arXiv subjects

Ming Lei

Publications and source records attributed to Ming Lei.

At least 55 records · Page 3Linked to original sources

arXiv2023

High-speed image reconstruction for nonlinear structured illumination microscopy

By exploiting the nonlinear responses of the fluorescent probes, the spatial resolution of structured illumination microscopy(SIM) can be further increased. However, due to the complex reconstruction process, the traditional reconstruction method of nonlinear structured illumination microscopy (NL-SIM) is relatively slow, which brings a great challenge to realizing real-time display of super-resolution results. To address these issues, an accelerated NL-SIM reconstruction algorithm was developed by extending a high-speed reconstruction framework, Joint Space and Frequency Reconstruction (JSFR) to NL-SIM. We anticipate that this algorithm will facilitate NL- SIM becoming a routine tool in biomedical laboratories.

physics.optics↗

arXiv2023

Versatile manipulation of low-refractive-index particles using customized optical building blocks

Low-refractive-index (LRI) particles play significant roles in physics, drug delivery, biomedical science, and other fields. However, they have not attained sufficient utilization in active manipulation due to the repulsive effect of light. Here, we demonstrate the establishment of optical building blocks (OBBs) to fulfill the demands of versatile manipulation of LRI particles. The OBBs are generated by assembling generalized perfect optical vortices based on the free lens modulation (FLM) method, by which the beams shape, intensity, and position can be elaborately designed with size independent of topological charge. Using the OBBs with high quality and high efficiency, we realized rotating LRI particles along arbitrary trajectories with controllable speed and parallel manipulation of multiple LRI particles. Importantly, we further achieved the sorting of LRI particles by size with specially structured OBBs. With unprecedented flexibility and quality, OBBs provide tremendous potential in optical trapping, lithography, and biomedicine.

physics.optics↗

arXiv2023

Fast List Decoding of High-Rate Polar Codes

Due to the ability to provide superior error-correction performance, the successive cancellation list (SCL) algorithm is widely regarded as one of the most promising decoding algorithms for polar codes with short-to-moderate code lengths. However, the application of SCL decoding in low-latency communication scenarios is limited due to its sequential nature. To reduce the decoding latency, developing tailored fast and efficient list decoding algorithms of specific polar substituent codes (special nodes) is a promising solution. Recently, fast list decoding algorithms are proposed by considering special nodes with low code rates. Aiming to further speedup the SCL decoding, this paper presents fast list decoding algorithms for two types of high-rate special nodes, namely single-parity-check (SPC) nodes and sequence rate one or single-parity-check (SR1/SPC) nodes. In particular, we develop two classes of fast list decoding algorithms for these nodes, where the first class uses a sequential decoding procedure to yield decoding latency that is linear with the list size, and the second further parallelizes the decoding process by pre-determining the redundant candidate paths offline. Simulation results show that the proposed list decoding algorithms are able to achieve up to 70.7\% lower decoding latency than state-of-the-art fast SCL decoders, while exhibiting the same error-correction performance.

cs.IT↗

arXiv2023

Joint Scattering Environment Sensing and Channel Estimation Based on Non-stationary Markov Random Field

This paper considers an integrated sensing and communication system, where some radar targets also serve as communication scatterers. A location domain channel modeling method is proposed based on the position of targets and scatterers in the scattering environment, and the resulting radar and communication channels exhibit a two-dimensional (2-D) joint burst sparsity. We propose a joint scattering environment sensing and channel estimation scheme to enhance the target/scatterer localization and channel estimation performance simultaneously, where a spatially non-stationary Markov random field (MRF) model is proposed to capture the 2-D joint burst sparsity. An expectation maximization (EM) based method is designed to solve the joint estimation problem, where the E-step obtains the Bayesian estimation of the radar and communication channels and the M-step automatically learns the dynamic position grid and prior parameters in the MRF. However, the existing sparse Bayesian inference methods used in the E-step involve a high-complexity matrix inverse per iteration. Moreover, due to the complicated non-stationary MRF prior, the complexity of M-step is exponentially large. To address these difficulties, we propose an inverse-free variational Bayesian inference algorithm for the E-step and a low-complexity method based on pseudo-likelihood approximation for the M-step. In the simulations, the proposed scheme can achieve a better performance than the state-of-the-art method while reducing the computational overhead significantly.

eess.SP↗

arXiv2023

Hybrid organic-inorganic two-dimensional metal carbide MXenes with amido- and imido-terminated surfaces

