arXiv subjects

Jian Tang

Publications and source records attributed to Jian Tang.

At least 379 records · Page 21Linked to original sources

arXiv2017

PixelSNE: Visualizing Fast with Just Enough Precision via Pixel-Aligned Stochastic Neighbor Embedding

Embedding and visualizing large-scale high-dimensional data in a two-dimensional space is an important problem since such visualization can reveal deep insights out of complex data. Most of the existing embedding approaches, however, run on an excessively high precision, ignoring the fact that at the end, embedding outputs are converted into coarse-grained discrete pixel coordinates in a screen space. Motivated by such an observation and directly considering pixel coordinates in an embedding optimization process, we accelerate Barnes-Hut tree-based t-distributed stochastic neighbor embedding (BH-SNE), known as a state-of-the-art 2D embedding method, and propose a novel method called PixelSNE, a highly-efficient, screen resolution-driven 2D embedding method with a linear computational complexity in terms of the number of data items. Our experimental results show the significantly fast running time of PixelSNE by a large margin against BH-SNE, while maintaining the minimal degradation in the embedding quality. Finally, the source code of our method is publicly available at https://github.com/awesome-davian/PixelSNE

cs.LG↗

arXiv2016

Less is More: Learning Prominent and Diverse Topics for Data Summarization

Statistical topic models efficiently facilitate the exploration of large-scale data sets. Many models have been developed and broadly used to summarize the semantic structure in news, science, social media, and digital humanities. However, a common and practical objective in data exploration tasks is not to enumerate all existing topics, but to quickly extract representative ones that broadly cover the content of the corpus, i.e., a few topics that serve as a good summary of the data. Most existing topic models fit exactly the same number of topics as a user specifies, which have imposed an unnecessary burden to the users who have limited prior knowledge. We instead propose new models that are able to learn fewer but more representative topics for the purpose of data summarization. We propose a reinforced random walk that allows prominent topics to absorb tokens from similar and smaller topics, thus enhances the diversity among the top topics extracted. With this reinforced random walk as a general process embedded in classical topic models, we obtain \textit{diverse topic models} that are able to extract the most prominent and diverse topics from data. The inference procedures of these diverse topic models remain as simple and efficient as the classical models. Experimental results demonstrate that the diverse topic models not only discover topics that better summarize the data, but also require minimal prior knowledge of the users.

cs.LG↗

arXiv2016

Identity-sensitive Word Embedding through Heterogeneous Networks

Most existing word embedding approaches do not distinguish the same words in different contexts, therefore ignoring their contextual meanings. As a result, the learned embeddings of these words are usually a mixture of multiple meanings. In this paper, we acknowledge multiple identities of the same word in different contexts and learn the \textbf{identity-sensitive} word embeddings. Based on an identity-labeled text corpora, a heterogeneous network of words and word identities is constructed to model different-levels of word co-occurrences. The heterogeneous network is further embedded into a low-dimensional space through a principled network embedding approach, through which we are able to obtain the embeddings of words and the embeddings of word identities. We study three different types of word identities including topics, sentiments and categories. Experimental results on real-world data sets show that the identity-sensitive word embeddings learned by our approach indeed capture different meanings of words and outperforms competitive methods on tasks including text classification and word similarity computation.

cs.CL↗

arXiv2016

Context-aware Natural Language Generation with Recurrent Neural Networks

This paper studied generating natural languages at particular contexts or situations. We proposed two novel approaches which encode the contexts into a continuous semantic representation and then decode the semantic representation into text sequences with recurrent neural networks. During decoding, the context information are attended through a gating mechanism, addressing the problem of long-range dependency caused by lengthy sequences. We evaluate the effectiveness of the proposed approaches on user review data, in which rich contexts are available and two informative contexts, sentiments and products, are selected for evaluation. Experiments show that the fake reviews generated by our approaches are very natural. Results of fake review detection with human judges show that more than 50\% of the fake reviews are misclassified as the real reviews, and more than 90\% are misclassified by existing state-of-the-art fake review detection algorithm.

