arXiv subjects

Jing Liu

Publications and source records attributed to Jing Liu.

At least 451 records · Page 25Linked to original sources

arXiv2023

TOME: A Two-stage Approach for Model-based Retrieval

Recently, model-based retrieval has emerged as a new paradigm in text retrieval that discards the index in the traditional retrieval model and instead memorizes the candidate corpora using model parameters. This design employs a sequence-to-sequence paradigm to generate document identifiers, which enables the complete capture of the relevance between queries and documents and simplifies the classic indexretrieval-rerank pipeline. Despite its attractive qualities, there remain several major challenges in model-based retrieval, including the discrepancy between pre-training and fine-tuning, and the discrepancy between training and inference. To deal with the above challenges, we propose a novel two-stage model-based retrieval approach called TOME, which makes two major technical contributions, including the utilization of tokenized URLs as identifiers and the design of a two-stage generation architecture. We also propose a number of training strategies to deal with the training difficulty as the corpus size increases. Extensive experiments and analysis on MS MARCO and Natural Questions demonstrate the effectiveness of our proposed approach, and we investigate the scaling laws of TOME by examining various influencing factors.

cs.IR↗

arXiv2023

Two Low-complexity Efficient Beamformers for IRS-and-UAV-aided Directional Modulation Networks

As the excellent tools for aiding communication,intelligent reflecting surface (IRS) and unmanned aerial vehicle (UAV) can extend the coverage area, remove blind area, and achieve a dramatic rate improvement. In this paper, we improve the secrecy rate (SR) performance at directional modulation (DM) networks using IRS and UAV in combination. To fully explore the benefits of IRS and UAV, two efficient methods are proposed to enhance SR performance. The first approach computes the confidential message (CM) beamforming vector by maximizing the SR, and the signal-to-leakage-noise ratio (SLNR) method is used to optimize the IRS phase shift matrix, which is called Max-SR-SLNR. Here, Eve is maximally interfered by transmitting artificial noise (AN) along the direct path and null-space projection (NSP) on the remaining two channels. To reduce the computational complexity, the CM, AN beamforming and IRS phase shift design are independently designed in the following methods. The CM beamforming vector is constructed based on maximum ratio transmission (MRT) criteria along the channel from Alice-to-IRS, and phase shift matrix of IRS is directly given by phase alignment (PA) method. This method is called MRT-NSP-PA. Simulation results show that the SR performance of the Max-SR-SLNR method outperforms the MRT-NSP-PA method in the cases of small-scale and medium-scale IRSs, and the latter approaches the former in performance as IRS tends to lager-scale.

cs.IT↗

arXiv2023

Enhanced-rate Iterative Beamformers for Active IRS-assisted Wireless Communications

Compared to passive intelligent reflecting surface (IRS), active IRS is viewed as a more efficient promising technique to combat the double-fading impact in IRS-aided wireless network. In this paper, in order to boost the achievable rate of user in such a wireless network, three enhanced-rate iterative beamforming methods are proposed by designing the amplifying factors and the corresponding phases at active IRS. The first method, maximizing the simplified signal-to-noise ratio (Max-SSNR) is designed by omitting the cross-term in the definition of rate. Using the Rayleigh-Ritz (RR) theorem, Max-SSNR-RR is proposed to iteratively optimize the norm of beamforming vector and its associated normalized vector. In addition, generalized maximum ratio reflection (GMRR) is presented with a closed-form expression, which is motivated by the maximum ratio combining. To further improve rate, maximizing SNR (Max-SNR) is designed by fractional programming (FP), which is called Max-SNR-FP. Simulation results show that the proposed three methods make an obvious rate enhancement over Max-reflecting signal-to-noise ratio (Max-RSNR), maximum ratio reflection (MRR), selective ratio reflecting (SRR), equal gain reflection (EGR) and passive IRS, and are in increasing order of rate performance as follows: Max-SSNR-RR, GMRR, and Max-SNR-FP.

eess.SP↗

arXiv2023

Robust Acoustic and Semantic Contextual Biasing in Neural Transducers for Speech Recognition

