Search arXiv⌕ Search

arXiv subjects

Ya Li

Publications and source records attributed to Ya Li.

115 records · Page 7Linked to original sources
arXiv2017

Cost-Effective Active Learning for Deep Image Classification

Recent successes in learning-based image classification, however, heavily rely on the large number of annotated training samples, which may require considerable human efforts. In this paper, we propose a novel active learning framework, which is capable of building a competitive classifier with optimal feature representation via a limited amount of labeled training instances in an incremental learning manner. Our approach advances the existing active learning methods in two aspects. First, we incorporate deep convolutional neural networks into active learning. Through the properly designed framework, the feature representation and the classifier can be simultaneously updated with progressively annotated informative samples. Second, we present a cost-effective sample selection strategy to improve the classification performance with less manual annotations. Unlike traditional methods focusing on only the uncertain samples of low prediction confidence, we especially discover the large amount of high confidence samples from the unlabeled set for feature learning. Specifically, these high confidence samples are automatically selected and iteratively assigned pseudo-labels. We thus call our framework "Cost-Effective Active Learning" (CEAL) standing for the two advantages.Extensive experiments demonstrate that the proposed CEAL framework can achieve promising results on two challenging image classification datasets, i.e., face recognition on CACD database [1] and object categorization on Caltech-256 [2].

cs.CV↗
arXiv2016

The $S$-wave resonance contributions to the three-body decays $B^0_{(s)}\to η_c f_0(X)\to η_cπ^+π^-$ in perturbative QCD approach

In this paper, we study the three-body decays $B^0/B^0_s \to η_c f_0(X)\to η_c π^+π^-$ by employing the perturbative QCD (PQCD) factorization approach. We evaluate the $S$-wave resonance contributions by using the two-pion distribution amplitude $Φ_{ππ}^{\rm S}$. The Breit-Wigner formula for the $f_0(500)$, $f_0(1500)$, and $f_0(1790)$ resonances and the Flatté model for the $f_0(980)$ resonance are adopted to parameterize the time-like scalar form factors $F_{s}(ω^2)$. We also use the D.~V.~Bugg model to parameterize the $f_0(500)$ and compare the relevant theoretical predictions from different models. We found the following results: (a) the PQCD predictions for the branching ratios are ${\cal B}(B^0\to η_c f_0(500)[π^+π^-])= \left ( 1.53 ^{+0.76}_{-0.35} \right ) \times 10^{-6}$ for Breit-Wigner model and ${\cal B}(B^0\to η_c f_0(500)[π^+π^-])= \left ( 2.31 ^{+0.96}_{-0.48} \right ) \times 10^{-6}$ for D.~V.~Bugg model; (b) $ {\cal B}(B_s\to η_c f_0(X)[π^+π^-] ) =\left ( 5.02^{+1.49}_{-1.08} \right )\times 10^{-5}$ when the contributions from $f_0(X)=(f_0(980),f_0(1500),f_0(1790))$ are all taken into account; and (c) The considered decays could be measured at the ongoing LHCb experiment, consequently, the formalism of two-hadron distribution amplitudes could also be tested by such experiments.

hep-ph↗
arXiv2016

DARI: Distance metric And Representation Integration for Person Verification

The past decade has witnessed the rapid development of feature representation learning and distance metric learning, whereas the two steps are often discussed separately. To explore their interaction, this work proposes an end-to-end learning framework called DARI, i.e. Distance metric And Representation Integration, and validates the effectiveness of DARI in the challenging task of person verification. Given the training images annotated with the labels, we first produce a large number of triplet units, and each one contains three images, i.e. one person and the matched/mismatch references. For each triplet unit, the distance disparity between the matched pair and the mismatched pair tends to be maximized. We solve this objective by building a deep architecture of convolutional neural networks. In particular, the Mahalanobis distance matrix is naturally factorized as one top fully-connected layer that is seamlessly integrated with other bottom layers representing the image feature. The image feature and the distance metric can be thus simultaneously optimized via the one-shot backward propagation. On several public datasets, DARI shows very promising performance on re-identifying individuals cross cameras against various challenges, and outperforms other state-of-the-art approaches.

