Search

This paper addresses the problem of localizing change points in high-dimensional linear regression models with piecewise constant regression coefficients. We develop a dynamic programming approach to estimate the locations of the change points whose performance improves upon the current state-of-the-art, even as the dimensionality, the sparsity of the regression coefficients, the temporal spacing between two consecutive change points, and the magnitude of the difference of two consecutive regression coefficient vectors are allowed to vary with the sample size. Furthermore, we devise a computationally-efficient refinement procedure that provably reduces the localization error of preliminary estimates of the change points. We demonstrate minimax lower bounds on the localization error that nearly match the upper bound on the localization error of our methodology and show that the signal-to-noise condition we impose is essentially the weakest possible based on information-theoretic arguments. Extensive numerical results support our theoretical findings, and experiments on real air quality data reveal change points supported by historical information not used by the algorithm.

Conditional sequence generation aims to instruct the generation procedure by conditioning the model with additional context information, which is a self-supervised learning issue (a form of unsupervised learning with supervision information from data itself). Unfortunately, the current state-of-the-art generative models have limitations in sequence generation with multiple attributes. In this paper, we propose a novel conditional hybrid GAN (C-Hybrid-GAN) to solve this issue. Discrete sequence with triplet attributes are separately generated when conditioned on the same context. Most importantly, relational reasoning technique is exploited to model not only the dependency inside each sequence of the attribute during the training of the generator but also the consistency among the sequences of attributes during the training of the discriminator. To avoid the non-differentiability problem in GANs encountered during discrete data generation, we exploit the Gumbel-Softmax technique to approximate the distribution of discrete-valued sequences.Through evaluating the task of generating melody (associated with note, duration, and rest) from lyrics, we demonstrate that the proposed C-Hybrid-GAN outperforms the existing methods in context-conditioned discrete-valued sequence generation.

In this paper, we propose and analyze the sparsity-aware sign subband adaptive filtering with individual weighting factors (S-IWF-SSAF) algorithm, and consider its application in acoustic echo cancellation (AEC). Furthermore, we design a joint optimization scheme of the step-size and the sparsity penalty parameter to enhance the S-IWF-SSAF performance in terms of convergence rate and steady-state error. A theoretical analysis shows that the S-IWF-SSAF algorithm outperforms the previous sign subband adaptive filtering with individual weighting factors (IWF-SSAF) algorithm in sparse scenarios. In particular, compared with the existing analysis on the IWF-SSAF algorithm, the proposed analysis does not require the assumptions of large number of subbands, long adaptive filter, and paraunitary analysis filter bank, and matches well the simulated results. Simulations in both system identification and AEC situations have demonstrated our theoretical analysis and the effectiveness of the proposed algorithms., Comment: 17 pages, 12 figures

Fashion is an important part of human experience. Events such as interviews, meetings, marriages, etc. are often based on clothing styles. The rise in the fashion industry and its effect on social influencing have made outfit compatibility a need. Thus, it necessitates an outfit compatibility model to aid people in clothing recommendation. However, due to the highly subjective nature of compatibility, it is necessary to account for personalization. Our paper devises an attribute-wise interpretable compatibility scheme with personal preference modelling which captures user-item interaction along with general item-item interaction. Our work solves the problem of interpretability in clothing matching by locating the discordant and harmonious attributes between fashion items. Extensive experiment results on IQON3000, a publicly available real-world dataset, verify the effectiveness of the proposed model., Comment: 10 pages, 5 figures, to be published in IEEE BigMM, 2020

