• ### Fast Radio Bursts from the collapse of Strange Star Crusts(1805.04448)

May 11, 2018 astro-ph.HE
Fast radio bursts (FRBs) are transient radio sources at cosmological distances. No counterparts in other bands have been observed for { non-repeating FRBs}. Here we suggest the collapse of strange star crusts as a possible origin for FRBs. Strange stars, which are composed of almost equal numbers of u, d, and s quarks, may be encapsulated by a thin crust of normal hadronic matter. When a strange star accretes matter from its environment, the crust becomes heavier and heavier. It may finally collapse, leading to the release of a large amount of magnetic energy and plenty of electron/positron pairs on a very short timescale. Electron/positron pairs in the polar cap region of the strange star can be accelerated to relativistic velocities, streaming along the magnetic field lines to form a thin shell. FRBs are produced by coherent emission from these electrons when the shell is expanding. Basic characteristics of observed FRBs can be explained in our model.
• ### Sentence-State LSTM for Text Representation(1805.02474)

May 7, 2018 cs.CL, cs.LG, stat.ML
Bi-directional LSTMs are a powerful tool for text representation. On the other hand, they have been shown to suffer various limitations due to their sequential nature. We investigate an alternative LSTM structure for encoding text, which consists of a parallel state for each word. Recurrent steps are used to perform local and global information exchange between words simultaneously, rather than incremental reading of a sequence of words. Results on various classification and sequence labelling benchmarks show that the proposed model has strong representation power, giving highly competitive performances compared to stacked BiLSTM models with similar parameter numbers.
• ### A Graph-to-Sequence Model for AMR-to-Text Generation(1805.02473)

May 7, 2018 cs.CL
The problem of AMR-to-text generation is to recover a text representing the same meaning as an input AMR graph. The current state-of-the-art method uses a sequence-to-sequence model, leveraging LSTM for encoding a linearized AMR structure. Although being able to model non-local semantic information, a sequence LSTM can lose information from the AMR graph structure, and thus faces challenges with large graphs, which result in long sequences. We introduce a neural graph-to-sequence model, using a novel LSTM structure for directly encoding graph-level semantics. On a standard benchmark, our model shows superior results to existing methods in the literature.
• ### Chinese NER Using Lattice LSTM(1805.02023)

May 5, 2018 cs.CL
We investigate a lattice-structured LSTM model for Chinese NER, which encodes a sequence of input characters as well as all potential words that match a lexicon. Compared with character-based methods, our model explicitly leverages word and word sequence information. Compared with word-based methods, lattice LSTM does not suffer from segmentation errors. Gated recurrent cells allow our model to choose the most relevant characters and words from a sentence for better NER results. Experiments on various datasets show that lattice LSTM outperforms both word-based and character-based LSTM baselines, achieving the best results.
• ### A Rank-Based Similarity Metric for Word Embeddings(1805.01923)

May 4, 2018 cs.CL
Word Embeddings have recently imposed themselves as a standard for representing word meaning in NLP. Semantic similarity between word pairs has become the most common evaluation benchmark for these representations, with vector cosine being typically used as the only similarity metric. In this paper, we report experiments with a rank-based metric for WE, which performs comparably to vector cosine in similarity estimation and outperforms it in the recently-introduced and challenging task of outlier detection, thus suggesting that rank-based measures can improve clustering quality.
• ### Domain Wall Motion Driven by Laplace Pressure in CoFeB-MgO Nanodots with Perpendicular Anisotropy(1711.11334)

We have studied the magnetization reversal of CoFeB-MgO nanodots with perpendicular anisotropy for size ranging from w=400 nm to 1 {\mu}m. Contrary to previous experiments, the switching field distribution is shifted toward lower magnetic fields as the size of the elements is reduced with a mean switching field varying as 1/w. We show that this mechanism can be explained by the nucleation of a pinned magnetic domain wall (DW) at the edges of the nanodots where damages are introduced by the patterning process. As the surface tension (Laplace pressure) applied on the DW increases when reducing the size of the nanodots, we demonstrate that the depinning field to reverse the entire elements varies as 1/w. These results suggest that the presence of DWs has to be considered in the switching process of nanoscale elements and open a path toward scalable spintronic devices.
• ### Learning How to Self-Learn: Enhancing Self-Training Using Neural Reinforcement Learning(1804.05734)

