Confidence measure is a performance index of particular importance for automatic speech recognition (ASR) systems deployed in real-world scenarios. In the present study, utterance-level neural confidence measure (NCM) in end-to-end automatic speech r
ecognition (E2E ASR) is investigated. The E2E system adopts the joint CTC-attention Transformer architecture. The prediction of NCM is formulated as a task of binary classification, i.e., accept/reject the input utterance, based on a set of predictor features acquired during the ASR decoding process. The investigation is focused on evaluating and comparing the efficacies of predictor features that are derived from different internal and external modules of the E2E system. Experiments are carried out on children speech, for which state-of-the-art ASR systems show less than satisfactory performance and robust confidence measure is particularly useful. It is noted that predictor features related to acoustic information of speech play a more important role in estimating confidence measure than those related to linguistic information. N-best score features show significantly better performance than single-best ones. It has also been shown that the metrics of EER and AUC are not appropriate to evaluate the NCM of a mismatched ASR with significant performance gap.
We explore the possibility that the dark matter relic density is not produced by thermal mechanism directly, but by the decay of other heavier dark sector particles which on the other hand can be produced by the thermal freeze-out mechanism. Using a
concrete model with a light dark matter from dark sector decay, we study the collider signature of the dark sector particles in association with Higgs production processes. We find that the future lepton colliders can be a better place to probe the signature of this kind of light dark matter model than the hadron collider such as LHC. Meanwhile, it is found that a Higgs factory with center of mass energy 250 GeV has a better potential to resolve the signature of this kind of light dark matter model than the Higgs factory with center of mass energy 350 GeV.
Privacy amplification is an indispensable step in the post-processing of quantum key distribution, which can be used to compress the redundancy of shared key and improve the security level of the key. The commonly used privacy amplification is based
on the random selection of universal hash functions, which needs the help of an additional random source, while it does not exist in general. In this paper, we propose a privacy amplification scheme based on composite coding, which is an extension of quantum CSS codes to classical linear codes. Compared with the universal hashing function, the proposed scheme does not need other random sources, and the randomness can be completely provided by the qubit string. Furthermore, the information-theoretic bound for the extraction of the key is obvious in composite coding.
Semantic Role Labeling (SRL) aims at recognizing the predicate-argument structure of a sentence and can be decomposed into two subtasks: predicate disambiguation and argument labeling. Prior work deals with these two tasks independently, which ignore
s the semantic connection between the two tasks. In this paper, we propose to use the machine reading comprehension (MRC) framework to bridge this gap. We formalize predicate disambiguation as multiple-choice machine reading comprehension, where the descriptions of candidate senses of a given predicate are used as options to select the correct sense. The chosen predicate sense is then used to determine the semantic roles for that predicate, and these semantic roles are used to construct the query for another MRC model for argument labeling. In this way, we are able to leverage both the predicate semantics and the semantic role semantics for argument labeling. We also propose to select a subset of all the possible semantic roles for computational efficiency. Experiments show that the proposed framework achieves state-of-the-art results on both span and dependency benchmarks.
We stabilize the full Arthur-Selberg trace formula for the metaplectic covering of symplectic groups over a number field. This provides a decomposition of the invariant trace formula for metaplectic groups, which encodes information about the genuine
$L^2$-automorphic spectrum, into a linear combination of stable trace formulas of products of split odd orthogonal groups via endoscopic transfer. By adapting the strategies of Arthur and Moeglin-Waldspurger from the linear case, the proof is built on a long induction process that mixes up local and global, geometric and spectral data. As a by-product, we also stabilize the local trace formula for metaplectic groups over any local field of characteristic zero.
Both classification and regression tasks are susceptible to the biased distribution of training data. However, existing approaches are focused on the class-imbalanced learning and cannot be applied to the problems of numerical regression where the le
arning targets are continuous values rather than discrete labels. In this paper, we aim to improve the accuracy of the regression analysis by addressing the data skewness/bias during model training. We first introduce two metrics, uniqueness and abnormality, to reflect the localized data distribution from the perspectives of their feature (i.e., input) space and target (i.e., output) space. Combining these two metrics we propose a Variation-Incentive Loss re-weighting method (VILoss) to optimize the gradient descent-based model training for regression analysis. We have conducted comprehensive experiments on both synthetic and real-world data sets. The results show significant improvement in the model quality (reduction in error by up to 11.9%) when using VILoss as the loss criterion in training.
