endpoint in $S\subset V$ and the other endpoint in $V\setminus S$, and whenever $G[S]$ is connected, the cut $[S,V\setminus S]$ of $G$ is called a connected cut. A bond of a graph $G$ is an inclusion-wise minimal disconnecting set of $G$, i.e., bonds are cut-sets that determine cuts $[S,V\setminus S]$ of $G$ such that $G[S]$ and $G[V\setminus S]$ are both connected. Contrasting with a large number of studies related to maximum cuts, there exist very few results regarding the largest bond of general graphs. In this paper, we aim to reduce this gap on the complexity of computing the largest bond, and the maximum connected cut of a graph. Although cuts and bonds are similar, we remark that computing the largest bond and the maximum connected cut of a graph tends to be harder than computing its maximum cut. We show that it does not exist a constant-factor approximation algorithm to compute the largest bond, unless P = NP. Also, we show that {\sc Largest Bond} and {\sc Maximum Connected Cut} are NP-hard even for planar bipartite graphs, whereas \textsc{Maximum Cut} is trivial on bipartite graphs and polynomial-time solvable on planar graphs. In addition, we show that {\sc Largest Bond} and {\sc Maximum Connected Cut} are NP-hard on split graphs, and restricted to graphs of clique-width $w$ they can not be solved in time $f(w)\times n^{o(w)}$ unless the Exponential Time Hypothesis fails, but they can be solved in time $f(w)\times n^{O(w)}$. Finally, we show that both problems are fixed-parameter tractable when parameterized by the size of the solution, the treewidth, and the twin-cover number., Comment: This paper resulted from a merge of two papers submitted to arXiv (arXiv:1908.03389 and arXiv:1910.01071). Both preliminary versions were presented at the 14th International Symposium on Parameterized and Exact Computation (IPEC 2019)

In semantic segmentation knowing about all existing classes is essential to yield effective results with the majority of existing approaches. However, these methods trained in a Closed Set of classes fail when new classes are found in the test phase. It means that they are not suitable for Open Set scenarios, which are very common in real-world computer vision and remote sensing applications. In this paper, we discuss the limitations of Closed Set segmentation and propose two fully convolutional approaches to effectively address Open Set semantic segmentation: OpenFCN and OpenPCS. OpenFCN is based on the well-known OpenMax algorithm, configuring a new application of this approach in segmentation settings. OpenPCS is a fully novel approach based on feature-space from DNN activations that serve as features for computing PCA and multi-variate gaussian likelihood in a lower dimensional space. Experiments were conducted on the well-known Vaihingen and Potsdam segmentation datasets. OpenFCN showed little-to-no improvement when compared to the simpler and much more time efficient SoftMax thresholding, while being between some orders of magnitude slower. OpenPCS achieved promising results in almost all experiments by overcoming both OpenFCN and SoftMax thresholding. OpenPCS is also a reasonable compromise between the runtime performances of the extremely fast SoftMax thresholding and the extremely slow OpenFCN, being close able to run close to real-time. Experiments also indicate that OpenPCS is effective, robust and suitable for Open Set segmentation, being able to improve the recognition of unknown class pixels without reducing the accuracy on the known class pixels., Comment: Submitted to the Machine Learning Journal

A bond of a graph $G$ is an inclusion-wise minimal disconnecting set of $G$, i.e., bonds are cut-sets that determine cuts $[S,V\setminus S]$ of $G$ such that $G[S]$ and $G[V\setminus S]$ are both connected. Given $s,t\in V(G)$, an $st$-bond of $G$ is a bond whose removal disconnects $s$ and $t$. Contrasting with the large number of studies related to maximum cuts, there are very few results regarding the largest bond of general graphs. In this paper, we aim to reduce this gap on the complexity of computing the largest bond and the largest $st$-bond of a graph. Although cuts and bonds are similar, we remark that computing the largest bond of a graph tends to be harder than computing its maximum cut. We show that {\sc Largest Bond} remains NP-hard even for planar bipartite graphs, and it does not admit a constant-factor approximation algorithm, unless $P = NP$. We also show that {\sc Largest Bond} and {\sc Largest $st$-Bond} on graphs of clique-width $w$ cannot be solved in time $f(w)\times n^{o(w)}$ unless the Exponential Time Hypothesis fails, but they can be solved in time $f(w)\times n^{O(w)}$. In addition, we show that both problems are fixed-parameter tractable when parameterized by the size of the solution, but they do not admit polynomial kernels unless NP $\subseteq$ coNP/poly., Comment: An extended abstract of this paper was published in IPEC 2019