Two-dimensional (2D) transition-metal carbides and nitrides (MXenes) show impressive performance in applications, such as supercapacitors, batteries, electromagnetic interference shielding, or electrocatalysis. These materials combine the electronic and mechanical properties of 2D inorganic crystals with chemically modifiable surfaces, and surface-engineered MXenes represent an ideal platform for fundamental and applied studies of interfaces in 2D functional materials. A natural step in structural engineering of MXene compounds is the development and understanding of MXenes with various organic functional groups covalently bound to inorganic 2D sheets. Such hybrid structures have the potential to unite the tailorability of organic molecules with the unique electronic properties of inorganic 2D solids. Here, we introduce a new family of hybrid MXenes (h-MXenes) with amido- and imido-bonding between organic and inorganic parts. The description of h-MXene structure requires an intricate mix of concepts from the fields of coordination chemistry, self-assembled monolayers (SAMs) and surface science. The optical properties of h-MXenes reveal coherent coupling between the organic and inorganic components. h-MXenes also show superior stability against hydrolysis in aqueous solutions.

cond-mat.mtrl-sci↗

arXiv2023

FedSpeech: Federated Text-to-Speech with Continual Learning

Federated learning enables collaborative training of machine learning models under strict privacy restrictions and federated text-to-speech aims to synthesize natural speech of multiple users with a few audio training samples stored in their devices locally. However, federated text-to-speech faces several challenges: very few training samples from each speaker are available, training samples are all stored in local device of each user, and global model is vulnerable to various attacks. In this paper, we propose a novel federated learning architecture based on continual learning approaches to overcome the difficulties above. Specifically, 1) we use gradual pruning masks to isolate parameters for preserving speakers' tones; 2) we apply selective masks for effectively reusing knowledge from tasks; 3) a private speaker embedding is introduced to keep users' privacy. Experiments on a reduced VCTK dataset demonstrate the effectiveness of FedSpeech: it nearly matches multi-task training in terms of multi-speaker speech quality; moreover, it sufficiently retains the speakers' tones and even outperforms the multi-task training in the speaker similarity experiment.

eess.AS↗

arXiv2022

Fast Successive-Cancellation Decoding of Polar Codes with Sequence Nodes

Due to the sequential nature of the successive-cancellation (SC) algorithm, the decoding of polar codes suffers from significant decoding latencies. Fast SC decoding is able to speed up the SC decoding process, by implementing parallel decoders at the intermediate levels of the SC decoding tree for some special nodes with specific information and frozen bit patterns. To further improve the parallelism of SC decoding, this paper present a new class of special nodes composed of a sequence of rate one or single-parity-check (SR1/SPC) nodes, which can be easily found especially in high-rate polar code and is able to envelop a wide variety of existing special node types. Then, we analyse the parity constraints caused by the frozen bits in each descendant node, such that the decoding performance of the SR1/SPC node can be preserved once the parity constraints are satisfied. Finally, a generalized fast decoding algorithm is proposed to decode SR1/SPC nodes efficiently, where the corresponding parity constraints are taken into consideration. Simulation results show that the proposed decoding algorithm of the SR1/SPC node can achieve near-ML performance, and the overall decoding latency can be reduced by 43.8% as compared to the state-of-the-art fast SC decoder.

cs.IT↗

arXiv2022

ProsoSpeech: Enhancing Prosody With Quantized Vector Pre-training in Text-to-Speech

Expressive text-to-speech (TTS) has become a hot research topic recently, mainly focusing on modeling prosody in speech. Prosody modeling has several challenges: 1) the extracted pitch used in previous prosody modeling works have inevitable errors, which hurts the prosody modeling; 2) different attributes of prosody (e.g., pitch, duration and energy) are dependent on each other and produce the natural prosody together; and 3) due to high variability of prosody and the limited amount of high-quality data for TTS training, the distribution of prosody cannot be fully shaped. To tackle these issues, we propose ProsoSpeech, which enhances the prosody using quantized latent vectors pre-trained on large-scale unpaired and low-quality text and speech data. Specifically, we first introduce a word-level prosody encoder, which quantizes the low-frequency band of the speech and compresses prosody attributes in the latent prosody vector (LPV). Then we introduce an LPV predictor, which predicts LPV given word sequence. We pre-train the LPV predictor on large-scale text and low-quality speech data and fine-tune it on the high-quality TTS dataset. Finally, our model can generate expressive speech conditioned on the predicted LPV. Experimental results show that ProsoSpeech can generate speech with richer prosody compared with baseline methods.

