Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks140 from public data
- Scaling Distributed Machine Learning with the Parameter Server979
View on new challenges identified are shared, and some of the application scenarios such as micro-blog data analysis and data processing in building next generation search engines are covered.
- Shift-Net: Image Inpainting via Deep Feature Rearrangement511
A special shift-connection layer to the U-Net architecture, namely Shift-Net, is introduced for filling in missing regions of any shape with sharp structures and fine-detailed textures and an end-to-end learning algorithm is further developed to train the Shift- net.
- Learning Convolutional Networks for Content-Weighted Image Compression421
The bit rate of the different parts of the image is adapted to local content, and the content-aware bit rate is allocated under the guidance of a content-weighted importance map so that the sum of the importance map can serve as a continuous alternative of discrete entropy estimation to control compression rate.
- Earthformer: Exploring Space-Time Transformers for Earth System Forecasting416
The idea is to decompose the data into cuboids and apply cuboid-level self-attention in parallel, which achieves state-of-the-art performance on real-world benchmarks about precipitation nowcasting and El Nino/Southern Oscillation forecasting.
- AIM: Adapting Image Models for Efficient Video Action Recognition273
This work proposes a novel method to Adapt pre-trained Image Models (AIM) for efficient video understanding, and shows that the proposed AIM can achieve competitive or even better performance than prior arts with substantially fewer tunable parameters on four video action recognition benchmarks.
- Hierarchical Recurrent Neural Network for Document Modeling170
A novel hierarchical recurrent neural network language model (HRNNLM) for document modeling that integrates it as the sentence history information into the word level RNN to predict the word sequence with cross-sentence contextual information.
- A Recursive Recurrent Neural Network for Statistical Machine Translation166
A novel recursive recurrent neural network (R 2 NN) is proposed to model the end-to-end decoding process for statistical machine translation and can outperform the state of theart baseline by about 1.5 points in BLEU.
- Exploring distributional similarity based models for query spelling correction143
Novel methods for use of distributional similarity estimated from query logs in learning improved query spelling correction models are described, which can significantly outperform their baseline systems in the web query spelling Correction task.
- MixGen: A New Multi-Modal Data Augmentation136
MixGen is presented: a joint data augmentation for vision-language representation learning to further improve data efficiency and generates new image-text pairs with semantic relationships preserved by interpolating images and concatenating text.
- Iterative Row Sampling127
This work shows that alternating between computing a short matrix estimate and finding more accurate approximate leverage scores leads to a series of geometrically smaller instances that gives an algorithm whose runtime is input sparsity plus an overhead comparable to the cost of solving a regression problem on the smaller approximation.
- 125
- Effective Tag Set Selection in Chinese Word Segmentation via Conditional Random Field Modeling122
It is shown that there is a significant performance difference as different tag sets are selected, and the proposed method gives the state-of-the-art performance.
- Bilingually-constrained Phrase Embeddings for Machine Translation120
This work proposes Bilingually-constrained Recursive Auto-encoders (BRAE) to learn semantic phrase embeddings (compact vector representations for phrases), which can distinguish the phrases with different semantic meanings.
- 117
- Sequence-to-Dependency Neural Machine Translation107
A novel Sequence-to-Dependency Neural Machine Translation (SD-NMT) method, in which the target word sequence and its corresponding dependency structure are jointly constructed and modeled, and this structure is used as context to facilitate word generations.
- 105
- LayoutDiffuse: Adapting Foundational Diffusion Models for Layout-to-Image Generation98
This paper proposes LayoutDiffuse, a novel neural adaptor that adapts a foundational diffusion model pretrained on large-scale image or text-image datasets for layout-to-image generation that significantly outperforms other 10 generative models based on GANs, VQ-VAE, and diffusion models.
- Learning Content-Weighted Deep Image Compression95
An upper-triangular masked convolutional network (triuMCN) for large context modeling is proposed, where the channel-wise multi-valued quantization is deployed for the discretization of the encoder features, and an importance map subnet is introduced to generate the importance masks for spatially varying code pruning.