Attention-based contextual biasing approaches have shown significant improvements in the recognition of generic and/or personal rare-words in End-to-End Automatic Speech Recognition (E2E ASR) systems like neural transducers. These approaches employ cross-attention to bias the model towards specific contextual entities injected as bias-phrases to the model. Prior approaches typically relied on subword encoders for encoding the bias phrases. However, subword tokenizations are coarse and fail to capture granular pronunciation information which is crucial for biasing based on acoustic similarity. In this work, we propose to use lightweight character representations to encode fine-grained pronunciation features to improve contextual biasing guided by acoustic similarity between the audio and the contextual entities (termed acoustic biasing). We further integrate pretrained neural language model (NLM) based encoders to encode the utterance's semantic context along with contextual entities to perform biasing informed by the utterance's semantic context (termed semantic biasing). Experiments using a Conformer Transducer model on the Librispeech dataset show a 4.62% - 9.26% relative WER improvement on different biasing list sizes over the baseline contextual model when incorporating our proposed acoustic and semantic biasing approach. On a large-scale in-house dataset, we observe 7.91% relative WER improvement compared to our baseline model. On tail utterances, the improvements are even more pronounced with 36.80% and 23.40% relative WER improvements on Librispeech rare words and an in-house testset respectively.

cs.CL↗

arXiv2023

Single-path Bit Sharing for Automatic Loss-aware Model Compression

Network pruning and quantization are proven to be effective ways for deep model compression. To obtain a highly compact model, most methods first perform network pruning and then conduct network quantization based on the pruned model. However, this strategy may ignore that they would affect each other and thus performing them separately may lead to sub-optimal performance. To address this, performing pruning and quantization jointly is essential. Nevertheless, how to make a trade-off between pruning and quantization is non-trivial. Moreover, existing compression methods often rely on some pre-defined compression configurations. Some attempts have been made to search for optimal configurations, which however may take unbearable optimization cost. To address the above issues, we devise a simple yet effective method named Single-path Bit Sharing (SBS). Specifically, we first consider network pruning as a special case of quantization, which provides a unified view for pruning and quantization. We then introduce a single-path model to encode all candidate compression configurations. In this way, the configuration search problem is transformed into a subset selection problem, which significantly reduces the number of parameters, computational cost and optimization difficulty. Relying on the single-path model, we further introduce learnable binary gates to encode the choice of bitwidth. By jointly training the binary gates in conjunction with network parameters, the compression configurations of each layer can be automatically determined. Extensive experiments on both CIFAR-100 and ImageNet show that SBS is able to significantly reduce computational cost while achieving promising performance. For example, our SBS compressed MobileNetV2 achieves 22.6x Bit-Operation (BOP) reduction with only 0.1% drop in the Top-1 accuracy.

cs.CV↗

arXiv2023

A Survey on Efficient Training of Transformers

Recent advances in Transformers have come with a huge requirement on computing resources, highlighting the importance of developing efficient training techniques to make Transformer training faster, at lower cost, and to higher accuracy by the efficient use of computation and memory resources. This survey provides the first systematic overview of the efficient training of Transformers, covering the recent progress in acceleration arithmetic and hardware, with a focus on the former. We analyze and compare methods that save computation and memory costs for intermediate tensors during training, together with techniques on hardware/algorithm co-design. We finally discuss challenges and promising areas for future research.

cs.LG↗

arXiv2023

SLSG: Industrial Image Anomaly Detection by Learning Better Feature Embeddings and One-Class Classification