Sheet music, audio, and lyrics are three main modalities during writing a song. In this paper, we propose an unsupervised generative adversarial alignment representation (UGAAR) model to learn deep discriminative representations shared across three major musical modalities: sheet music, lyrics, and audio, where a deep neural network based architecture on three branches is jointly trained. In particular, the proposed model can transfer the strong relationship between audio and sheet music to audio-lyrics and sheet-lyrics pairs by learning the correlation in the latent shared subspace. We apply CCA components of audio and sheet music to establish new ground truth. The generative (G) model learns the correlation of two couples of transferred pairs to generate new audio-sheet pair for a fixed lyrics to challenge the discriminative (D) model. The discriminative model aims at distinguishing the input which is from the generative model or the ground truth. The two models simultaneously train in an adversarial way to enhance the ability of deep alignment representation learning. Our experimental results demonstrate the feasibility of our proposed UGAAR for alignment representation learning among sheet music, audio, and lyrics., Comment: 5 pages, 2 figures, 2 tables

In this paper we are concerned with a sequence of univariate random variables with piecewise polynomial means and independent sub-Gaussian noise. The underlying polynomials are allowed to be of arbitrary but fixed degrees. All the other model parameters are allowed to vary depending on the sample size. We propose a two-step estimation procedure based on the $\ell_0$-penalisation and provide upper bounds on the localisation error. We complement these results by deriving a global information-theoretic lower bounds, which show that our two-step estimators are nearly minimax rate-optimal. We also show that our estimator enjoys near optimally adaptive performance by attaining individual localisation errors depending on the level of smoothness at individual change points of the underlying signal. In addition, under a special smoothness constraint, we provide a minimax lower bound on the localisation errors. This lower bound is independent of the polynomial orders and is sharper than the global minimax lower bound.

High-dimensional self-exciting point processes have been widely used in many application areas to model discrete event data in which past and current events affect the likelihood of future events. In this paper, we are concerned with detecting abrupt changes of the coefficient matrices in discrete-time high-dimensional self-exciting Poisson processes, which have yet to be studied in the existing literature due to both theoretical and computational challenges rooted in the non-stationary and high-dimensional nature of the underlying process. We propose a penalized dynamic programming approach which is supported by a theoretical rate analysis and numerical evidence.

Graph matching is a fruitful area in terms of both algorithms and theories. In this paper, we exploit the degree information, which was previously used only in noiseless graphs and perfectly-overlapping Erd\H{o}s--R\'enyi random graphs matching. We are concerned with graph matching of partially-overlapping graphs and stochastic block models, which are more useful in tackling real-life problems. We propose the edge exploited degree profile graph matching method and two refined varations. We conduct a thorough analysis of our proposed methods' performances in a range of challenging scenarios, including a zebrafish neuron activity data set and a coauthorship data set. Our methods are proved to be numerically superior than the state-of-the-art methods.

We investigate sequential change point estimation and detection in univariate nonparametric settings, where a stream of independent observations from sub-Gaussian distributions with a common variance factor and piecewise-constant but otherwise unknown means are collected. We develop a simple CUSUM-based methodology that provably control the probability of false alarms or the average run length while minimizing, in a minimax sense, the detection delay. We allow for all the model parameters to vary in order to capture a broad range of levels of statistical hardness for the problem at hand. We further show how our methodology is applicable to the case in which multiple change points are to be estimated sequentially.

Named entity recognition (NER) from text has been a widely studied problem and usually extracts semantic information from text. Until now, NER from speech is mostly studied in a two-step pipeline process that includes first applying an automatic speech recognition (ASR) system on an audio sample and then passing the predicted transcript to a NER tagger. In such cases, the error does not propagate from one step to another as both the tasks are not optimized in an end-to-end (E2E) fashion. Recent studies confirm that integrated approaches (e.g., E2E ASR) outperform sequential ones (e.g., phoneme based ASR). In this paper, we introduce a first publicly available NER annotated dataset for English speech and present an E2E approach, which jointly optimizes the ASR and NER tagger components. Experimental results show that the proposed E2E approach outperforms the classical two-step approach. We also discuss how NER from speech can be used to handle out of vocabulary (OOV) words in an ASR system., Comment: submitted to Interspeech-2020