- A Probabilistic Approach to Syntax-based Reordering for Statistical Machine Translation92
A novel, probabilistic approach to reordering which combines the merits of syntax and phrase-based SMT is proposed, which leads to BLEU improvement of 1.56% for the NIST MT-05 task of Chinese-toEnglish translation.
- CrossNorm and SelfNorm for Generalization under Distribution Shifts75
CrossNorm exchanges channel-wise mean and variance between feature maps to enlarge training distribution, while SelfNorm uses attention to recalibrate the statistics to bridge gaps between training and test distributions.
- Video Contrastive Learning with Global Context70
This paper proposes a new video-level contrastive learning method based on segments to formulate positive pairs that is able to capture the global context in a video, thus robust to temporal content change and can outperform previous state-of-the-arts efforts.
- Knowledge-Based Semantic Embedding for Machine Translation67
This paper builds and formulate a semantic space to connect the source and target languages, and applies it to the sequence-to-sequence framework to pro-pose a Knowledge-Based Semantic Em-bedding (KBSE) method.
- Removing Batch Normalization Boosts Adversarial Training53
This work proposes to completely eliminate this bottleneck by removing all BN layers in AT and achieves adversarial robustness with only a minor sacrifice on clean sample accuracy, and observes better model smoothness and larger decision margins from NoFrost, which make the models less sensitive to input perturbations and thus more robust.
- Modular Blind Video Quality Assessment50
A modular BVQA model that comprises a base quality predictor, a spatial rectifier, and a temporal rectifier, responding to the visual content and distortion, spatial resolution, and frame rate changes on video quality, respectively is proposed.
- 48
- 43
- 41
- 39
- 39
- Targeted Chromosomal Translocations and Essential Gene Knockout Using CRISPR/Cas9 Technology in Caenorhabditis elegans39
The method provides an easy and efficient approach to generate and maintain loss-of-function alleles of essential genes with detailed genetic background information and is applied to edit essential genes in Caenorhabditis elegans.
- 36
- 35
- Phrase Reordering Model Integrating Syntactic Knowledge for SMT35
Syntactic knowledge is introduced to improve global reordering capability of SMT system and constraints in syntax tree are proposed to avoid the reordering errors, but also the modification of syntax tree is made to strengthen the capability of capturing phrase reordering.
- 34
- Learning Multimodal Data Augmentation in Feature Space33
LeMDA, Learning Multimodal Data Augmentation, an easy-to-use method that automatically learns to jointly augment multimodal data in feature space, with no constraints on the identities of the modalities or the relationship between modalities is introduced.
- 32
- A Probabilistic Hierarchical Model for Multi-View and Multi-Feature Classification30
A probabilistic hierarchical model is proposed to address the issue of hierarchical structure in classification and applied for classification, and the Expectation-Maximization algorithm is applied to estimate the parameters and latent variables.
- 29
- Efficient Collective Entity Linking with Stacking29
A fast collective disambiguation approach based on stacking that learns a rich semantic relatedness measure between entity categories and context document, which shows effectiveness over various algorithms on several public datasets.
- Bilingual Data Cleaning for SMT using Graph-based Random Walk29
The method leverages the mutual reinforcement between the sentence pairs and the extracted phrase pairs, based on the observation that better sentence pairs often lead to better phrase extraction and vice versa and substantially improves the performance in largescale Chinese-to-English translation tasks.
- Tencent Neural Machine Translation Systems for the WMT20 News Translation Task28
This paper describes Tencent Neural Machine Translation systems for the WMT 2020 news translation tasks, built on deep Transformer and several data augmentation methods, and proposes a boosted in-domain finetuning method to improve single models.
- 24
- Adaptive Development Data Selection for Log-linear Model in Statistical Machine Translation24
Experimental results show that the proposed principled method can effectively adapt log-linear model parameters to different test data, and consistently achieves good translation performance compared with conventional methods that use a fixed model parameter setting across different data sets.
- 22
- 21
- 21
- Similarity and diversity induced paired projection for cross-modal retrieval20
A similarity and diversity induced paired projection method which not only extracts the correlation in a common subspace, but also removes the view-specific information which does not contribute to the authors' task.
- Twofold Video Hashing With Automatic Synchronization19
Experiments on real video collections show that such a hash extraction and fusion method enables unprecedented robustness under both spatial and temporal attacks.