Industrial image anomaly detection under the setting of one-class classification has significant practical value. However, most existing models struggle to extract separable feature representations when performing feature embedding and struggle to build compact descriptions of normal features when performing one-class classification. One direct consequence of this is that most models perform poorly in detecting logical anomalies which violate contextual relationships. Focusing on more effective and comprehensive anomaly detection, we propose a network based on self-supervised learning and self-attentive graph convolution (SLSG) for anomaly detection. SLSG uses a generative pre-training network to assist the encoder in learning the embedding of normal patterns and the reasoning of position relationships. Subsequently, SLSG introduces the pseudo-prior knowledge of anomaly through simulated abnormal samples. By comparing the simulated anomalies, SLSG can better summarize the normal features and narrow down the hypersphere used for one-class classification. In addition, with the construction of a more general graph structure, SLSG comprehensively models the dense and sparse relationships among elements in the image, which further strengthens the detection of logical anomalies. Extensive experiments on benchmark datasets show that SLSG achieves superior anomaly detection performance, demonstrating the effectiveness of our method.

cs.CV↗

arXiv2023

RocketQAv2: A Joint Training Method for Dense Passage Retrieval and Passage Re-ranking

In various natural language processing tasks, passage retrieval and passage re-ranking are two key procedures in finding and ranking relevant information. Since both the two procedures contribute to the final performance, it is important to jointly optimize them in order to achieve mutual improvement. In this paper, we propose a novel joint training approach for dense passage retrieval and passage re-ranking. A major contribution is that we introduce the dynamic listwise distillation, where we design a unified listwise training approach for both the retriever and the re-ranker. During the dynamic distillation, the retriever and the re-ranker can be adaptively improved according to each other's relevance information. We also propose a hybrid data augmentation strategy to construct diverse training instances for listwise training approach. Extensive experiments show the effectiveness of our approach on both MSMARCO and Natural Questions datasets. Our code is available at https://github.com/PaddlePaddle/RocketQA.

cs.CL↗

arXiv2023

A Thorough Examination on Zero-shot Dense Retrieval

Recent years have witnessed the significant advance in dense retrieval (DR) based on powerful pre-trained language models (PLM). DR models have achieved excellent performance in several benchmark datasets, while they are shown to be not as competitive as traditional sparse retrieval models (e.g., BM25) in a zero-shot retrieval setting. However, in the related literature, there still lacks a detailed and comprehensive study on zero-shot retrieval. In this paper, we present the first thorough examination of the zero-shot capability of DR models. We aim to identify the key factors and analyze how they affect zero-shot retrieval performance. In particular, we discuss the effect of several key factors related to source training set, analyze the potential bias from the target dataset, and review and compare existing zero-shot DR models. Our findings provide important evidence to better understand and develop zero-shot DR models.

cs.CL↗

arXiv2023

Approximate Order-Preserving Pattern Mining for Time Series

The order-preserving pattern mining can be regarded as discovering frequent trends in time series, since the same order-preserving pattern has the same relative order which can represent a trend. However, in the case where data noise is present, the relative orders of many meaningful patterns are usually similar rather than the same. To mine similar relative orders in time series, this paper addresses an approximate order-preserving pattern (AOP) mining method based on (delta-gamma) distance to effectively measure the similarity, and proposes an algorithm called AOP-Miner to mine AOPs according to global and local approximation parameters. AOP-Miner adopts a pattern fusion strategy to generate candidate patterns generation and employs the screening strategy to calculate the supports of candidate patterns. Experimental results validate that AOP-Miner outperforms other competitive methods and can find more similar trends in time series.

cs.DB↗

arXiv2023

Production of dibaryon $d_{NΩ}$ in kaon induced reactions

In this work, we propose to investigate the $d_{NΩ}$ dibaryon production in the process $K^- p \rightarrow d_{NΩ} \barΞ^0$ by utilizing the kaon beam with the typical momentum to be around 10 GeV, which may be available at COMPASS, OKA@U-70 and SPS@CERN. The cross sections for $K^- p \rightarrow d_{NΩ} \barΞ^0$ are estimated and in particular, the magnitude of the cross sections are evaluated to be several hundreds nanobarn at $P_K=20$ GeV. Considering that $d_{NΩ}$ dibaryon dominantly decay into $ΞΛ$ and $ΞΣ$, we also estimate the cross sections for $K^- p \to Ξ^0 Λ\barΞ^0$ and $K^- p \to Ξ^- Σ^+ \barΞ^0$, where the $d_{NΩ}$ dibaryon can be observed in the invariant mass distributions of $Ξ^0 Λ$ and $Ξ^- Σ^+ $, respectively.