April 16, 2018 cs.CL
Self-training is a useful strategy for semi-supervised learning, leveraging raw texts for enhancing model performances. Traditional self-training methods depend on heuristics such as model confidence for instance selection, the manual adjustment of which can be expensive. To address these challenges, we propose a deep reinforcement learning method to learn the self-training strategy automatically. Based on neural network representation of sentences, our model automatically learns an optimal policy for instance selection. Experimental results show that our approach outperforms the baseline solutions in terms of better tagging performances and stability.
• ### YEDDA: A Lightweight Collaborative Text Span Annotation Tool(1711.03759)

March 22, 2018 cs.CL
In this paper, we introduce YEDDA, a lightweight but efficient and comprehensive open-source tool for text span annotation. YEDDA provides a systematic solution for text span annotation, ranging from collaborative user annotation to administrator evaluation and analysis. It overcomes the low efficiency of traditional text annotation tools by annotating entities through both command line and shortcut keys, which are configurable with custom labels. YEDDA also gives intelligent recommendations by learning the up-to-date annotated text. An administrator client is developed to evaluate annotation quality of multiple annotators and generate detailed comparison report for each annotator pair. Experiments show that the proposed system can reduce the annotation time by half compared with existing annotation tools. And the annotation time can be further compressed by 16.47\% through intelligent recommendation.
• ### Experimental investigation of coaxial-gun-formed plasmas injected into a background transverse magnetic field or plasma(1712.05829)

March 21, 2018 physics.plasm-ph
Injection of coaxial-gun-formed magnetized plasmas into a background transverse vacuum magnetic field or into a background magnetized plasma has been studied in the helicon-cathode (HelCat) linear plasma device at the University of New Mexico [M. Gilmore et al., J. Plasma Phys.81, 345810104 (2015)]. Magnetized plasma jet launched into a background transverse magnetic field shows emergent kink stabilization of the jet due to the formation of a sheared flow in the jet above the kink-stabilization threshold $0.1kV_A$ [Y. Zhang et al., Phys. Plasmas 24, 110702 (2017)]. Injection of a spheromak-like plasma into a transverse background magnetic field led to the observation of finger-like structures on the side with a stronger magnetic field null between the spheromak and background field. The finger-like structures are consistent with magneto-Rayleigh-Taylor instability. Jets or spheromaks launched into a background, low-$\beta$ magnetized plasma show similar behavior as above, respectively, in both cases.
• ### The Penetration Effect of Connected Automated Vehicles in Urban Traffic: An Energy Impact Study(1803.05577)

March 15, 2018 math.OC
Earlier work has established a decentralized framework of optimally controlling connected and automated vehicles (CAVs) crossing an urban intersection without using explicit traffic signaling. The proposed solution is capable of minimizing energy consumption subject to a throughput maximization requirement. In this paper, we address the problem of optimally controlling CAVs under mixed traffic conditions where both CAVs and human-driven vehicles (non-CAVs) travel on the roads, so as to minimize energy consumption while guaranteeing safety constraints. The impact of CAVs on overall energy consumption is also investigated under different traffic scenarios. The benefit from CAV penetration (i.e., the fraction of CAVs relative to all vehicles) is validated through simulation in MATLAB and VISSIM. The results indicate that the energy efficiency improvement becomes more significant as the CAV penetration rate increases, while the significance diminishes as the traffic becomes heavier
• ### Skyrmions in magnetic tunnel junctions(1803.05138)

In this work, we demonstrate that skyrmions can be nucleated in the free layer of a magnetic tunnel junction (MTJ) with Dzyaloshinskii-Moriya interactions (DMI) by a spin-polarized current with the assistance of stray fields from the pinned layer. The size, stability and number of created skyrmions can be tuned by either the DMI strength or the stray field distribution. The interaction between the stray field and the DMI effective field is discussed. A device with multi-level tunneling magnetoresistance is proposed, which could pave the ways for skyrmion-MTJ-based multi-bit storage and artificial neural network computation. Our results may facilitate the efficient nucleation and electrical detection of skyrmions.
• ### Ultra-Fast Ferrimagnetic All Spin Logic Device(1803.02552)