cs.CL↗

arXiv2016

PandaX-III: Searching for Neutrinoless Double Beta Decay with High Pressure $^{136}$Xe Gas Time Projection Chambers

Searching for the Neutrinoless Double Beta Decay (NLDBD) is now regarded as the topmost promising technique to explore the nature of neutrinos after the discovery of neutrino masses in oscillation experiments. PandaX-III (Particle And Astrophysical Xenon Experiment III) will search for the NLDBD of $^{136}$Xe at the China Jin Ping underground Laboratory (CJPL). In the first phase of the experiment, a high pressure gas Time Projection Chamber (TPC) will contain 200 kg, 90% $^{136}$Xe enriched gas operated at 10 bar. Fine pitch micro-pattern gas detector (Microbulk Micromegas) will be used at both ends of the TPC for the charge readout with a cathode in the middle. Charge signals can be used to reconstruct tracks of NLDBD events and provide good energy and spatial resolution. The detector will be immersed in a large water tank to ensure $\sim$5 m of water shielding in all directions. The second phase, a ton-scale experiment, will consist of five TPCs in the same water tank, with improved energy resolution and better control over backgrounds.

physics.ins-det↗

arXiv2016

Letter of Intent: Jinping Neutrino Experiment

Jinping Neutrino Experiment (Jinping) is proposed to significantly improve measurements on solar neutrinos and geoneutrinos in China Jinping Laboratory - a lab with a number of unparalleled features, thickest overburden, lowest reactor neutrino background, etc., which identify it as the world-best low-energy neutrino laboratory. The proposed experiment will have target mass of 4 kilotons of liquid scintillator or water-based liquid scintillator, with a fiducial mass of 2 kilotons for neutrino-electron scattering events and 3 kilotons for inverse-beta interaction events. A number of initial sensitivities studies have been carried out, including on the transition phase for the solar neutrinos oscillation from the vacuum to the matter effect, the discovery of solar neutrinos from the carbon-nitrogen-oxygen (CNO) cycle, the resolution of the high and low metallicity hypotheses, and the unambiguous separation on U and Th cascade decays from the dominant crustal anti-electron neutrinos in China.

physics.ins-det↗

arXiv2016

Visualizing Large-scale and High-dimensional Data

We study the problem of visualizing large-scale and high-dimensional data in a low-dimensional (typically 2D or 3D) space. Much success has been reported recently by techniques that first compute a similarity structure of the data points and then project them into a low-dimensional space with the structure preserved. These two steps suffer from considerable computational costs, preventing the state-of-the-art methods such as the t-SNE from scaling to large-scale and high-dimensional data (e.g., millions of data points and hundreds of dimensions). We propose the LargeVis, a technique that first constructs an accurately approximated K-nearest neighbor graph from the data and then layouts the graph in the low-dimensional space. Comparing to t-SNE, LargeVis significantly reduces the computational cost of the graph construction step and employs a principled probabilistic model for the visualization step, the objective of which can be effectively optimized through asynchronous stochastic gradient descent with a linear time complexity. The whole procedure thus easily scales to millions of high-dimensional data points. Experimental results on real-world data sets demonstrate that the LargeVis outperforms the state-of-the-art methods in both efficiency and effectiveness. The hyper-parameters of LargeVis are also much more stable over different data sets.

cs.LG↗

arXiv2016

Energy-Efficient Power Allocation in Cognitive Radio Systems with Imperfect Spectrum Sensing