Visual Question Generation (VQG) is the task of generating natural questions based on an image. Popular methods in the past have explored image-to-sequence architectures trained with maximum likelihood which have demonstrated meaningful generated questions given an image and its associated ground-truth answer. VQG becomes more challenging if the image contains rich contextual information describing its different semantic categories. In this paper, we try to exploit the different visual cues and concepts in an image to generate questions using a variational autoencoder (VAE) without ground-truth answers. Our approach solves two major shortcomings of existing VQG systems: (i) minimize the level of supervision and (ii) replace generic questions with category relevant generations. Most importantly, by eliminating expensive answer annotations, the required supervision is weakened. Using different categories enables us to exploit different concepts as the inference requires only the image and the category. Mutual information is maximized between the image, question, and answer category in the latent space of our VAE. A novel category consistent cyclic loss is proposed to enable the model to generate consistent predictions with respect to the answer category, reducing redundancies and irregularities. Additionally, we also impose supplementary constraints on the latent space of our generative model to provide structure based on categories and enhance generalization by encapsulating decorrelated features within each dimension. Through extensive experiments, the proposed model, C3VQG outperforms state-of-the-art VQG methods with weak supervision.

Singing voice detection (SVD), to recognize vocal parts in the song, is an essential task in music information retrieval (MIR). The task remains challenging since singing voice varies and intertwines with the accompaniment music, especially for some complicated polyphonic music such as choral music recordings. To address this problem, we investigate singing voice detection while discarding the interference from the accompaniment. The proposed SVD has two steps: i. The singing voice separation (SVS) technique is first utilized to filter out the singing voice's potential part coarsely. ii. Upon the continuity of vocal in the time domain, Long-term Recurrent Convolutional Networks (LRCN) is used to learn compositional features. Moreover, to eliminate the outliers, we choose to use a median filter for time-domain smoothing. Experimental results show that the proposed method outperforms the existing state-of-the-art works on two public datasets, the Jamendo Corpus and the RWC pop dataset., Comment: Accepted by CSMT (The 9th Conference on Sound and Music Technology)

Opinion formation in complex social networks may exhibit complex system dynamics even when based on some simplest system evolution models. An interesting and important issue is the effects of the initial state on the final steady-state opinion distribution. We show that, while different initial opinion distributions certainly make differences to opinion evolution in social systems without noises, in systems with noises, given enough time, different initial states basically do not contribute to making any significant differences in the final steady state. Instead, it is the basal distribution of the preferred opinions that contributes to deciding the final state of the systems. Such an observation contradicts with a long-term belief on the roles of system initial state in opinion formation. We propose some brief discussions on the reasons supporting our statement, and the implications of such an observation in real-life applications., Comment: 4 pages, 5 figures

The strong coupling between two subsystems consisting of quantum emitters and photonic modes, at which the level splitting of mixed quantum states occurs, has been a central subject of quantum physics and nanophotonics due to various important applications. The spectral Rabi-splitting of photon emission or absorption has been adopted to experimentally characterize the strong coupling under the equality assumption that it is identical to the level splitting. Here, we for the first time reveal that the equality assumption is not valid. It is the invalidity that results in the relativity and diversity of the strong coupling characterized by the spectral Rabi-splitting to the measured subsystems, highly correlated with their dissipative decays. The strong coupling is easier to be observed from the subsystem with larger decay, and can be classified into pseudo-, dark-, middle-, and super-strong interaction regimes. We also suggest a prototype of coupled plasmon-exciton system for possibly future experiment observations on these novel predictions. Our work brings new fundamental insight to the light-matter interaction in nanostructures, which will stimulate further researches in this field., Comment: 40 pages, 4 figures