hep-ph↗

arXiv2023

PAIR: Leveraging Passage-Centric Similarity Relation for Improving Dense Passage Retrieval

Recently, dense passage retrieval has become a mainstream approach to finding relevant information in various natural language processing tasks. A number of studies have been devoted to improving the widely adopted dual-encoder architecture. However, most of the previous studies only consider query-centric similarity relation when learning the dual-encoder retriever. In order to capture more comprehensive similarity relations, we propose a novel approach that leverages both query-centric and PAssage-centric sImilarity Relations (called PAIR) for dense passage retrieval. To implement our approach, we make three major technical contributions by introducing formal formulations of the two kinds of similarity relations, generating high-quality pseudo labeled data via knowledge distillation, and designing an effective two-stage training procedure that incorporates passage-centric similarity relation constraint. Extensive experiments show that our approach significantly outperforms previous state-of-the-art models on both MSMARCO and Natural Questions datasets.

cs.IR↗

arXiv2023

Video Event Restoration Based on Keyframes for Video Anomaly Detection

Video anomaly detection (VAD) is a significant computer vision problem. Existing deep neural network (DNN) based VAD methods mostly follow the route of frame reconstruction or frame prediction. However, the lack of mining and learning of higher-level visual features and temporal context relationships in videos limits the further performance of these two approaches. Inspired by video codec theory, we introduce a brand-new VAD paradigm to break through these limitations: First, we propose a new task of video event restoration based on keyframes. Encouraging DNN to infer missing multiple frames based on video keyframes so as to restore a video event, which can more effectively motivate DNN to mine and learn potential higher-level visual features and comprehensive temporal context relationships in the video. To this end, we propose a novel U-shaped Swin Transformer Network with Dual Skip Connections (USTN-DSC) for video event restoration, where a cross-attention and a temporal upsampling residual skip connection are introduced to further assist in restoring complex static and dynamic motion object features in the video. In addition, we propose a simple and effective adjacent frame difference loss to constrain the motion consistency of the video sequence. Extensive experiments on benchmarks demonstrate that USTN-DSC outperforms most existing methods, validating the effectiveness of our method.

cs.CV↗

arXiv2023

Dual-Attention Neural Transducers for Efficient Wake Word Spotting in Speech Recognition

We present dual-attention neural biasing, an architecture designed to boost Wake Words (WW) recognition and improve inference time latency on speech recognition tasks. This architecture enables a dynamic switch for its runtime compute paths by exploiting WW spotting to select which branch of its attention networks to execute for an input audio frame. With this approach, we effectively improve WW spotting accuracy while saving runtime compute cost as defined by floating point operations (FLOPs). Using an in-house de-identified dataset, we demonstrate that the proposed dual-attention network can reduce the compute cost by $90\%$ for WW audio frames, with only $1\%$ increase in the number of parameters. This architecture improves WW F1 score by $16\%$ relative and improves generic rare word error rate by $3\%$ relative compared to the baselines.

cs.SD↗

arXiv2023

PROCTER: PROnunciation-aware ConTextual adaptER for personalized speech recognition in neural transducers

End-to-End (E2E) automatic speech recognition (ASR) systems used in voice assistants often have difficulties recognizing infrequent words personalized to the user, such as names and places. Rare words often have non-trivial pronunciations, and in such cases, human knowledge in the form of a pronunciation lexicon can be useful. We propose a PROnunCiation-aware conTextual adaptER (PROCTER) that dynamically injects lexicon knowledge into an RNN-T model by adding a phonemic embedding along with a textual embedding. The experimental results show that the proposed PROCTER architecture outperforms the baseline RNN-T model by improving the word error rate (WER) by 44% and 57% when measured on personalized entities and personalized rare entities, respectively, while increasing the model size (number of trainable parameters) by only 1%. Furthermore, when evaluated in a zero-shot setting to recognize personalized device names, we observe 7% WER improvement with PROCTER, as compared to only 1% WER improvement with text-only contextual attention

eess.AS↗

arXiv2023

Sounding Video Generator: A Unified Framework for Text-guided Sounding Video Generation