This paper studies energy-efficient power allocation schemes for secondary users in sensing-based spectrum sharing cognitive radio systems. It is assumed that secondary users first perform channel sensing possibly with errors and then initiate data transmission with different power levels based on sensing decisions. The circuit power is taken into account in total power consumption. In this setting, the optimization problem is to maximize energy efficiency (EE) subject to peak/average transmission power constraints and peak/average interference constraints. By exploiting quasiconcave property of EE maximization problem, the original problem is transformed into an equivalent parameterized concave problem and an iterative power allocation algorithm based on Dinkelbach's method is proposed. The optimal power levels are identified in the presence of different levels of channel side information (CSI) regarding the transmission and interference links at the secondary transmitter, namely perfect CSI of both transmission and interference links, perfect CSI of the transmission link and imperfect CSI of the interference link, imperfect CSI of both links or only statistical CSI of both links. Through numerical results, the impact of sensing performance, different types of CSI availability, and transmit and interference power constraints on the EE of the secondary users is analyzed.

cs.IT↗

arXiv2015

PTE: Predictive Text Embedding through Large-scale Heterogeneous Text Networks

Unsupervised text embedding methods, such as Skip-gram and Paragraph Vector, have been attracting increasing attention due to their simplicity, scalability, and effectiveness. However, comparing to sophisticated deep learning architectures such as convolutional neural networks, these methods usually yield inferior results when applied to particular machine learning tasks. One possible reason is that these text embedding methods learn the representation of text in a fully unsupervised way, without leveraging the labeled information available for the task. Although the low dimensional representations learned are applicable to many different tasks, they are not particularly tuned for any task. In this paper, we fill this gap by proposing a semi-supervised representation learning method for text data, which we call the \textit{predictive text embedding} (PTE). Predictive text embedding utilizes both labeled and unlabeled data to learn the embedding of text. The labeled information and different levels of word co-occurrence information are first represented as a large-scale heterogeneous text network, which is then embedded into a low dimensional space through a principled and efficient algorithm. This low dimensional embedding not only preserves the semantic closeness of words and documents, but also has a strong predictive power for the particular task. Compared to recent supervised approaches based on convolutional neural networks, predictive text embedding is comparable or more effective, much more efficient, and has fewer parameters to tune.

cs.CL↗

arXiv2015

LINE: Large-scale Information Network Embedding

This paper studies the problem of embedding very large information networks into low-dimensional vector spaces, which is useful in many tasks such as visualization, node classification, and link prediction. Most existing graph embedding methods do not scale for real world information networks which usually contain millions of nodes. In this paper, we propose a novel network embedding method called the "LINE," which is suitable for arbitrary types of information networks: undirected, directed, and/or weighted. The method optimizes a carefully designed objective function that preserves both the local and global network structures. An edge-sampling algorithm is proposed that addresses the limitation of the classical stochastic gradient descent and improves both the effectiveness and the efficiency of the inference. Empirical experiments prove the effectiveness of the LINE on a variety of real-world information networks, including language networks, social networks, and citation networks. The algorithm is very efficient, which is able to learn the embedding of a network with millions of vertices and billions of edges in a few hours on a typical single machine. The source code of the LINE is available online.

cs.LG↗

arXiv2014

"Look Ma, No Hands!" A Parameter-Free Topic Model

It has always been a burden to the users of statistical topic models to predetermine the right number of topics, which is a key parameter of most topic models. Conventionally, automatic selection of this parameter is done through either statistical model selection (e.g., cross-validation, AIC, or BIC) or Bayesian nonparametric models (e.g., hierarchical Dirichlet process). These methods either rely on repeated runs of the inference algorithm to search through a large range of parameter values which does not suit the mining of big data, or replace this parameter with alternative parameters that are less intuitive and still hard to be determined. In this paper, we explore to "eliminate" this parameter from a new perspective. We first present a nonparametric treatment of the PLSA model named nonparametric probabilistic latent semantic analysis (nPLSA). The inference procedure of nPLSA allows for the exploration and comparison of different numbers of topics within a single execution, yet remains as simple as that of PLSA. This is achieved by substituting the parameter of the number of topics with an alternative parameter that is the minimal goodness of fit of a document. We show that the new parameter can be further eliminated by two parameter-free treatments: either by monitoring the diversity among the discovered topics or by a weak supervision from users in the form of an exemplar topic. The parameter-free topic model finds the appropriate number of topics when the diversity among the discovered topics is maximized, or when the granularity of the discovered topics matches the exemplar topic. Experiments on both synthetic and real data prove that the parameter-free topic model extracts topics with a comparable quality comparing to classical topic models with "manual transmission". The quality of the topics outperforms those extracted through classical Bayesian nonparametric models.

