Searcharxiv⌕ Search

arXiv subjects

Ya Li

Publications and source records attributed to Ya Li.

At least 73 records · Page 4Linked to original sources
arXiv2024

CRB Minimization for RIS-aided mmWave Integrated Sensing and Communications

In this paper, reconfigurable intelligent surface (RIS) is employed in a millimeter wave (mmWave) integrated sensing and communications (ISAC) system. To alleviate the multi-hop attenuation, the semi-self sensing RIS approach is adopted, wherein sensors are configured at the RIS to receive the radar echo signal. Focusing on the estimation accuracy, the Cramer-Rao bound (CRB) for estimating the direction-of-the-angles is derived as the metric for sensing performance. A joint optimization problem on hybrid beamforming and RIS phaseshifts is proposed to minimize the CRB, while maintaining satisfactory communication performance evaluated by the achievable data rate. The CRB minimization problem is first transformed as a more tractable form based on Fisher information matrix (FIM). To solve the complex non-convex problem, a double layer loop algorithm is proposed based on penalty concave-convex procedure (penalty-CCCP) and block coordinate descent (BCD) method with two sub-problems. Successive convex approximation (SCA) algorithm and second order cone (SOC) constraints are employed to tackle the non-convexity in the hybrid beamforming optimization. To optimize the unit modulus constrained analog beamforming and phase shifts, manifold optimization (MO) is adopted. Finally, the numerical results verify the effectiveness of the proposed CRB minimization algorithm, and show the performance improvement compared with other baselines. Additionally, the proposed hybrid beamforming algorithm can achieve approximately 96% of the sensing performance exhibited by the full digital approach within only a limited number of radio frequency (RF) chains.

eess.SP↗
arXiv2023

Frame-level emotional state alignment method for speech emotion recognition

Speech emotion recognition (SER) systems aim to recognize human emotional state during human-computer interaction. Most existing SER systems are trained based on utterance-level labels. However, not all frames in an audio have affective states consistent with utterance-level label, which makes it difficult for the model to distinguish the true emotion of the audio and perform poorly. To address this problem, we propose a frame-level emotional state alignment method for SER. First, we fine-tune HuBERT model to obtain a SER system with task-adaptive pretraining (TAPT) method, and extract embeddings from its transformer layers to form frame-level pseudo-emotion labels with clustering. Then, the pseudo labels are used to pretrain HuBERT. Hence, the each frame output of HuBERT has corresponding emotional information. Finally, we fine-tune the above pretrained HuBERT for SER by adding an attention layer on the top of it, which can focus only on those frames that are emotionally more consistent with utterance-level label. The experimental results performed on IEMOCAP indicate that our proposed method performs better than state-of-the-art (SOTA) methods.

cs.SD↗
arXiv2023

Hypergraph Enhanced Knowledge Tree Prompt Learning for Next-Basket Recommendation

Next-basket recommendation (NBR) aims to infer the items in the next basket given the corresponding basket sequence. Existing NBR methods are mainly based on either message passing in a plain graph or transition modelling in a basket sequence. However, these methods only consider point-to-point binary item relations while item dependencies in real world scenarios are often in higher order. Additionally, the importance of the same item to different users varies due to variation of user preferences, and the relations between items usually involve various aspects. As pretrained language models (PLMs) excel in multiple tasks in natural language processing (NLP) and computer vision (CV), many researchers have made great efforts in utilizing PLMs to boost recommendation. However, existing PLM-based recommendation methods degrade when encountering Out-Of-Vocabulary (OOV) items. OOV items are those whose IDs are out of PLM's vocabulary and thus unintelligible to PLM. To settle the above challenges, we propose a novel method HEKP4NBR, which transforms the knowledge graph (KG) into prompts, namely Knowledge Tree Prompt (KTP), to help PLM encode the OOV item IDs in the user's basket sequence. A hypergraph convolutional module is designed to build a hypergraph based on item similarities measured by an MoE model from multiple aspects and then employ convolution on the hypergraph to model correlations among multiple items. Extensive experiments are conducted on HEKP4NBR on two datasets based on real company data and validate its effectiveness against multiple state-of-the-art methods.