eess.AS↗

arXiv2021

Ultra-wideband electrostrictive mechanical antenna

Conventional mechanical antennas provide a strategy in long-wave communication with a surprisingly compact size below 1/1,000 of the wavelength. However, the narrow bandwidth and weak field intensity seriously hamper its practical applications. Here, we present a mechanical antenna based on the electrostrictive effect of PMN-PT-based relaxor ferroelectric ceramic to improve radiation capacity and achieve ultra-wideband characteristics (10 kHz - 1 MHz, the relative bandwidth is beyond 196%). Determined by the different underlying mechanism, the mechanical antenna based on the electrostrictive effect exhibits excellent communication properties from traditional mechanical antennas. The functions of signal coding, transmitting, receiving, and decoding were experimentally demonstrated. This approach offers a promising way of constructing mechanical antennas for long-wave communication.

physics.ins-det↗

arXiv2021

Speaker Embedding-aware Neural Diarization for Flexible Number of Speakers with Textual Information

Overlapping speech diarization is always treated as a multi-label classification problem. In this paper, we reformulate this task as a single-label prediction problem by encoding the multi-speaker labels with power set. Specifically, we propose the speaker embedding-aware neural diarization (SEND) method, which predicts the power set encoded labels according to the similarities between speech features and given speaker embeddings. Our method is further extended and integrated with downstream tasks by utilizing the textual information, which has not been well studied in previous literature. The experimental results show that our method achieves lower diarization error rate than the target-speaker voice activity detection. When textual information is involved, the diarization errors can be further reduced. For the real meeting scenario, our method can achieve 34.11% relative improvement compared with the Bayesian hidden Markov model based clustering algorithm.

cs.SD↗

arXiv2021

Zero-order-free complex beam shaping

The unwanted zero-order diffraction is still an issue in beam shaping using pixelated spatial light modulators. In this paper, we report a new approach for zero-order free beam shaping by designing an asymmetric triangle reflector and introducing a digital blazed grating and a digital lens to the phase hologram addressed onto the spatial light modulator. By adding the digital lens phase to the previously reported complex-amplitude coding algorithms, we realized the generation of complex beams without the burden of zero-order diffraction. We comparatively investigated the produced complex light fields using the modified complex-amplitude coding algorithms to validate the proposed method.

physics.optics↗

arXiv2021

BeamTransformer: Microphone Array-based Overlapping Speech Detection

We propose BeamTransformer, an efficient architecture to leverage beamformer's edge in spatial filtering and transformer's capability in context sequence modeling. BeamTransformer seeks to optimize modeling of sequential relationship among signals from different spatial direction. Overlapping speech detection is one of the tasks where such optimization is favorable. In this paper we effectively apply BeamTransformer to detect overlapping segments. Comparing to single-channel approach, BeamTransformer exceeds in learning to identify the relationship among different beam sequences and hence able to make predictions not only from the acoustic signals but also the localization of the source. The results indicate that a successful incorporation of microphone array signals can lead to remarkable gains. Moreover, BeamTransformer takes one step further, as speech from overlapped speakers have been internally separated into different beams.

cs.SD↗

arXiv2021

Extremely Low Footprint End-to-End ASR System for Smart Device

Recently, end-to-end (E2E) speech recognition has become popular, since it can integrate the acoustic, pronunciation and language models into a single neural network, which outperforms conventional models. Among E2E approaches, attention-based models, e.g. Transformer, have emerged as being superior. Such models have opened the door to deployment of ASR on smart devices, however they still suffer from requiring a large number of model parameters. We propose an extremely low footprint E2E ASR system for smart devices, to achieve the goal of satisfying resource constraints without sacrificing recognition accuracy. We design cross-layer weight sharing to improve parameter efficiency and further exploit model compression methods including sparsification and quantization, to reduce memory storage and boost decoding efficiency. We evaluate our approaches on the public AISHELL-1 and AISHELL-2 benchmarks. On the AISHELL-2 task, the proposed method achieves more than 10x compression (model size reduces from 248 to 24MB), at the cost of only minor performance loss (CER reduces from 6.49% to 6.92%).

cs.SD↗

arXiv2021

EMOVIE: A Mandarin Emotion Speech Dataset with a Simple Emotional Text-to-Speech Model

Recently, there has been an increasing interest in neural speech synthesis. While the deep neural network achieves the state-of-the-art result in text-to-speech (TTS) tasks, how to generate a more emotional and more expressive speech is becoming a new challenge to researchers due to the scarcity of high-quality emotion speech dataset and the lack of advanced emotional TTS model. In this paper, we first briefly introduce and publicly release a Mandarin emotion speech dataset including 9,724 samples with audio files and its emotion human-labeled annotation. After that, we propose a simple but efficient architecture for emotional speech synthesis called EMSpeech. Unlike those models which need additional reference audio as input, our model could predict emotion labels just from the input text and generate more expressive speech conditioned on the emotion embedding. In the experiment phase, we first validate the effectiveness of our dataset by an emotion classification task. Then we train our model on the proposed dataset and conduct a series of subjective evaluations. Finally, by showing a comparable performance in the emotional speech synthesis task, we successfully demonstrate the ability of the proposed model.