- 17
- 17
- Benchmarking Robustness of Multimodal Image-Text Models under Distribution Shift17
It is observed that multimodal models are not robust to image and text perturbations, especially to image perturbations, so several new multimodal robustness benchmarks are proposed by applying 17 image perturbation and 16 text perturbation techniques on top of existing datasets.
- Chunk-based Decoder for Neural Machine Translation17
Experimental results show that the proposed chunk-based decoders can significantly improve translation performance in a WAT ‘16 English-to-Japanese translation task.
- Learning Topic Representation for SMT with Neural Networks17
Experimental results show that the proposed novel approach to learning topic representation for parallel data using a neural network architecture significantly improves translation accuracy in the NIST Chinese-to-English translation task compared to a state-of-the-art baseline.
- 16
- Perceptual Quality Assessment of Virtual Reality Videos in the Wild16
The VR Video Quality in the Wild (VRVQW) database is constructed, containing 502 user-generated videos with diverse content and distortion characteristics, and an objective quality assessment model for VR videos based on pseudocylindrical representation and convolution is developed.
- 16
- Secure outsourcing of large matrix determinant computation16
This study proposes a secure outsourcing method for large matrix determinant computation using permutation and mix-row/mix-column operations and achieves a higher security level with greater cloud efficiency.
- Multi-Domain Adaptation for SMT Using Multi-Task Learning15
This paper proposes a novel multi-domain adaptation approach for SMT using Multi-Task Learning (MTL), with in-domain models tailored for each specific domain and a general-domain model shared by different domains.
- A Joint Rule Selection Model for Hierarchical Phrase-Based Translation15
A joint model to predict the selection of hierarchical rules based on four sub-models where the rich context knowledge from both source and target sides is leveraged and can yield significant improvements in performance.
- Mixture Model-based Minimum Bayes Risk Decoding using Multiple Machine Translation Systems15
Experimental results on the NIST Chinese-to-English MT evaluation tasks show that the MMMBR approach brings significant improvements to single system-based MBR decoding and outperforms a state-of-the-art system combination method.
- 14
- 14
- Pseudocylindrical Convolutions for Learned Omnidirectional Image Compression13
This work proposes parametric pseudocylindrical representations that generalize common pseudocylindrical map projections and develops pseudocylindrical convolutions, which can be efficiently implemented by standard convolutions with “pseudocylindrical padding".
- 13
- 13
- 13
- Learning Translation Consensus with Structured Label Propagation13
This work forms the problem as structured labeling over a much smaller graph, and proposes a novel structured label propagation for the task, and converts such graph-based translation consensus from similar source strings into useful features both for n-best output re-ranking and for decoding algorithm.
- 12
- Spherical Pseudo-Cylindrical Representation for Omnidirectional Image Super-resolution12
This paper makes one of the first attempts to design a spherical pseudo-cylindrical representation, which not only allows pixels at different latitudes to adaptively adopt the best distinct sampling density but also is model-agnostic to most off-the-shelf SR methods, enhancing their performances.
- ShiftLIC: Lightweight Learned Image Compression With Spatial-Channel Shift Operations11
Experimental results demonstrate that ShiftLIC outperforms leading compression methods, such as VVC Intra and GMM, in terms of computational cost, parameter count, and decoding latency, showcasing its potential for practical deployment in resource-constrained environments.
- Recurrent Attention for Neural Machine Translation11
Recurrent AtteNtion (RAN), a novel substitute mechanism for self-attention, is proposed and found that RAN models are competitive and outperform their Transformer counterpart in certain scenarios, with fewer parameters and inference time.
- Online Multi-view Subspace Learning with Mixed Noise11
This paper proposes a novel online multi-view subspace learning method that not only estimates a transformation for each view to extract the correlation among various views, but also introduces a Mixture of Gausssians (MoG) model into the multi- view data, successfully exploiting numbers of Gaussian Distributions to adaptively fit a wider range of the complex noise.