cs.LG↗

arXiv2013

Study on perturbation schemes for achieving the real PMNS matrix from various symmetric textures

The PMNS matrix displays an obvious symmetry, but not exact. There are several textures proposed in literature, which possess various symmetry patterns and seem to originate from different physics scenarios at high energy scales. To be consistent with the experimental measurement, all of the regularities slightly decline, i.e. the symmetry must be broken. Following the schemes given in literature, we modify the matrices (9 in total) to gain the real PMNS matrix by perturbative rotations. The transformations may provide hints about the underlying physics at high energies and the breaking mechanisms which apply during the evolution to the low energy scale, especially the results may be useful for the future model builders.

hep-ph↗

arXiv2012

Requirements for a New Detector at the South Pole Receiving an Accelerator Neutrino Beam

There are recent considerations to increase the photomultiplier density in the IceCube detector array beyond that of DeepCore, which will lead to a lower detection threshold and a huge fiducial mass for the neutrino detection. This initiative is known as "Phased IceCube Next Generation Upgrade" (PINGU). We discuss the possibility to send a neutrino beam from one of the major accelerator laboratories in the Northern hemisphere to such a detector. Such an experiment would be unique in the sense that it would be the only neutrino beam where the baseline crosses the Earth's core. We study the detector requirements for a beta beam, a neutrino factory beam, and a superbeam, where we consider both the cases of small theta_13 and large theta_13, as suggested by the recent T2K and Double Chooz results. We illustrate that a flavor-clean beta beam best suits the requirements of such a detector, in particular, that PINGU may replace a magic baseline detector for small values of theta_13 -- even in the absence of any energy resolution capability. For large theta_13, however, a single-baseline beta beam experiment cannot compete if it is constrained by the CERN-SPS. For a neutrino factory, because of the missing charge identification possibility in the detector, a very good energy resolution is required. If this can be achieved, especially a low energy neutrino factory, which does not suffer from the tau contamination, may be an interesting option for large theta_13. For the superbeam, where we use the LBNE beam as a reference, electron neutrino flavor identification and statistics are two of the main limitations. Finally, we demonstrate that, at least in principle, neutrino factory and superbeam can measure the density of the Earth's core to the sub-percent level for sin^2 2theta_13 larger than 0.01.

hep-ph↗

arXiv2011

Phenomenology of Neutrino Oscillations at the Neutrino Factory

We consider the prospects for a neutrino factory to measure mixing angles, the CP violating phase and mass-squared differences by detecting wrong-charged muons arising from the chain μ^+ to ν_e to ν_μ to μ^- and the right-charged muons coming from the chain μ^+ to \barν_μ to \barν_μ to μ^+ (similar to μ^- chains), where ν_e to ν_μ and \barν_μ to \barν_μ are neutrino oscillation channels through a long baseline. First, we perform the baseline and energy optimization of the neutrino factory including the latest simulation results from the magnetized iron neutrino detector (MIND). Second, we study physics with near detectors and consider the treatment of systematic errors including cross section errors, flux errors, and background uncertainties. Third, the effects of one additional massive sterile neutrino are investigated in the context of near and far detector combinations.