Compared to LiDAR-based localization methods, which provide high accuracy but rely on expensive sensors, visual localization approaches only require a camera and thus are more cost-effective while their accuracy and reliability typically is inferior to LiDAR-based methods. In this work, we propose a vision-based localization approach that learns from LiDAR-based localization methods by using their output as training data, thus combining a cheap, passive sensor with an accuracy that is on-par with LiDAR-based localization. The approach consists of two deep networks trained on visual odometry and topological localization, respectively, and a successive optimization to combine the predictions of these two networks. We evaluate the approach on a new challenging pedestrian-based dataset captured over the course of six months in varying weather conditions with a high degree of noise. The experiments demonstrate that the localization errors are up to 10 times smaller than with traditional vision-based localization methods., Comment: 16 pages, 7 figures, ISRR 2017 submission

Robots are expected to operate autonomously in dynamic environments. Understanding the underlying dynamic characteristics of objects is a key enabler for achieving this goal. In this paper, we propose a method for pointwise semantic classification of 3D LiDAR data into three classes: non-movable, movable and dynamic. We concentrate on understanding these specific semantics because they characterize important information required for an autonomous system. Non-movable points in the scene belong to unchanging segments of the environment, whereas the remaining classes corresponds to the changing parts of the scene. The difference between the movable and dynamic class is their motion state. The dynamic points can be perceived as moving, whereas movable objects can move, but are perceived as static. To learn the distinction between movable and non-movable points in the environment, we introduce an approach based on deep neural network and for detecting the dynamic points, we estimate pointwise motion. We propose a Bayes filter framework for combining the learned semantic cues with the motion cues to infer the required semantic classification. In extensive experiments, we compare our approach with other methods on a standard benchmark dataset and report competitive results in comparison to the existing state-of-the-art. Furthermore, we show an improvement in the classification of points by combining the semantic cues retrieved from the neural network with the motion cues., Comment: 8 pages to be published in IROS 2017

General human action recognition requires understanding of various visual cues. In this paper, we propose a network architecture that computes and integrates the most important visual cues for action recognition: pose, motion, and the raw images. For the integration, we introduce a Markov chain model which adds cues successively. The resulting approach is efficient and applicable to action classification as well as to spatial and temporal action localization. The two contributions clearly improve the performance over respective baselines. The overall approach achieves state-of-the-art action classification performance on HMDB51, J-HMDB and NTU RGB+D datasets. Moreover, it yields state-of-the-art spatio-temporal action localization results on UCF101 and J-HMDB., Comment: 10 pages, 7 figures, ICCV 2017 submission

We use analogies to give introductory laboratory students intuition about measurement uncertainties. Using a battery-resistor circuit we discuss uncertainty concepts and derive expressions for uncertainty of the mean and sums of uncertainties. Finally, we draw attention to the fact that the interpretation of standard deviation as uncertainty depends on the statistical distribution of the data, while the interpretation of uncertainty of the mean is largely insensitive to such distribution, especially for large samples. If the resistor in the battery-resistor circuit is a resistive loudspeaker, the uncertainty or noise is literally the acoustic power of the sound produced by such noise. In the sound analogy the statistical distribution of the noise is related to its timbre., Comment: 13 pages, 6 figures

HTML5 WebSocket protocol brings real time communication in web browsers to a new level. Daily, new products are designed to stay permanently connected to the web. WebSocket is the technology enabling this revolution. WebSockets are supported by all current browsers, but it is still a new technology in constant evolution. WebSockets are slowly replacing older client-server communication technologies. As opposed to comet-like technologies WebSockets' remarkable performances is a result of the protocol's fully duplex nature and because it doesn't rely on HTTP communications. To begin with this paper studies the WebSocket protocol and different WebSocket servers implementations. This first theoretic part focuses more deeply on heterogeneous implementations and OpenCL. The second part is a benchmark of a new promising library. The real-time engine used for testing purposes is SocketCluster. SocketCluster provides a highly scalable WebSocket server that makes use of all available cpu cores on an instance. The scope of this work is reduced to vertical scaling of SocketCluster.