March 7, 2018 physics.app-ph
All spin logic device (ASLD) blazes an alternative path for realizing ultra-low power computing in the Post-Moore era. However, initial device structure relying on ferromagnetic input/output and spin transfer torque (STT) driven magnetization switching degrades its performance and even hinders its realization. In this paper, we propose an ASLD based on rare-earth (RE)-transition-metal (TM) ferromagnetic alloy that can achieve an ultra-high frequency up to terahertz. The spin orbit torque (SOT) induced fast precession near the spin angular momentum compensated point is investigated through the macrospin model. Combining the non-local spin current diffusing from the input to the output, a deterministic picosecond switching can be realized without any external magnetic field. Our results show that ASLD has the potential to exceed the performance of mainstream computing.
• ### Lepton Flavorful Fifth Force and Depth-dependent Neutrino Matter Interactions(1803.00591)

March 1, 2018 hep-ph, hep-ex
We consider a fifth force to be an interaction that couples to matter with a strength that grows with the number of atoms. In addition to competing with the strength of gravity a fifth force can give rise to violations of the equivalence principle. Current long range constraints on the strength and range of fifth forces are very impressive. Amongst possible fifth forces are those that couple to lepton flavorful charges $L_e-L_{\mu}$ or $L_e-L_{\tau}$. They have the property that their range and strength are also constrained by neutrino interactions with matter. In this brief note we review the existing constraints on the allowed parameter space in gauged $U(1)_{L_e-L_{\mu}, L_{\tau}}$. We find two regions where neutrino oscillation experiments are at the frontier of probing such a new force. In particular, there is an allowed range of parameter space where neutrino matter interactions relevant for long baseline oscillation experiments depend on the depth of the neutrino beam below the surface of the earth.
• ### Supermassive Black Holes with High Accretion Rates in Active Galactic Nuclei. IX 10 New Observations of Reverberation Mapping and Shortened H$\beta$ Lags(1802.03022)

Feb. 8, 2018 astro-ph.GA
As one of the series of papers reporting on a large reverberation mapping campaign of super-Eddington accreting massive black holes (SEAMBHs) in active galactic nuclei (AGNs), we present the results of 10 SEAMBHs monitored spectroscopically during 2015-2017. Six of them are observed for the first time, and have generally higher 5100 \AA\ luminosities than the SEAMBHs monitored in our campaign from 2012 to 2015; the remaining four are repeat observations to check if their previous lags change. Similar to the previous SEAMBHs, the H$\beta$ time lags of the newly observed objects are shorter than the values predicted by the canonical $R_{\mathrm{H\beta}}$-$L_{5100}$ relation of sub-Eddington AGNs, by factors of $\sim2-6$, depending on the accretion rate. The four previously observed objects have lags consistent with previous measurements. We provide linear regressions for the $R_{\mathrm{H\beta}}$-$L_{5100}$ relation, solely for the SEAMBH sample and for low-accretion AGNs. We find that the relative strength of Fe II and the profile of the H$\beta$ emission line can be used as proxies of accretion rate, showing that the shortening of H$\beta$ lags depends on accretion rates. The recent SDSS-RM discovery of shortened H$\beta$ lags in AGNs with low accretion rates provides compelling evidence for retrograde accretion onto the black hole. These evidences show that the canonical $R_{\mathrm{H\beta}}$-$L_{5100}$ relation holds only in AGNs with moderate accretion rates. At low accretion rates, it should be revised to include the effects of black hole spin, whereas the accretion rate itself becomes a key factor in the regime of high accretion rates.
• ### Applying Cooperative Machine Learning to Speed Up the Annotation of Social Signals in Large Multi-modal Corpora(1802.02565)