In this paper, we study the effect of population imbalance and its interplay with pairing strength and lattice effect in atomic Fermi gases in a one-dimensional optical lattice. We compute various phase diagrams as the system undergoes BCS-BEC crossover, using the same pairing fluctuation theory as in Part I. We find widespread pseudogap phenomena beyond the BCS regime and intermediate temperature superfluid states for relatively low population imbalances. The Fermi surface topology plays an important role in the behavior of $T_\text{c}$. For large $d$ and/or small $t$, which yield an open Fermi surface, superfluidity can be readily destroyed by a small amount of population imbalance $p$. The superfluid phase, especially in the BEC regime, can exist only for a highly restricted volume of the parameter space. Due to the continuum-lattice mixing, population imbalance gives rise to a new mechanism for pair hopping, as assisted by excessive majority fermions, which may lead to significant enhancement of $T_\text{c}$ on the BEC side of the Feshbach resonance, and also render $T_\text{c}$ approaching a constant asymptote in the BEC limit, when it exists. Furthermore, we find that not all minority fermions will be paired up in BEC limit, unlike the 3D continuum case. These predictions can be tested in future experiments., Comment: 14 pages, 14 figures in color

The superfluidity and pairing phenomena in ultracold atomic Fermi gases have been of great interest in recent years, with multiple tunable parameters. Here we study the BCS-BEC crossover behavior of balanced two-component Fermi gases in a one-dimensional optical lattice, which is distinct from the simple three-dimensional (3D) continuum and a fully 3D lattice often found in a condensed matter system. We use a pairing fluctuation theory which includes self-consistent feedback effects at finite temperatures, and find widespread pseudogap phenomena beyond the BCS regime. As a consequence of the lattice periodicity, the superfluid transition temperature $T_c$ decreases with pairing strength in the BEC regime, where it approaches asymptotically $T_c = \pi an/2m$, with $a$ being the $s$-wave scattering length, and $n$ ($m$) the fermion density (mass). In addition, the quasi-two dimensionality leads to fast growing (absolute value of the) fermionic chemical potential $\mu$ and pairing gap $\Delta$, which depends exponentially on the ratio $d/a$. Importantly, $T_c$ at unitarity increases with the lattice constant $d$ and hopping integral $t$. The effect of the van Hove singularity on $T_c$ is identified. The superfluid density exhibits $T^{3/2}$ power laws at low $T$, away from the extreme BCS limit. These predictions can be tested in future experiments., Comment: 11 pages, 7 figures in color

Powerful generative adversarial networks (GAN) have been developed to automatically synthesize realistic images from text. However, most existing tasks are limited to generating simple images such as flowers from captions. In this work, we extend this problem to the less addressed domain of face generation from fine-grained textual descriptions of face, e.g., "A person has curly hair, oval face, and mustache". We are motivated by the potential of automated face generation to impact and assist critical tasks such as criminal face reconstruction. Since current datasets for the task are either very small or do not contain captions, we generate captions for images in the CelebA dataset by creating an algorithm to automatically convert a list of attributes to a set of captions. We then model the highly multi-modal problem of text to face generation as learning the conditional distribution of faces (conditioned on text) in same latent space. We utilize the current state-of-the-art GAN (DC-GAN with GAN-CLS loss) for learning conditional multi-modality. The presence of more fine-grained details and variable length of the captions makes the problem easier for a user but more difficult to handle compared to the other text-to-image tasks. We flipped the labels for real and fake images and added noise in discriminator. Generated images for diverse textual descriptions show promising results. In the end, we show how the widely used inceptions score is not a good metric to evaluate the performance of generative models used for synthesizing faces from text.

In this paper, we study the offline change point localization problem in a sequence of dependent nonparametric random dot product graphs. To be specific, assume that at every time point, a network is generated from a nonparametric random dot product graph model \citep[see e.g.][]{athreya2017statistical}, where the latent positions are generated from unknown underlying distributions. The underlying distributions are piecewise constant in time and change at unknown locations, called change points. Most importantly, we allow for dependence among networks generated between two consecutive change points. This setting incorporates edge-dependence within networks and temporal dependence between networks, which is the most flexible setting in the published literature. To accomplish the task of consistently localizing change points, we propose a novel change point detection algorithm, consisting of two steps. First, we estimate the latent positions of the random dot product model, our theoretical result being a refined version of the state-of-the-art results, allowing the dimension of the latent positions to diverge. Subsequently, we construct a nonparametric version of the CUSUM statistic \citep[e.g.][]{Page1954, padilla2019optimal} that allows for temporal dependence. Consistent localization is proved theoretically and supported by extensive numerical experiments, which illustrate state-of-the-art performance. We also provide in depth discussion of possible extensions to give more understanding and insights.