cs.IR↗
arXiv2023

The coherence of wave-packet-tunable photons

The wave-packet-tunable photons [Optics Express 30, 2792-2802 (2022)] generated by spontaneous Raman scattering (SRS) based on atomic ensemble lay a foundation for the hybrid quantum network to successfully connect quantum nodes with different bandwidths, but the coherence time of wave-packet photons becomes the key factor limiting the distance of entanglement distribution. The coherence of photons deteriorates with the propagation distance of the entanglement distribution. So far, the coherence of wave-packet-tunable photons entangled with an atomic memory has remained unexplored. An unequal arm fiber interferometer is constructed to measure the interference visibility of 150 ns-1.06 μs pulse width wave-packet-tunable photons. The coherence time and bandwidth of the photons can be directly derived from the decay of the visibility in the interferogram as the wave-packet photons length increases. The measured results show that the coherence time of write laser is 2.36 μs and bandwidth is 78 KHz, which interact onto atoms can generate Stokes photons with the coherence time is 1.14μs and bandwidth is 156 KHz. The measurement of coherence of wave-packet-tunable photons lays the foundation for establishing the distribution of entanglement between spatially separated memories in hybrid quantum networks, and for establishing a baseline telescope of arbitrary length through wave-packet-tunable photon interference.

physics.optics↗
arXiv2023

CONCSS: Contrastive-based Context Comprehension for Dialogue-appropriate Prosody in Conversational Speech Synthesis

Conversational speech synthesis (CSS) incorporates historical dialogue as supplementary information with the aim of generating speech that has dialogue-appropriate prosody. While previous methods have already delved into enhancing context comprehension, context representation still lacks effective representation capabilities and context-sensitive discriminability. In this paper, we introduce a contrastive learning-based CSS framework, CONCSS. Within this framework, we define an innovative pretext task specific to CSS that enables the model to perform self-supervised learning on unlabeled conversational datasets to boost the model's context understanding. Additionally, we introduce a sampling strategy for negative sample augmentation to enhance context vectors' discriminability. This is the first attempt to integrate contrastive learning into CSS. We conduct ablation studies on different contrastive learning strategies and comprehensive experiments in comparison with prior CSS systems. Results demonstrate that the synthesized speech from our proposed method exhibits more contextually appropriate and sensitive prosody.

cs.CL↗
arXiv2023

Study of four-body decays $B_{(s)} \to (ππ)(ππ)$ in the perturbative QCD approach

In this work, we make a systematical study on the four-body $B_{(s)} \to (ππ)(ππ)$ decays in the perturbative QCD approach, where the $ππ$ invariant mass spectra are dominated by the vector resonance $ρ(770)$ and the scalar resonance $f_0(980)$. We improve the Gengenbauer moments for the longitudinal $P$-wave two-pion distribution amplitudes (DAs) by fitting the PQCD factorization formulas to measured branching ratios of three-body and four-body $B$ decays. With the fitted Gegenbauer moments, we make predictions for the branching ratios and direct $CP$ asymmetries of four-body $B_{(s)} \to (ππ)(ππ)$ decays. We extract the branching ratios of two-body $B_{(s)} \to ρρ$ from the corresponding four-body decay modes and calculate the relevant polarization fractions. We find that the ${\cal B}(B^0 \to ρ^+ρ^-)$ is consistent with the previous theoretical predictions and data. The leading-order PQCD calculations of the ${\cal B}(B^+\to ρ^+ρ^0)$, ${\cal B}(B^0\to ρ^0ρ^0)$ and the $f_0(B^0\to ρ^0ρ^0)$ are a bit lower than the experimental measurements, which should be further examined. In addition, the "true" and "fake" triple-product asymmetries (TPAs) in the $B_{(s)}\to (ππ)(ππ)$ decays are also analyzed. The sizable averaged TPA ${\cal A}_{\text{T-true}}^{1, \text{ave}}=25.26\%$ of the color-suppressed decay $B^0\to ρ^0ρ^0 \to (π^+π^-)(π^+π^-)$ is predicted for the first time, which deviates a lot from the so-called "true" TPA $\mathcal{A}_\text{T-true}^1=7.92\%$ due to the large direct $CP$ violation. A large "fake" TPA $\mathcal{A}_\text{T-fake}^1=24.96\%$ of the decay $B^0\to ρ^0ρ^0 \to (π^+π^-)(π^+π^-)$ is also found, which indicates the significance of the final-state interactions. The predictions in this work can be tested by LHCb and Belle-II experiments in the near future.

hep-ph↗
arXiv2023

Decoherence of Single-Excitation Entanglement over DLCZ Quantum Networks Caused by Slow-Magnetic-Field Fluctuations and Protection Approach