Feb. 7, 2018 cs.AI, cs.HC, cs.LG, stat.ML
Scientific disciplines, such as Behavioural Psychology, Anthropology and recently Social Signal Processing are concerned with the systematic exploration of human behaviour. A typical work-flow includes the manual annotation (also called coding) of social signals in multi-modal corpora of considerable size. For the involved annotators this defines an exhausting and time-consuming task. In the article at hand we present a novel method and also provide the tools to speed up the coding procedure. To this end, we suggest and evaluate the use of Cooperative Machine Learning (CML) techniques to reduce manual labelling efforts by combining the power of computational capabilities and human intelligence. The proposed CML strategy starts with a small number of labelled instances and concentrates on predicting local parts first. Afterwards, a session-independent classification model is created to finish the remaining parts of the database. Confidence values are computed to guide the manual inspection and correction of the predictions. To bring the proposed approach into application we introduce NOVA - an open-source tool for collaborative and machine-aided annotations. In particular, it gives labellers immediate access to CML strategies and directly provides visual feedback on the results. Our experiments show that the proposed method has the potential to significantly reduce human labelling efforts.
• ### Lepton-Number-Charged Scalars and Neutrino Beamstrahlung(1802.00009)

Jan. 31, 2018 hep-ph, hep-ex
Experimentally, baryon number minus lepton number, $B-L$, appears to be a good global symmetry of nature. We explore the consequences of the existence of gauge-singlet scalar fields charged under $B-L$ -- dubbed lepton-number-charged scalars, LeNCS -- and postulate that these couple to the standard model degrees of freedom in such a way that $B-L$ is conserved even at the non-renormalizable level. In this framework, neutrinos are Dirac fermions. Including only the lowest mass-dimension effective operators, some of the LeNCS couple predominantly to neutrinos and may be produced in terrestrial neutrino experiments. We examine several existing constraints from particle physics, astrophysics, and cosmology to the existence of a LeNCS carrying $B-L$ charge equal to two, and discuss the emission of LeNCS's via "neutrino beamstrahlung," which occurs every once in a while when neutrinos scatter off of ordinary matter. We identify regions of the parameter space where existing and future neutrino experiments, including the Deep Underground Neutrino Experiment, are at the frontier of searches for such new phenomena.
• ### Improved English to Russian Translation by Neural Suffix Prediction(1801.03615)

Jan. 11, 2018 cs.CL
Neural machine translation (NMT) suffers a performance deficiency when a limited vocabulary fails to cover the source or target side adequately, which happens frequently when dealing with morphologically rich languages. To address this problem, previous work focused on adjusting translation granularity or expanding the vocabulary size. However, morphological information is relatively under-considered in NMT architectures, which may further improve translation quality. We propose a novel method, which can not only reduce data sparsity but also model morphology through a simple but effective mechanism. By predicting the stem and suffix separately during decoding, our system achieves an improvement of up to 1.98 BLEU compared with previous work on English to Russian translation. Our method is orthogonal to different NMT architectures and stably gains improvements on various domains.
• ### Top-philic $Z'$ Forces at the LHC(1801.03505)

Jan. 10, 2018 hep-ph, hep-ex
Despite extensive searches for an additional neutral massive gauge boson at the LHC, a $Z^\prime$ at the weak scale could still be present if its couplings to the first two generations of quarks are suppressed, in which case the production in hadron colliders relies on tree-level processes in association with heavy flavors or one-loop processes in association with a jet. We consider the low-energy effective theory of a top-philic $Z'$ and present possible UV completions. We clarify theoretical subtleties in evaluating the production of a top-philic $Z'$ at the LHC and examine carefully the treatment of an anomalous $Z'$ current in the low-energy effective theory. Recipes for properly computing the production rate in the $Z'+j$ channel are given. We discuss constraints from colliders and low-energy probes of new physics. As an application, we apply these considerations to models that use a weak-scale $Z'$ to explain possible violations of lepton universality in $B$ meson decays, and show that the future running of a high luminosity LHC can potentially cover much of the remaining parameter space favored by this particular interpretation of the $B$ physics anomaly.
• ### Insight-HXMT observations of the first binary neutron star merger GW170817(1710.06065)