- Mind the Gap: Machine Translation by Minimizing the Semantic Gap in Embedding Space11
A Recursive Neural Network (RNN) based model is proposed that converts each translation rule into a compact real-valued vector in the semantic embedding space and performs the decoding process by minimizing the semantic gap between the source language string and its translation candidates at each state in a bottom-up structure.
- Better synchronous binarization for machine translation11
An algorithm is presented that iteratively improves the resulting binary SCFG and can improve a string-to-tree statistical machine translations system based on the synchronous binarization method in Zhang et al. (2006) on the NIST machine translation evaluation tasks.
- Single character Chinese named entity recognition11
This paper formulates the SCNE recognition within the source-channel model framework, constructing two classifiers based on maximum entropy model (ME) and vector space model (VSM), respectively, and comparing all proposed approaches, showing that the source -channel model performs the best in most cases.
- Deep dual incomplete multi-view multi-label classification via label semantic-guided contrastive learning10
A label semantic-guided contrastive learning method named LSGC for the dual incomplete multi-view multi-label classification problem, which employs deep neural networks to extract high-level features of samples and introduces a graph convolutional network to effectively capture label semantics.
- Emotion Classification by Jointly Learning to Lexiconize and Classify10
This paper proposes an emotional network (EmNet) to jointly learn sentence emotions and construct emotion lexicons which are dynamically adapted to a given context and achieves new state-of-the-art on the benchmark Twitter dataset.
- Extracting Keyphrases from Chinese News Articles Using TextRank and Query Log Knowledge10
This paper proposes a practical approach to extracting keyphrases for Chinese news articles using the TextRank and query log knowl- edge, and generates phrases by employing several statistical criteria with the huge amount of queries as a training corpus.
- 9
- Translation Model Size Reduction for Hierarchical Phrase-based Statistical Machine Translation9
Experimental results on Chinese-to-English machine translation tasks show that the proposed method is able to reduce almost the half size of the translation model with very tiny degradation of translation performance.
- Measure Word Generation for English-Chinese SMT Systems9
This paper proposes a statistical model to generate appropriate measure words of nouns for an English-to-Chinese SMT system, and model the probability of measure word generation by utilizing lexical and syntactic knowledge from both source and target sentences.
- Omnidirectional image super-resolution via position attention network8
A novel position attention network (PAN) is proposed for ODISR that exploits a long-term memory module (LM), promoting information interactions and fusions between the branches to enhance the perception of the distortion, aggregating the prior hierarchical features to keep the long-term memory and boosting the ODISR performance.
- Joint distance and similarity measure learning based on triplet-based constraints8
A novel model is proposed to learn combined distance and similarity measure (CDSM) based on triplet-based constraints with a unified regularizer of the Frobenius norm and two effective strategies are adopted to speed up training and testing of kernelized CDSM.
- 8
- Scanpath Prediction in Panoramic Videos via Expected Code Length Minimization6
This work presents a simple criterion for scanpath prediction based on principles from lossy data compression, and introduces a proportional–integral–derivative (PID) controller-based sampler to generate realistic human-like scanpaths from the learned probability model.
- Learned Scanpaths Aid Blind Panoramic Video Quality Assessment6
This work proposes an end-to-end optimized, blind PVQA method with explicit modeling of user viewing patterns through visual scanpaths, and enables direct quality assessment of panoramic images by treating them as videos composed of identical frames.
- A novel multi-feature learning model for disease diagnosis using face skin images6
This study represents the first to demonstrate concept of multi-feature learning using only facial skin images as an effective non-invasive approach for simultaneously identifying DM, FL and CRF in Han Chinese, the largest ethnic group in the world.
- Efficient Trimmed Convolutional Arithmetic Encoding for Lossless Image Compression.6
This work proposes a trimmed convolutional network for arithmetic encoding (TCAE), which attains better compression ratio in lossless gray image compression, and can be adopted in CNN-based lossy image compression to achieve state-of-the-art rate-distortion performance with real-time encoding speed.
- A Lexicalized Reordering Model for Hierarchical Phrase-based Translation6
This paper proposes a novel lexicalized reordering model which is built directly on synchronous rules for hierarchical phrase-based translation and achieves good results on NIST machine translation test sets.