cs.CV↗
arXiv2016

Audio Visual Emotion Recognition with Temporal Alignment and Perception Attention

This paper focuses on two key problems for audio-visual emotion recognition in the video. One is the audio and visual streams temporal alignment for feature level fusion. The other one is locating and re-weighting the perception attentions in the whole audio-visual stream for better recognition. The Long Short Term Memory Recurrent Neural Network (LSTM-RNN) is employed as the main classification architecture. Firstly, soft attention mechanism aligns the audio and visual streams. Secondly, seven emotion embedding vectors, which are corresponding to each classification emotion type, are added to locate the perception attentions. The locating and re-weighting process is also based on the soft attention mechanism. The experiment results on EmotiW2015 dataset and the qualitative analysis show the efficiency of the proposed two techniques.

cs.CV↗
arXiv2015

Deep Boosting: Joint Feature Selection and Analysis Dictionary Learning in Hierarchy

This work investigates how the traditional image classification pipelines can be extended into a deep architecture, inspired by recent successes of deep neural networks. We propose a deep boosting framework based on layer-by-layer joint feature boosting and dictionary learning. In each layer, we construct a dictionary of filters by combining the filters from the lower layer, and iteratively optimize the image representation with a joint discriminative-generative formulation, i.e. minimization of empirical classification error plus regularization of analysis image generation over training images. For optimization, we perform two iterating steps: i) to minimize the classification error, select the most discriminative features using the gentle adaboost algorithm; ii) according to the feature selection, update the filters to minimize the regularization on analysis image representation using the gradient descent method. Once the optimization is converged, we learn the higher layer representation in the same way. Our model delivers several distinct advantages. First, our layer-wise optimization provides the potential to build very deep architectures. Second, the generated image representation is compact and meaningful. In several visual recognition tasks, our framework outperforms existing state-of-the-art approaches.

cs.CV↗
arXiv2014

$\bar{B}^0_s \to (π^0 η^{(*)}, η^{(*)}η^{(*)})$ decays and the effects of next-to-leading order contributions in the perturbative QCD approach

In this paper, we calculate the branching ratios and CP violating asymmetries of the five $\bar{B}^0_s \to (π^0η^{(*)},η^{(*)}η^{(*)})$ decays, by employing the perturbative QCD (pQCD) factorization approach and with the inclusion of all currently known next-to-leading order (NLO) contributions. We find that (a) the NLO contributions can provide about 100% enhancements to the LO pQCD predictions for the decay rates of $\bar{B}_s^0 \to ηη^\prime$ and $η^\prime η^\prime$ decays, but result in small changes to $Br(\bar{B}_s \to π^0 η^{(*)})$ and $Br(\bar{B}_s \to ηη)$; (b) the newly known NLO twist-2 and twist-3 contributions to the relevant form factors can provide about 10% enhancements to the decay rates of the considered decays; (c) for $\bar{B}_s \to π^0 η^{(*)}$ decays, their direct CP-violating asymmetries $\cala_f^{dir}$ could be enhanced significantly by the inclusion of the NLO contributions; and (d) the pQCD predictions for $Br(\bar{B}_s \to ηη^{(*)})$ and $Br(\bar{B}_s \to η^\prime η^\prime)$ can be as large as $4\times 10^{-5}$, which may be measurable at LHCb or the forthcoming super-B experiments.

hep-ph↗
arXiv2014

Defuzzify firstly or finally: Dose it matter in fuzzy DEMATEL under uncertain environment?

Decision-Making Trial and Evaluation Laboratory (DEMATEL) method is widely used in many real applications. With the desirable property of efficient handling with the uncertain information in decision making, the fuzzy DEMATEL is heavily studied. Recently, Dytczak and Ginda suggested to defuzzify the fuzzy numbers firstly and then use the classical DEMATEL to obtain the final result. In this short paper, we show that it is not reasonable in some situations. The results of defuzzification at the first step are not coincide with the results of defuzzification at the final step.It seems that the alternative is to defuzzification in the final step in fuzzy DEMATEL.

cs.AI↗