Atomic spin-wave (SW) memory is a building block for quantum repeater. However, due to decoherence caused by atomic motions and magnetic-field gradients, SW retrieval efficiency decays with storage time. Slow magnetic-field fluctuations (SMFFs) lead to shot-to-shot phase noises of SWs, but have not been considered in previous repeater protocols. Here, we develop a SW model including a homogeneous phase noise caused by SMFFs, and then reveal that such phase noise induces decoherence of single-excitation entanglement between two atomic ensembles in a repeater link. For verify this conclusion, we experimentally prepare single-excitation entanglement between two SWs in a cold atomic ensemble and observe entanglement decoherence induced by SMFFs. Limited by SMFFs, the evaluated lifetime of the single-excitation entanglement over a repeater link is ~135 ms, even if the link uses optical-lattice atoms as nodes. We then present an approach that may overcome such limitation and extend this lifetime to ~1.7 s.

quant-ph↗
arXiv2023

Rhythm-controllable Attention with High Robustness for Long Sentence Speech Synthesis

Regressive Text-to-Speech (TTS) system utilizes attention mechanism to generate alignment between text and acoustic feature sequence. Alignment determines synthesis robustness (e.g, the occurence of skipping, repeating, and collapse) and rhythm via duration control. However, current attention algorithms used in speech synthesis cannot control rhythm using external duration information to generate natural speech while ensuring robustness. In this study, we propose Rhythm-controllable Attention (RC-Attention) based on Tracotron2, which improves robustness and naturalness simultaneously. Proposed attention adopts a trainable scalar learned from four kinds of information to achieve rhythm control, which makes rhythm control more robust and natural, even when synthesized sentences are extremely longer than training corpus. We use word errors counting and AB preference test to measure robustness of proposed method and naturalness of synthesized speech, respectively. Results shows that RC-Attention has the lowest word error rate of nearly 0.6%, compared with 11.8% for baseline system. Moreover, nearly 60% subjects prefer to the speech synthesized with RC-Attention to that with Forward Attention, because the former has more natural rhythm.

cs.AI↗
arXiv2023

M2-CTTS: End-to-End Multi-scale Multi-modal Conversational Text-to-Speech Synthesis

Conversational text-to-speech (TTS) aims to synthesize speech with proper prosody of reply based on the historical conversation. However, it is still a challenge to comprehensively model the conversation, and a majority of conversational TTS systems only focus on extracting global information and omit local prosody features, which contain important fine-grained information like keywords and emphasis. Moreover, it is insufficient to only consider the textual features, and acoustic features also contain various prosody information. Hence, we propose M2-CTTS, an end-to-end multi-scale multi-modal conversational text-to-speech system, aiming to comprehensively utilize historical conversation and enhance prosodic expression. More specifically, we design a textual context module and an acoustic context module with both coarse-grained and fine-grained modeling. Experimental results demonstrate that our model mixed with fine-grained context information and additionally considering acoustic features achieves better prosody performance and naturalness in CMOS tests.

cs.SD↗
arXiv2023

Multi-Stage Coarse-to-Fine Contrastive Learning for Conversation Intent Induction

Intent recognition is critical for task-oriented dialogue systems. However, for emerging domains and new services, it is difficult to accurately identify the key intent of a conversation due to time-consuming data annotation and comparatively poor model transferability. Therefore, the automatic induction of dialogue intention is very important for intelligent dialogue systems. This paper presents our solution to Track 2 of Intent Induction from Conversations for Task-Oriented Dialogue at the Eleventh Dialogue System Technology Challenge (DSTC11). The essence of intention clustering lies in distinguishing the representation of different dialogue utterances. The key to automatic intention induction is that, for any given set of new data, the sentence representation obtained by the model can be well distinguished from different labels. Therefore, we propose a multi-stage coarse-to-fine contrastive learning model training scheme including unsupervised contrastive learning pre-training, supervised contrastive learning pre-training, and fine-tuning with joint contrastive learning and clustering to obtain a better dialogue utterance representation model for the clustering task. In the released DSTC11 Track 2 evaluation results, our proposed system ranked first on both of the two subtasks of this Track.

cs.CL↗
arXiv2023

DKT-STDRL: Spatial and Temporal Representation Learning Enhanced Deep Knowledge Tracing for Learning Performance Prediction

Knowledge tracing (KT) serves as a primary part of intelligent education systems. Most current KTs either rely on expert judgments or only exploit a single network structure, which affects the full expression of learning features. To adequately mine features of students' learning process, Deep Knowledge Tracing Based on Spatial and Temporal Deep Representation Learning for Learning Performance Prediction (DKT-STDRL) is proposed in this paper. DKT-STDRL extracts spatial features from students' learning history sequence, and then further extracts temporal features to extract deeper hidden information. Specifically, firstly, the DKT-STDRL model uses CNN to extract the spatial feature information of students' exercise sequences. Then, the spatial features are connected with the original students' exercise features as joint learning features. Then, the joint features are input into the BiLSTM part. Finally, the BiLSTM part extracts the temporal features from the joint learning features to obtain the prediction information of whether the students answer correctly at the next time step. Experiments on the public education datasets ASSISTment2009, ASSISTment2015, Synthetic-5, ASSISTchall, and Statics2011 prove that DKT-STDRL can achieve better prediction effects than DKT and CKT.