Point clouds-based Networks have achieved great attention in 3D object classification, segmentation and indoor scene semantic parsing. In terms of face recognition, 3D face recognition method which directly consume point clouds as input is still under study. Two main factors account for this: One is how to get discriminative face representations from 3D point clouds using deep network; the other is the lack of large 3D training dataset. To address these problems, a data-free 3D face recognition method is proposed only using synthesized unreal data from statistical 3D Morphable Model to train a deep point cloud network. To ease the inconsistent distribution between model data and real faces, different point sampling methods are used in train and test phase. In this paper, we propose a curvature-aware point sampling(CPS) strategy replacing the original furthest point sampling(FPS) to hierarchically down-sample feature-sensitive points which are crucial to pass and aggregate features deeply. A PointNet++ like Network is used to extract face features directly from point clouds. The experimental results show that the network trained on generated data generalizes well for real 3D faces. Fine tuning on a small part of FRGCv2.0 and Bosphorus, which include real faces in different poses and expressions, further improves recognition accuracy., Comment: 10 pages, 8 figures

We study the multivariate nonparametric change point detection problem, where the data are a sequence of independent $p$-dimensional random vectors whose distributions are piecewise-constant with Lipschitz densities changing at unknown times, called change points. We quantify the size of the distributional change at any change point with the supremum norm of the difference between the corresponding densities. We are concerned with the localization task of estimating the positions of the change points. In our analysis, we allow for the model parameters to vary with the total number of time points, including the minimal spacing between consecutive change points and the magnitude of the smallest distributional change. We provide information-theoretic lower bounds on both the localization rate and the minimal signal-to-noise ratio required to guarantee consistent localization. We formulate a novel algorithm based on kernel density estimation that nearly achieves the minimax lower bound, save possibly for logarithm factors. We have provided extensive numerical evidence to support our theoretical findings.

In solid state physics, the Gr\"{u}neisen parameter (GP), originally introduced in the study of the effect of changing the volume of a crystal lattice on its vibrational frequency, has been widely used to investigate the characteristic energy scales of systems with respect to the changes of external potentials. On the other hand, the GP is little investigated in a strongly interacting quantum gas systems. Here we report on our general results on the origin of GP, new identity and caloric effects in quantum gases of ultracold atoms. We prove that the symmetry of the dilute quantum gas systems leads to a simple identity among three different types of GPs, quantifying caloric effect induced respectively by variations of volume, magnetic field and interaction. Using exact Bethe ansatz solutions, we present a rigorous study of these different GPs and the quantum refrigeration in one-dimensional Bose and Femi gases. Based on the exact equations of states of these systems, we obtain analytic results for the singular behaviour of the GPs and the caloric effects at quantum criticality. We also predict the existence of the lowest temperature for cooling near a quantum phase transition. It turns out that the interaction ramp-up and -down in quantum gases provides a promising protocol of quantum refrigeration in addition to the usual adiabatic demagnetization cooling in solid state materials., Comment: 11+7 pages, 4+1 figures

Autoregressive models capture stochastic processes in which past realizations determine the generative distribution of new data; they arise naturally in a variety of industrial, biomedical, and financial settings. A key challenge when working with such data is to determine when the underlying generative model has changed, as this can offer insights into distinct operating regimes of the underlying system. This paper describes a novel dynamic programming approach to localizing changes in high-dimensional autoregressive processes and associated error rates that improve upon the prior state of the art. When the model parameters are piecewise constant over time and the corresponding process is piecewise stable, the proposed dynamic programming algorithm consistently localizes change points even as the dimensionality, the sparsity of the coefficient matrices, the temporal spacing between two consecutive change points, and the magnitude of the difference of two consecutive coefficient matrices are allowed to vary with the sample size. Furthermore, the accuracy of initial, coarse change point localization estimates can be boosted via a computationally-efficient refinement algorithm that provably improves the localization error rate. Finally, a comprehensive simulation experiments and a real data analysis are provided to show the numerical superiority of our proposed methods., Comment: 53 pages; 4 figure