hep-ph↗

arXiv2011

Optimization of the Neutrino Factory, revisited

We perform the baseline and energy optimization of the Neutrino Factory including the latest simulation results on the magnetized iron detector (MIND). We also consider the impact of tau decays, generated by nu_mu to nu_tau or nu_e to nu_tau appearance, on the mass hierarchy, CP violation, and theta_{13} discovery reaches, which we find to be negligible for the considered detector. For the baseline-energy optimization for small theta_{13}, we qualitatively recover the results with earlier simulations of the MIND detector. We find optimal baselines of about 2500 km to 5000 km for the CP violation measurement, where now values of E_mu as low as about 12 GeV may be possible. However, for large theta_{13}, we demonstrate that the lower threshold and the backgrounds reconstructed at lower energies allow in fact for muon energies as low as 5 GeV at considerably shorter baselines, such as FNAL-Homestake. This implies that with the latest MIND analysis, low- and high-energy versions of the Neutrino Factory are just two different versions of the same experiment optimized for different parts of the parameter space. Apart from a green-field study of the updated detector performance, we discuss specific implementations for the two-baseline Neutrino Factory, where the considered detector sites are taken to be currently discussed underground laboratories. We find that reasonable setups can be found for the Neutrino Factory source in Asia, Europe, and North America, and that a triangular-shaped storage ring is possible in all cases based on geometrical arguments only.

hep-ph↗

arXiv2010

Sterile neutrinos beyond LSND at the Neutrino Factory

We discuss the effects of one additional sterile neutrino at the Neutrino Factory. Compared to earlier analyses, which have been motivated by LSND results, we do not impose any constraint on the additional mass squared splitting. This means that the additional mass eigenstate could, with small mixings, be located among the known ones, as it is suggested by the recent analysis of cosmological data. We use a self-consistent framework at the Neutrino Factory without any constraints on the new parameters. We demonstrate for a combined short and long baseline setup that near detectors can provide the expected sensitivity at the LSND-motivated Δm_{41}^2-range, while some sensitivity can also be obtained in the region of the atmospheric mass splitting from the long baselines. We point out that limits on such very light sterile neutrinos may also be obtained from a re-analysis of atmospheric and solar neutrino oscillation data, as well as from supernova neutrino observations. In the second part of the analysis, we compare our sensitivity with the existing literature using additional assumptions, such as |Δm_{41}^2| \gg |Δm_{31}^2| leading to averaging of the fast oscillations in the far detectors. We demonstrate that while the Neutrino Factory has excellent sensitivity compared to existing studies using similar assumptions, one has to be very careful interpreting these results for a combined short and long baseline setup where oscillations could occur in the near detectors. We also test the impact of additional ν_τdetectors at the short and long baselines, and we do not find a substantial improvement of the sensitivities.

hep-ph↗

arXiv2010

Neutrino factory in stages: Low energy, high energy, off-axis

We discuss neutrino oscillation physics with a neutrino factory in stages, including the possibility of upgrading the muon energy within the same program. We point out that a detector designed for the low energy neutrino factory may be used off-axis in a high energy neutrino factory beam. We include the re-optimization of the experiment depending on the value of theta_13 found. As upgrade options, we consider muon energy, additional baselines, a detector mass upgrade, an off-axis detector, and the platinum (muon to electron neutrino) channels. In addition, we test the impact of Daya Bay data on the optimization. We find that for large theta_13 (theta_13 discovered by the next generation of experiments), a low energy neutrino factory might be the most plausible minimal version to test the unknown parameters. However, if a higher muon energy is needed for new physics searches, a high energy version including an off-axis detector may be an interesting alternative. For small theta_13 (theta_13 not discovered by the next generation), a plausible program could start with a low energy neutrino factory, followed by energy upgrade, and then baseline or detector mass upgrade, depending on the outcome of the earlier phases.

hep-ph↗

arXiv2009

On near detectors at a neutrino factory

The geometric effects of the beam in near detectors at a neutrino factory are discussed. The refined systematics treatment, including cross section errors, flux errors and background uncertainties, is compared with the IDS-NF one. Different near detector setups are included. We also probe their effects both at the measurements of standard neutrino oscillation parameters and constraints of the non-standard neutrino interaction.

hep-ph↗