As a combination of visual and audio signals, video is inherently multi-modal. However, existing video generation methods are primarily intended for the synthesis of visual frames, whereas audio signals in realistic videos are disregarded. In this work, we concentrate on a rarely investigated problem of text guided sounding video generation and propose the Sounding Video Generator (SVG), a unified framework for generating realistic videos along with audio signals. Specifically, we present the SVG-VQGAN to transform visual frames and audio melspectrograms into discrete tokens. SVG-VQGAN applies a novel hybrid contrastive learning method to model inter-modal and intra-modal consistency and improve the quantized representations. A cross-modal attention module is employed to extract associated features of visual frames and audio signals for contrastive learning. Then, a Transformer-based decoder is used to model associations between texts, visual frames, and audio signals at token level for auto-regressive sounding video generation. AudioSetCap, a human annotated text-video-audio paired dataset, is produced for training SVG. Experimental results demonstrate the superiority of our method when compared with existing textto-video generation methods as well as audio generation methods on Kinetics and VAS datasets.

cs.CV↗

arXiv2023

Dynamic Focus-aware Positional Queries for Semantic Segmentation

The DETR-like segmentors have underpinned the most recent breakthroughs in semantic segmentation, which end-to-end train a set of queries representing the class prototypes or target segments. Recently, masked attention is proposed to restrict each query to only attend to the foreground regions predicted by the preceding decoder block for easier optimization. Although promising, it relies on the learnable parameterized positional queries which tend to encode the dataset statistics, leading to inaccurate localization for distinct individual queries. In this paper, we propose a simple yet effective query design for semantic segmentation termed Dynamic Focus-aware Positional Queries (DFPQ), which dynamically generates positional queries conditioned on the cross-attention scores from the preceding decoder block and the positional encodings for the corresponding image features, simultaneously. Therefore, our DFPQ preserves rich localization information for the target segments and provides accurate and fine-grained positional priors. In addition, we propose to efficiently deal with high-resolution cross-attention by only aggregating the contextual tokens based on the low-resolution cross-attention scores to perform local relation aggregation. Extensive experiments on ADE20K and Cityscapes show that with the two modifications on Mask2former, our framework achieves SOTA performance and outperforms Mask2former by clear margins of 1.1%, 1.9%, and 1.1% single-scale mIoU with ResNet-50, Swin-T, and Swin-B backbones on the ADE20K validation set, respectively. Source code is available at https://github.com/ziplab/FASeg

cs.CV↗

arXiv2023

OmniAvatar: Geometry-Guided Controllable 3D Head Synthesis

We present OmniAvatar, a novel geometry-guided 3D head synthesis model trained from in-the-wild unstructured images that is capable of synthesizing diverse identity-preserved 3D heads with compelling dynamic details under full disentangled control over camera poses, facial expressions, head shapes, articulated neck and jaw poses. To achieve such high level of disentangled control, we first explicitly define a novel semantic signed distance function (SDF) around a head geometry (FLAME) conditioned on the control parameters. This semantic SDF allows us to build a differentiable volumetric correspondence map from the observation space to a disentangled canonical space from all the control parameters. We then leverage the 3D-aware GAN framework (EG3D) to synthesize detailed shape and appearance of 3D full heads in the canonical space, followed by a volume rendering step guided by the volumetric correspondence map to output into the observation space. To ensure the control accuracy on the synthesized head shapes and expressions, we introduce a geometry prior loss to conform to head SDF and a control loss to conform to the expression code. Further, we enhance the temporal realism with dynamic details conditioned upon varying expressions and joint poses. Our model can synthesize more preferable identity-preserved 3D heads with compelling dynamic details compared to the state-of-the-art methods both qualitatively and quantitatively. We also provide an ablation study to justify many of our system design choices.

cs.CV↗