In analog to counterparts widely used in electronic circuits, all optical non-reciprocal devices are basic building blocks for both classical and quantum optical information processing. Approaching the fundamental limit of such devices, where the propagation of a single photon exhibits a good non-reciprocal characteristic, requires an asymmetric strong coupling between a single photon and a matter. Unfortunately it has been not realized yet. Here, we propose and experimentally realize a quantum non-reciprocity device with low optical losses and a high isolation of larger than 14 dB based on the cold atoms. Besides, the non-reciprocal transmission of a quantum qubit and non-reciprocal quantum storage of a true single photon are also realized. All results achieved would be very promising in building up quantum non-reciprocal devices for quantum networks., Comment: The concept of quantum non-reciprocity is not suitable in this manuscript

Cavity-free optical nonreciprocity components, which have an inherent strong asymmetric interaction between the forward- and backward-propagation direction of the probe field, are key to produce such as optical isolators and circulators. According to the proposal presented by Xia et al., [Phys. Rev. Lett. 121, 203602 (2018)], we experimentally build a device that uses cross-Kerr nonlinearity to achieve a cavity-free optical isolator and circulator. Its nonreciprocal behavior arises from the thermal motion of N-type configuration atoms, which induces a strong chiral cross-Kerr nonlinear response for the weak probe beam. We obtain a two-port optical isolator for up to 20 dB of isolation ratio in a specially designed Sagnac interferometer. The distinct propagation directions of the weak probe field determine its cross-phase shift and transmission, by which we demonstrate the accessibility of a four-port optical circulator., Comment: 5 pages, 3 figures

Melody generation from lyrics has been a challenging research issue in the field of artificial intelligence and music, which enables to learn and discover latent relationship between interesting lyrics and accompanying melody. Unfortunately, the limited availability of paired lyrics-melody dataset with alignment information has hindered the research progress. To address this problem, we create a large dataset consisting of 12,197 MIDI songs each with paired lyrics and melody alignment through leveraging different music sources where alignment relationship between syllables and music attributes is extracted. Most importantly, we propose a novel deep generative model, conditional Long Short-Term Memory - Generative Adversarial Network (LSTM-GAN) for melody generation from lyrics, which contains a deep LSTM generator and a deep LSTM discriminator both conditioned on lyrics. In particular, lyrics-conditioned melody and alignment relationship between syllables of given lyrics and notes of predicted melody are generated simultaneously. Experimental results have proved the effectiveness of our proposed lyrics-to-melody generative model, where plausible and tuneful sequences can be inferred from lyrics.

Frequent group activities of human beings have become an indispensable part in their daily life. Group recommendation can recommend satisfactory activities to group members in the recommender systems, and the key issue is how to aggregate preferences in different group members. Most existing group recommendation employed the predefined static aggregation strategies to aggregate the preferences of different group members, but these static strategies cannot simulate the dynamic group decision-making. Meanwhile, most of these methods depend on intuitions or assumptions to analyze the influence of group members and lack of convincing theoretical support. We argue that the influence of group members plays a particularly important role in group decision-making and it can better assist group profile modeling and perform more accurate group recommendation. To tackle the issue of preference aggregation for group recommendation, we propose a novel attentive aggregation representation learning method based on sociological theory for group recommendation, namely SIAGR (short for "Social Influence-based Attentive Group Recommendation"), which takes attention mechanisms and the popular method (BERT) as the aggregation representation for group profile modeling. Specifically, we analyze the influence of group members based on social identity theory and two-step flow theory and exploit an attentive mavens mining method. In addition, we develop a BERT-based representation method to learn the interaction of group members. Lastly, we complete the group recommendation under the neural collaborative filtering framework and verify the effectiveness of the proposed method by experimenting.