Oct. 17, 2017 astro-ph.HE
Finding the electromagnetic (EM) counterpart of binary compact star merger, especially the binary neutron star (BNS) merger, is critically important for gravitational wave (GW) astronomy, cosmology and fundamental physics. On Aug. 17, 2017, Advanced LIGO and \textit{Fermi}/GBM independently triggered the first BNS merger, GW170817, and its high energy EM counterpart, GRB 170817A, respectively, resulting in a global observation campaign covering gamma-ray, X-ray, UV, optical, IR, radio as well as neutrinos. The High Energy X-ray telescope (HE) onboard \textit{Insight}-HXMT (Hard X-ray Modulation Telescope) is the unique high-energy gamma-ray telescope that monitored the entire GW localization area and especially the optical counterpart (SSS17a/AT2017gfo) with very large collection area ($\sim$1000 cm$^2$) and microsecond time resolution in 0.2-5 MeV. In addition, \textit{Insight}-HXMT quickly implemented a Target of Opportunity (ToO) observation to scan the GW localization area for potential X-ray emission from the GW source. Although it did not detect any significant high energy (0.2-5 MeV) radiation from GW170817, its observation helped to confirm the unexpected weak and soft nature of GRB 170817A. Meanwhile, \textit{Insight}-HXMT/HE provides one of the most stringent constraints (~10$^{-7}$ to 10$^{-6}$ erg/cm$^2$/s) for both GRB170817A and any other possible precursor or extended emissions in 0.2-5 MeV, which help us to better understand the properties of EM radiation from this BNS merger. Therefore the observation of \textit{Insight}-HXMT constitutes an important chapter in the full context of multi-wavelength and multi-messenger observation of this historical GW event.
• ### Dark Matter and Neutrino Mass from the Smallest Non-Abelian Chiral Dark Sector(1706.02722)

Oct. 16, 2017 hep-ph
All pieces of concrete evidence for phenomena outside the standard model (SM) - neutrino masses and dark matter - are consistent with the existence of new degrees of freedom that interact very weakly, if at all, with those in the SM. We propose that these new degrees of freedom organize themselves into a simple dark sector, a chiral SU(3) x SU(2) gauge theory with the smallest nontrivial fermion content. Similar to the SM, the dark SU(2) is spontaneously broken while the dark SU(3) confines at low energies. At the renormalizable level, the dark sector contains massless fermions - dark leptons - and stable massive particles - dark protons. We find that dark protons with masses between 10-100 TeV satisfy all current cosmological and astrophysical observations concerning dark matter even if dark protons are a symmetric thermal relic. The dark leptons play the role of right-handed neutrinos and allow simple realizations of the seesaw mechanism or the possibility that neutrinos are Dirac fermions. In the latter case, neutrino masses are also parametrically different from charged-fermion masses and the lightest neutrino is predicted to be massless. Since the new "neutrino" and "dark matter" degrees of freedom interact with one another, these two new-physics phenomena are intertwined. Dark leptons play a nontrivial role in early universe cosmology while indirect searches for dark matter involve, decisively, dark matter annihilations into dark leptons. These, in turn, may lead to observable signatures at high-energy neutrino and gamma-ray observatories, especially once one accounts for the potential Sommerfeld enhancement of the annihilation cross-section, derived from the low-energy dark-sector effective theory, a possibility we explore quantitatively in some detail.
• ### A Cosmic Selection Rule for Glueball Dark Matter Relic Density(1704.02347)

Oct. 16, 2017 hep-ph, astro-ph.CO
We point out a unique mechanism to produce the relic abundance for glueball dark matter from a gauged $SU(N)_d$ hidden sector which is bridged to the standard model sector through heavy vectorlike quarks colored under gauge interactions from both sides. A necessary ingredient of our assumption is that the vectorlike quarks, produced either thermally or non-thermally, are abundant enough to dominate the universe for some time in the early universe. They later undergo dark color confinement and form unstable vectorlike-quarkonium states which annihilate decay and reheat the visible and dark sectors. The ratio of entropy dumped into two sectors and the final energy budget in the dark glueballs is only determined by low energy parameters, including the intrinsic scale of the dark $SU(N)_d$, $\Lambda_d$, and number of dark colors, $N_d$, but depend weakly on parameters in the ultraviolet such as the vectorlike quark mass or the initial condition. We call this a cosmic selection rule for the glueball dark matter relic density.
• ### Demonstration of Laser-produced Neutron Diagnostic by Radiative Capture Gamma-rays(1709.06561)