cs.CL↗

arXiv2021

Large cross-polarized Raman signal in CrI$_3$: A first-principles study

We find unusually large cross-polarized (and anti-symmetric) Raman signature of A$_{\rm g}$ phonon mode in CrI$_3$, in agreement with experiments. The signal is present only when the following three effects are considered in concert: ferromagnetism on Cr atoms, spin-orbit interaction, and resonant effects. Somewhat surprisingly, we find that the relevant spin-orbit interaction potential originates from iodine atoms, despite magnetism being mostly on chromium atoms. We analyze the Raman signature as a function of magnetic order, the direction of the magnetic moment, energy and polarization of light used for Raman scattering, as well as carrier lifetime. In addition to a strong cross-polarized Raman signal, we also find unusually strong phonon modulated magneto-optical Kerr effect (MOKE) in CrI$_3$.

cond-mat.mtrl-sci↗

arXiv2021

DeviceTTS: A Small-Footprint, Fast, Stable Network for On-Device Text-to-Speech

With the number of smart devices increasing, the demand for on-device text-to-speech (TTS) increases rapidly. In recent years, many prominent End-to-End TTS methods have been proposed, and have greatly improved the quality of synthesized speech. However, to ensure the qualified speech, most TTS systems depend on large and complex neural network models, and it's hard to deploy these TTS systems on-device. In this paper, a small-footprint, fast, stable network for on-device TTS is proposed, named as DeviceTTS. DeviceTTS makes use of a duration predictor as a bridge between encoder and decoder so as to avoid the problem of words skipping and repeating in Tacotron. As we all know, model size is a key factor for on-device TTS. For DeviceTTS, Deep Feedforward Sequential Memory Network (DFSMN) is used as the basic component. Moreover, to speed up inference, mix-resolution decoder is proposed for balance the inference speed and speech quality. Experiences are done with WORLD and LPCNet vocoder. Finally, with only 1.4 million model parameters and 0.099 GFLOPS, DeviceTTS achieves comparable performance with Tacotron and FastSpeech. As far as we know, the DeviceTTS can meet the needs of most of the devices in practical application.

eess.AS↗

arXiv2020

Simplified Self-Attention for Transformer-based End-to-End Speech Recognition

Transformer models have been introduced into end-to-end speech recognition with state-of-the-art performance on various tasks owing to their superiority in modeling long-term dependencies. However, such improvements are usually obtained through the use of very large neural networks. Transformer models mainly include two submodules - position-wise feedforward layers and self-attention (SAN) layers. In this paper, to reduce the model complexity while maintaining good performance, we propose a simplified self-attention (SSAN) layer which employs FSMN memory block instead of projection layers to form query and key vectors for transformer-based end-to-end speech recognition. We evaluate the SSAN-based and the conventional SAN-based transformers on the public AISHELL-1, internal 1000-hour and 20,000-hour large-scale Mandarin tasks. Results show that our proposed SSAN-based transformer model can achieve over 20% relative reduction in model parameters and 6.7% relative CER reduction on the AISHELL-1 task. With impressively 20% parameter reduction, our model shows no loss of recognition performance on the 20,000-hour large-scale task.

cs.SD↗

arXiv2020

Universal ASR: Unifying Streaming and Non-Streaming ASR Using a Single Encoder-Decoder Model

Recently, online end-to-end ASR has gained increasing attention. However, the performance of online systems still lags far behind that of offline systems, with a large gap in quality of recognition. For specific scenarios, we can trade-off between performance and latency, and can train multiple systems with different delays to match the performance and latency requirements of various application scenarios. In this work, in contrast to trading-off between performance and latency, we envisage a single system that can match the needs of different scenarios. We propose a novel architecture, termed Universal ASR that can unify streaming and non-streaming ASR models into one system. The embedded streaming ASR model can configure different delays according to requirements to obtain real-time recognition results, while the non-streaming model is able to refresh the final recognition result for better performance. We have evaluated our approach on the public AISHELL-2 benchmark and an industrial-level 20,000-hour Mandarin speech recognition task. The experimental results show that the Universal ASR provides an efficient mechanism to integrate streaming and non-streaming models that can recognize speech quickly and accurately. On the AISHELL-2 task, Universal ASR comfortably outperforms other state-of-the-art systems.

cs.SD↗