cs.LG↗
arXiv2023

Generation of highly retrievable atom photon entanglement with a millisecond lifetime via a spatially multiplexed cavity

Qubit memory that is entangled with photonic qubit is the building block for long distance quantum repeaters. Cavity enhanced and long lived spin wave photon entanglement has been demonstrated by applying dual laser beams onto optical-lattice atoms. However, owing to cross readouts by two beams, retrieval efficiency of spin wave qubit is decreased by one quarter compared to that of single mode spin wave at all storage times. Here, by coupling cold atoms to two modes of a polarization interferometer based cavity, we achieve perfect qubit retrieval in cavity enhanced and long lived atom photon entanglement. A write laser beam is applied onto cold atoms, we then create a magnetic field insensitive spin wave qubit that is entangled with the photonic qubit encoded onto two arms of the interferometer. The spin wave qubit is retrieved by a read beam, which avoids the cross readouts. Our experiment demonstrates 540μs storage time at 50% intrinsic qubit retrieval efficiency, which is 13.5 times longer than the best reported result.

quant-ph↗
arXiv2022

Towards Lightweight Black-Box Attacks against Deep Neural Networks

Black-box attacks can generate adversarial examples without accessing the parameters of target model, largely exacerbating the threats of deployed deep neural networks (DNNs). However, previous works state that black-box attacks fail to mislead target models when their training data and outputs are inaccessible. In this work, we argue that black-box attacks can pose practical attacks in this extremely restrictive scenario where only several test samples are available. Specifically, we find that attacking the shallow layers of DNNs trained on a few test samples can generate powerful adversarial examples. As only a few samples are required, we refer to these attacks as lightweight black-box attacks. The main challenge to promoting lightweight attacks is to mitigate the adverse impact caused by the approximation error of shallow layers. As it is hard to mitigate the approximation error with few available samples, we propose Error TransFormer (ETF) for lightweight attacks. Namely, ETF transforms the approximation error in the parameter space into a perturbation in the feature space and alleviates the error by disturbing features. In experiments, lightweight black-box attacks with the proposed ETF achieve surprising results. For example, even if only 1 sample per category available, the attack success rate in lightweight black-box attacks is only about 3% lower than that of the black-box attacks with complete training data.

cs.LG↗
arXiv2022

A Keypoint Based Enhancement Method for Audio Driven Free View Talking Head Synthesis

Audio driven talking head synthesis is a challenging task that attracts increasing attention in recent years. Although existing methods based on 2D landmarks or 3D face models can synthesize accurate lip synchronization and rhythmic head pose for arbitrary identity, they still have limitations, such as the cut feeling in the mouth mapping and the lack of skin highlights. The morphed region is blurry compared to the surrounding face. A Keypoint Based Enhancement (KPBE) method is proposed for audio driven free view talking head synthesis to improve the naturalness of the generated video. Firstly, existing methods were used as the backend to synthesize intermediate results. Then we used keypoint decomposition to extract video synthesis controlling parameters from the backend output and the source image. After that, the controlling parameters were composited to the source keypoints and the driving keypoints. A motion field based method was used to generate the final image from the keypoint representation. With keypoint representation, we overcame the cut feeling in the mouth mapping and the lack of skin highlights. Experiments show that our proposed enhancement method improved the quality of talking-head videos in terms of mean opinion score.

cs.CV↗
arXiv2022

$Λ_b\to p$ transition form factors in perturbative QCD

We reanalyze the $Λ_b\to p$ transition form factors in the perturbative QCD (PQCD) approach by including higher-twist light-cone distribution amplitudes (LCDAs) of a $Λ_b$ baryon and a proton. The previous PQCD evaluation performed decades ago with only the leading-twist $Λ_b$ baryon and proton LCDAs gave the form factors, which are two orders of magnitude smaller than indicated by experimental data. We find that the twist-4 $Λ_b$ baryon LCDAs and the twist-4 and -5 proton LCDAs contribute dominantly, and the enhanced form factors become consistent with those from lattice QCD and other nonperturbative methods. The estimated branching ratios of the semileptonic decays $Λ_b\to p\ell\barν_\ell$ and the hadronic decay $Λ_b\to pπ$ are also close to the data. It implies that the $b$ quark mass is not really heavy enough, and higher-power contributions play a crucial role, similar to the observation made in analyses of $B$ meson transition form factors. With the formalism established in this work, we are ready to study various exclusive heavy baryon decays systematically in the PQCD approach.