The South African Library of Parliament and Research unit has devoted considerable time and money to improving its research and library services for members, committees, and management of Parliament. The Library and Research Units have recently re-structured and re-designed services and information products, engaging in collection development policy, stock-taking and weeding, upgrade of the information technology infrastructure, and changes of the organizational structure to reflect the newly formed Information Services Section (ISS). The aim was to create an ISS that will facilitate the work of members, committees, and management of Parliament by gathering, analyzing, and disseminating relevant information in a non-partisan, easily accessible, and understandable format. This paper presents a case study of the formation of the ISS. Topics covered include historical background, unified information service, information processing, the benefits of parliamentary information service, and information technology. (Author/MES)

A nozzle assembly for a dual gas turbine engine propulsion system includes a housing mountable proximate to a first bypass passage of a first gas turbine engine and a second bypass passage of a second gas turbine engine, first and second upper doors, and first and second lower doors. Each of the first and second upper doors and the first and second lower doors are pivotally mounted to the housing for movement between a stowed position and a deployed position in which airflow through the first and second bypass passages is redirected relative to respective centerline axes of the first and second gas turbine engines.

An aircraft assembly is disclosed and includes a fuselage including a turbine engine mounted within the aft fuselage. A burst zone is defined about the turbine engine and a tail is disposed at least partially with the burst zone. The tail includes primary control surfaces and sacrificial control surfaces. The sacrificial control surfaces can break away in a defined manner to maintain integrity of the primary control surfaces outside of the burst zone.

A propulsion system for an aircraft includes first and second turbine engines mounted within a fuselage of the aircraft. The first turbine engine includes a first engine core that drives a first propulsor disposed about a first propulsor axis. The second turbine engine includes a second engine core and a second propulsor disposed about a second propulsor axis parallel to the first propulsor axis. The first engine core and the second engine core are mounted at an angle relative to corresponding ones of the first and second propulsor axes.

An aircraft includes a propulsion supported within an aft portion of a fuselage A thrust reverser is mounted in the aft portion of the fuselage proximate the propulsion system for directing thrust in a direction to slow the aircraft. The thrust reverser includes an upper blocker door movable about a first pivot axis to a deployed position and a lower blocker door movable about a second pivot axis not parallel to the first pivot axis.

An engine system has a gas generator, a bi-fi wall surrounding at least a portion of the gas generator, a casing surrounding a fan, and the casing having first and second thrust reverser doors which in a deployed position abut each other and the bi-fi wall.

An aircraft includes a fuselage including a propulsion system supported within an aft portion. A thrust reverser is mounted proximate to the propulsion system for directing thrust in a direction to slow the aircraft. The thrust reverser directs thrust at an angle relative to a vertical plane to reduce interference on control surfaces and reduce generation of underbody lift.

A gas turbine engine has a fairing and an air intake that includes an air inlet embedded within the fairing for supplying free stream atmospheric air to a gas generator.

An aircraft includes a propulsor supported within an aft portion of the fuselage. A thrust reverser is supported proximate the propulsor for redirecting thrust forward to slow the aircraft upon landing. A tail extending from the aft portion of the fuselage is angled forward away from the aft portion and out of the discharge of airflow from the thrust reverser.

A gas generator for a reverse core engine propulsion system has a variable cycle intake for the gas generator, which variable cycle intake includes a duct system. The duct system is configured for being selectively disposed in a first position and a second position, wherein free stream air is fed to the gas generator when in the first position, and fan stream air is fed to the gas generator when in the second position.

A gas turbine engine has a core engine incorporating a core engine turbine. A fan rotor is driven by a fan rotor turbine. The fan rotor turbine is in the path of gases downstream from the core engine turbine. A bypass door is moveable from a closed position at which the gases from the core engine turbine pass over the fan rotor turbine, and moveable to a bypass position at which the gases are directed away from the fan rotor turbine. An aircraft is also disclosed.

A gas turbine engine provides a differential gear system coupling the turbine to the bypass fan and the compressor. In this manner, the power/speed split between the bypass fan and the compressor can be optimized under all conditions. In the example shown, the turbine drives a sun gear, which drives a planet carrier and a ring gear in a differential manner. One of the planet carrier and the ring gear is coupled to the bypass fan, while the other is coupled to the compressor.