Since many online music services emerged in recent years so that effective music recommendation systems are desirable. Some common problems in recommendation system like feature representations, distance measure and cold start problems are also challenges for music recommendation. In this paper, I proposed a triplet neural network, exploiting both positive and negative samples to learn the representation and distance measure between users and items, to solve the recommendation task., Comment: 1 figure; 1 table

Deep learning has successfully shown excellent performance in learning joint representations between different data modalities. Unfortunately, little research focuses on cross-modal correlation learning where temporal structures of different data modalities, such as audio and video, should be taken into account. Music video retrieval by given musical audio is a natural way to search and interact with music contents. In this work, we study cross-modal music video retrieval in terms of emotion similarity. Particularly, audio of an arbitrary length is used to retrieve a longer or full-length music video. To this end, we propose a novel audio-visual embedding algorithm by Supervised Deep CanonicalCorrelation Analysis (S-DCCA) that projects audio and video into a shared space to bridge the semantic gap between audio and video. This also preserves the similarity between audio and visual contents from different videos with the same class label and the temporal structure. The contribution of our approach is mainly manifested in the two aspects: i) We propose to select top k audio chunks by attention-based Long Short-Term Memory (LSTM)model, which can represent good audio summarization with local properties. ii) We propose an end-to-end deep model for cross-modal audio-visual learning where S-DCCA is trained to learn the semantic correlation between audio and visual modalities. Due to the lack of music video dataset, we construct 10K music video dataset from YouTube 8M dataset. Some promising results such as MAP and precision-recall show that our proposed model can be applied to music video retrieval., Comment: 8 pages, 9 figures. Accepted by ISM 2018

Cross-modal retrieval aims to retrieve data in one modality by a query in another modality, which has been a very interesting research issue in the field of multimedia, information retrieval, and computer vision, and database. Most existing works focus on cross-modal retrieval between text-image, text-video, and lyrics-audio.Little research addresses cross-modal retrieval between audio and video due to limited audio-video paired datasets and semantic information. The main challenge of audio-visual cross-modal retrieval task focuses on learning joint embeddings from a shared subspace for computing the similarity across different modalities, where generating new representations is to maximize the correlation between audio and visual modalities space. In this work, we propose a novel deep triplet neural network with cluster canonical correlation analysis(TNN-C-CCA), which is an end-to-end supervised learning architecture with audio branch and video branch.We not only consider the matching pairs in the common space but also compute the mismatching pairs when maximizing the correlation. In particular, two significant contributions are made: i) a better representation by constructing deep triplet neural network with triplet loss for optimal projections can be generated to maximize correlation in the shared subspace. ii) positive examples and negative examples are used in the learning stage to improve the capability of embedding learning between audio and video. Our experiment is run over 5-fold cross-validation, where average performance is applied to demonstrate the performance of audio-video cross-modal retrieval. The experimental results achieved on two different audio-visual datasets show the proposed learning architecture with two branches outperforms existing six CCA-based methods and four state-of-the-art based cross-modal retrieval methods., Comment: 21 pages,11 figures

Light-emitting diodes (LEDs) are efficient light sources extensively applied in people's daily life nowadays, with profound effect on their psychological and physiological state subtly and deeply. The correlated color temperature (CCT) the eyes perceive is a key parameter, but the CCT tuning technology of LEDs remains in the development stage. In this study, an electrochromic (EC) material is employed in the tuning device, exhibiting a wide CCT tuning range from 3,200 K to 6,900 K, varying almost along the black-body locus. The driving voltage is lower than 2.4 V, with luminosity loss less than 16%. Using the EC materials, the performance of CCT tuning technology is able to be further enhanced.