hep-ph↗
arXiv2022

Study of $B_{(s)}^0 \to ϕϕ\to (K^+K^-)(K^+K^-)$ decays in the perturbative QCD approach

In this work, we make a detailed analysis on the penguin-dominant processes $B_{(s)}^0 \to ϕϕ\to (K^+K^-)(K^+K^-)$ in the perturbative QCD (PQCD) approach. In addition to the dominant $P$-wave resonance, the scalar background $f_0(980) \to K^+K^-$ is also accounted for. We improve the Gegenbauer moments in $KK$ two-meson distribution amplitudes by fitting the PQCD factorization formulas to measured branching ratios of three-body and four-body $B$ decays. We extract the branching ratios of two-body $B_{(s)}^0 \to ϕϕ$ decays from the corresponding four-body decay modes and calculate the relevant polarization fractions together with two relative phases $ϕ_{\parallel,\perp}$, which are consistent with the previous theoretical predictions. The PQCD predictions for the "true" triple product asymmetries (TPAs) are zero which are expected in the standard model due to the vanishing weak phase difference, and support the current data reported by the CDF and LHCb Collaborations. A large "fake" TPA $\mathcal{A}_\text{T-fake}^1=30.4\%$ of the decay $B^0_s \to ϕϕ\to (K^+K^-)(K^+K^-)$ is predicted for the first time, which indicates the presence of the significant final-state interactions. The TPAs of the rare decay channel $B^0 \to ϕϕ\to (K^+K^-)(K^+K^-)$ are also predicted and can be tested in the near future.

hep-ph↗
arXiv2022

ECAPA-TDNN for Multi-speaker Text-to-speech Synthesis

In recent years, neural network based methods for multi-speaker text-to-speech synthesis (TTS) have made significant progress. However, the current speaker encoder models used in these methods still cannot capture enough speaker information. In this paper, we focus on accurate speaker encoder modeling and propose an end-to-end method that can generate high-quality speech and better similarity for both seen and unseen speakers. The proposed architecture consists of three separately trained components: a speaker encoder based on the state-of-the-art ECAPA-TDNN model which is derived from speaker verification task, a FastSpeech2 based synthesizer, and a HiFi-GAN vocoder. The comparison among different speaker encoder models shows our proposed method can achieve better naturalness and similarity. To efficiently evaluate our synthesized speech, we are the first to adopt deep learning based automatic MOS evaluation methods to assess our results, and these methods show great potential in automatic speech quality assessment.

cs.SD↗
arXiv2022

$CP$-violating observables in four-body $B\rightarrow ϕ(\rightarrow K\bar K)K^*(\rightarrow Kπ)$ decays

We analyse the four-body $B\rightarrow ϕ(\rightarrow K\bar K)K^*(\rightarrow Kπ)$ decays in the perturbative QCD approach,where the invariant mass of $K\bar K$($Kπ$) system is limited in a window of $\pm 15$ MeV ($\pm150$ MeV) around the $ϕ(K^*(892))$ mass.In addition to the P wave resonances,two important S wave backgrounds in the selected invariant mass region are also accounted for. Angular momentum conservation allows six helicity amplitudes to contribute,including three P waves, two single S waves,and one double S wave. We calculated the branching ratio for each component and found sizable S wave contributions,coincide with the experimental observation.The obtained branching ratios of $B^{0(+)}\rightarrow ϕK^{*0(+)}$ are comparable with the previous predictions and support the measurements, whereas the predicted $\mathcal{B}(B^0_s\rightarrow ϕ\bar K^{*0})$ is smaller than the world average. The longitudinal polarizations are predicted to be around 0.7,consistent with previous PQCD results but larger than the data. Aside from the direct CP asymmetries,the true and fake triple product asymmetries(TPAs) are calculated in this work. In the case of neutral modes, both direct CP asymmetries and true TPAs are expected to be zero due to the vanishing weak phase difference. The direct CP asymmetries for the $B^+$ mode are predicted to be tiny,since the tree contributions are suppressed with respect to the penguin ones. The true asymmetries have shown no significant deviations from zero.In contrast,large fake asymmetries are observed in these decays,indicating the presence of significant final state interactions.We give the predictions of the S wave induced TPAs for the first time,which is consistent with LHCb data and would be checked with future measurements from Belle and BABAR experiments if the S wave components can be properly taken into account in angular analysis.

hep-ph↗