Sept. 19, 2017 nucl-ex, physics.ins-det
We report a new scenario of time-of-flight (TOF) technique in which fast neutrons and delayed gamma-ray signals were both recorded in a millisecond time window in harsh environments induced by high-intensity lasers. The delayed gamma signals, arriving far later than the original fast neutron and often being ignored previously, were identified to be the results of radiative captures of thermalized neutrons. The linear correlation between gamma photon number and the fast neutron yield shows that these delayed gamma events can be employed for neutron diagnosis. This method can reduce the detecting efficiency dropping problem caused by prompt high-flux gamma radiation, and provides a new way for neutron diagnosing in high-intensity laser-target interaction experiments.
• ### Decentralized Optimal Control for Connected Automated Vehicles at Intersections Including Left and Right Turns(1703.06956)

Sept. 12, 2017 math.OC
In prior work, we addressed the problem of optimally controlling on line connected and automated vehicles crossing two adjacent intersections in an urban area to minimize fuel consumption while achieving maximal throughput without any explicit traffic signaling and without considering left and right turns. In this paper, we extend the solution of this problem to account for left and right turns under hard safety constraints. Furthermore, we formulate and solve another optimization problem to minimize a measure of passenger discomfort while the vehicle turns at the intersection and we investigate the associated tradeoff between minimizing fuel consumption and passenger discomfort.
• ### Combining Discrete and Neural Features for Sequence Labeling(1708.07279)

Aug. 24, 2017 cs.CL
Neural network models have recently received heated research attention in the natural language processing community. Compared with traditional models with discrete features, neural models have two main advantages. First, they take low-dimensional, real-valued embedding vectors as inputs, which can be trained over large raw data, thereby addressing the issue of feature sparsity in discrete models. Second, deep neural networks can be used to automatically combine input features, and including non-local features that capture semantic patterns that cannot be expressed using discrete indicator features. As a result, neural network models have achieved competitive accuracies compared with the best discrete models for a range of NLP tasks. On the other hand, manual feature templates have been carefully investigated for most NLP tasks over decades and typically cover the most useful indicator pattern for solving the problems. Such information can be complementary the features automatically induced from neural networks, and therefore combining discrete and neural features can potentially lead to better accuracy compared with models that leverage discrete or neural features only. In this paper, we systematically investigate the effect of discrete and neural feature combination for a range of fundamental NLP tasks based on sequence labeling, including word segmentation, POS tagging and named entity recognition for Chinese and English, respectively. Our results on standard benchmarks show that state-of-the-art neural models can give accuracies comparable to the best discrete models in the literature for most tasks and combing discrete and neural features unanimously yield better results.
• ### In-Order Transition-based Constituent Parsing(1707.05000)

July 17, 2017 cs.CL
Both bottom-up and top-down strategies have been used for neural transition-based constituent parsing. The parsing strategies differ in terms of the order in which they recognize productions in the derivation tree, where bottom-up strategies and top-down strategies take post-order and pre-order traversal over trees, respectively. Bottom-up parsers benefit from rich features from readily built partial parses, but lack lookahead guidance in the parsing process; top-down parsers benefit from non-local guidance for local decisions, but rely on a strong encoder over the input to predict a constituent hierarchy before its construction.To mitigate both issues, we propose a novel parsing system based on in-order traversal over syntactic trees, designing a set of transition actions to find a compromise between bottom-up constituent information and top-down lookahead information. Based on stack-LSTM, our psycholinguistically motivated constituent parsing system achieves 91.8 F1 on WSJ benchmark. Furthermore, the system achieves 93.6 F1 with supervised reranking and 94.2 F1 with semi-supervised reranking, which are the best results on the WSJ benchmark.