In the fourth paper of this series, we present the metallicity-dependent Sloan Digital Sky Survey (SDSS) stellar color loci of red giant stars, using a spectroscopic sample of red giants in the SDSS Stripe 82 region. The stars span a range of 0.55 --
1.2 mag in color g-i, -0.3 -- -2.5 in metallicity [Fe/H], and have values of surface gravity log g smaller than 3.5 dex. As in the case of main-sequence (MS) stars, the intrinsic widths of loci of red giants are also found to be quite narrow, a few mmag at maximum. There are however systematic differences between the metallicity-dependent stellar loci of red giants and MS stars. The colors of red giants are less sensitive to metallicity than those of MS stars. With good photometry, photometric metallicities of red giants can be reliably determined by fitting the u-g, g-r, r-i, and i-z colors simultaneously to an accuracy of 0.2 -- 0.25 dex, comparable to the precision achievable with low-resolution spectroscopy for a signal-to-noise ratio of 10. By comparing fitting results to the stellar loci of red giants and MS stars, we propose a new technique to discriminate between red giants and MS stars based on the SDSS photometry. The technique achieves completeness of ~ 70 per cent and efficiency of ~ 80 per cent in selecting metal-poor red giant stars of [Fe/H] $le$ -1.2. It thus provides an important tool to probe the structure and assemblage history of the Galactic halo using red giant stars.
Recently the splitting of elliptic flow $v_2$ at finite rapidities has been proposed as a result of the global vorticity in non-central relativistic heavy ion collisions. Using a multi-phase transport model that automatically includes the vorticity f
ield and flow fluctuations, we confirm the left-right (i.e., on opposite sides of the impact parameter axis) splitting of the elliptic flow at finite rapidities. However, we find that this $v_2$ splitting is a result of the non-zero directed flow $v_1$ at finite rapidities, with the splitting magnitude $approx 8v_1/3pi$. As a result, the $v_2$ splitting vanishes at zero transverse momentum ($p_{rm T}$), and its magnitude and sign may have non-trivial dependences on $p_{rm T}$, centrality, collision energy, and hadron species. Since the left-right $v_2$ splitting is a combined effect of $v_1$ and $v_2$, it will benefit studies of the three-dimensional structure and dynamics of the dense matter.
Object SLAM introduces the concept of objects into Simultaneous Localization and Mapping (SLAM) and helps understand indoor scenes for mobile robots and object-level interactive applications. The state-of-art object SLAM systems face challenges such
as partial observations, occlusions, unobservable problems, limiting the mapping accuracy and robustness. This paper proposes a novel monocular Semantic Object SLAM (SO-SLAM) system that addresses the introduction of object spatial constraints. We explore three representative spatial constraints, including scale proportional constraint, symmetrical texture constraint and plane supporting constraint. Based on these semantic constraints, we propose two new methods - a more robust object initialization method and an orientation fine optimization method. We have verified the performance of the algorithm on the public datasets and an author-recorded mobile robot dataset and achieved a significant improvement on mapping effects. We will release the code here: https://github.com/XunshanMan/SoSLAM.
Dimension reduction for high-dimensional compositional data plays an important role in many fields, where the principal component analysis of the basis covariance matrix is of scientific interest. In practice, however, the basis variables are latent
and rarely observed, and standard techniques of principal component analysis are inadequate for compositional data because of the simplex constraint. To address the challenging problem, we relate the principal subspace of the centered log-ratio compositional covariance to that of the basis covariance, and prove that the latter is approximately identifiable with the diverging dimensionality under some subspace sparsity assumption. The interesting blessing-of-dimensionality phenomenon enables us to propose the principal subspace estimation methods by using the sample centered log-ratio covariance. We also derive nonasymptotic error bounds for the subspace estimators, which exhibits a tradeoff between identification and estimation. Moreover, we develop efficient proximal alternating direction method of multipliers algorithms to solve the nonconvex and nonsmooth optimization problems. Simulation results demonstrate that the proposed methods perform as well as the oracle methods with known basis. Their usefulness is illustrated through an analysis of word usage pattern for statisticians.