A blade for a turbine engine having a centerline. The blade comprises: a root section extending at an angle relative to the centerline; and an airfoil section extending from the root section. The root section is directly adjacent said airfoil section. In other words, the blade is neckless. The blade is part of a rotor assembly, and is preferably a fan blade.

We report on measurements of the angular stability, commonly called 'jitter', of the STS Orbiter during normal operations in space. Measurements were carried out by measuring optically the Orbiter's roll and pitch orientation relative to the solar vector as the orbiter was held in a -Z(sub 0) solar inertial orientation (orbiter bay oriented toward the Sun). We also report observations of an interesting perturbation to the orbiter's orientation noted by the crew during the STS-60 mission. These data may be useful in analyzing the in-orbit response of the Orbiter to thruster firings and other applied torques, and may aid in the planning of future experiments that require fine-pointed operations by the orbiter.

The Orbiter Stability Experiment (OSE) was developed to evaluate the steadiness of the STS Orbiter as a potential platform for instrumentation that would image the Sun in its extreme ultraviolet and soft X-ray radiations. We were interested in any high frequency motions of the Orbiter's orientation due to normal operations and manned activities. Preliminary results are presented of the observations. Other than the expected slow motion of the Orbiter within the specified angular deadband of 0.1 degrees during the observations, it was found that high frequency (above 1 Hz) angular motions (jitter) were not detectable at the 0.25 arc sec detection limit of the most sensitive detector, for most of the period of observation. No high frequency motions were recorded during intervals that were identified with vernier thruster firings. However, one short interval with detectable spectral power to a frequency of 10 Hz has been found to date. It has not yet been correlated with a particular activity going on at the time. The results of the observations may also be of value in assessing perturbations to the Orbiter's micro-gravity environment produced by normal operations.

An extreme ultraviolet (EUV) imaging spectrograph for the wavelength range from 235 to 450 A has been developed and used for resolution observations of the sun. The instrument incorporates a glancing incidence Wolter Type II Telescope and a near-normal incidence toroidal grating spectrograph to achieve near-stigmatic performance over this spectral range. The design of the spectrograph entrance aperture enables both stigmatic spectra with spectral resolution adequate to observe emission line profiles and spectroheliograms of restricted portions of the sun to be obtained concurrently. The design and performance of the instrument are described, and an overview of results obtained during a sounding rocket flight on May 5, 1989 is provided.

Research conducted by Tadanobu Tsunoda on auditory and visual sensation, designed to test and understand the functions of the cerebral hemispheres, is discussed. Tsunoda discovered that the Japanese responses to sounds by the left and the right sides of the brain are very different from the responses obtained from people from other countries. His tests show that a Japanese left cerebral hemisphere receives a wide range of sounds: linguistics sounds (vowel and consonant sounds); and nonlinguistic sounds, including cries, emotional utterances, traditional Japanese musical instruments, sounds of wind, waves, and creeks. In comparison, the range of sounds received by the left hemisphere of people from western countries is narrower, apparently limited to syllables made up of consonant and vowel sounds. On the other hand, the right hemisphere of a Japanese receives a very limited amount of artificial or man-made sounds: mechanical sounds, Western musical instruments, and noise. Japanese living outside of Japan where Western languages are spoken do not show the typical Japanese pattern after the second or third generation. However, the children of Americans or Koreans who are exposed from their birth to a complete Japanese environment have the Japanese brain functioning system. It appears that the uniqueness of the Japanese brain functions comes from the influences of the environment (SW)

Suggestopedia is a form of educational dynamics based on the insights of suggestology. The process focuses on the whole person, and in so doing, it takes into account the conscious/unconscious duality of the student. In accordance with the principles of suggestology, the process of suggestopedia is conducted at three levels, the psychological, the didactic, and the artistic. While creating a state of psychological relaxation, it uses every possible suggestive interrelationship to overcome psychological resistance and to develop new mental behavior. The main principles of suggestopedic teaching are pseudo-passivity, synthesis, and suggestive inter-relation. The principal suggestive factors that can break down anti-suggestive barriers are prestige or authority, verbal and non-verbal levels in acts of communication, and the integration of the arts, especially music. In language courses based on this method, the initial aim is to enable students to express themselves to a degree which corresponds to their desires and capabilities, thus making them more comfortable. The suggestopedic manual helps to create the conditions for the learning process and for the suggestive atmosphere. It is a question of recreating in the classroom a situational context which the students can live and enjoy, and in which they can express themselves. (AMH)