- Analogue Implementation of Wavelet Transform Using Discrete Time Switched-Current Filters6
A novel scheme for the analogue implementation of wavelet transform (WT) using discrete time switched-current filters is presented and an excellent approximation of the mother wavelet is given by differential evolution (DE) algorithm.
- The feature subspace method for SMT system combination6
A simple and effective method to systematically derive an ensemble of SMT systems from one baseline linear SMT model for use in system combination, and experimental results show that the method can bring significant improvements to baseline systems with state-of-the-art performance.
- Information Retrieval for short documents6
For short document retrieval, RDM achieves significant improvements over the classical probabilistic models on the task of ad hoc retrieval on Text REtrieval Conference (TREC) test sets and results show that the shorter the document, the better the RDM performance.
- 5
- 5
- 5
- Reliability-aware automatic composition approach for web services5
This paper gives an exponential-time algorithm (CSM-NSA) which for a given formula ψ and a synthesis model, synthesizes available services in the library satisfying ψ over the synthesis model or responds with “not satisfiable” (otherwise).
- Hierarchical Chunk-to-String Translation5
A hierarchical chunk-to-string translation model, which can be seen as a compromise between the hierarchical phrase-based model and the tree- to-string model, to combine the merits of the two models.
- Hybrid Decoding: Decoding with Partial Hypotheses Combination over Multiple SMT Systems5
Experimental results on NIST evaluation data sets for Chinese-to-English machine translation (MT) task show that the hybrid decoding method can not only achieve significant improvements over individual decoders, but also bring substantial gains compared with a state-of-the-art word-level system combination method.
- A Comparative Study of Bing Web N-gram Language Models for Web Search and Natural Language Processing5
The significant improvements on search query spelling correction and search query reformulation, resulting from MWNLM, demonstrate the benefit of training multiple language models on different portions of web data and search queries in a principled way with zero count cutoff.
- Introduction to China’s CWMT2008 Machine Translation Evaluation5
This paper presents an overall introduction to the CWMT2008 evaluation and focuses on its two new metrics: BLEU-SBP and linguistic check-point method, and indicates that this method can be used to evaluate the capability of an MT system in translating various linguistic phenomena.
- 4
- 4
- Towards Machine Translation in Semantic Vector Space4
This work presents a recursive neural network (RNN)-based translation framework, which includes two submodels, and shows that the proposed vector representations for the translation rules are very reliable for application in translation modeling.
- Metric Learning with Relative Distance Constraints: A Modified SVM Approach4
A Relative Distance Constrained Metric Learning (RDCML) model is suggested which can be easily implemented and effectively solved by a modified support vector machine (SVM) approach and achieves better or comparable classification accuracy when compared with the state-of-the-art metric learning methods.
- Dataset Distillation as Data Compression: A Rate-Utility Perspective3
This work proposes a joint rate-utility optimization method for dataset distillation that achieves up to $170 \times greater compression than standard distillation at comparable accuracy on CIFAR-10, CIFAR100, and ImageNet-128.
- Efficient Mining of Contrast Patterns on Large Scale Imbalanced Real-Life Data3
A novel algorithm, ConvergMiner, which adopts T*-tree and branch bound pruning strategies to mine converging patterns efficiently, is proposed and greatly outperforms the existing cost-sensitive classification methods in terms of predicative accuracy.
- Hypothesis Mixture Decoding for Statistical Machine Translation3
This paper presents hypothesis mixture decoding (HM decoding), a new decoding scheme that performs translation reconstruction using hypotheses generated by multiple translation systems, and compares the approach with several related techniques, and demonstrates significant BLEU improvements in large-scale Chinese-to-English translation tasks.
- DynaQuant: Dynamic Mixed-Precision Quantization for Learned Image Compression2
DynaQuant is introduced, a novel framework for dynamic mixed-precision quantization that operates on two complementary levels, where learnable scaling and offset parameters dynamically adapt to the statistical variations of latent features.
- Temporal Boundary Analysis on Startup Algorithm for Time-Triggered Architecture with Bus Topology2
The upper boundary of startup time for systems based on TTA with arbitrary number of nodes towards the dedicated startup scheme by formal deduction is given.
- Woodpecker: An Automatic Methodology for Machine Translation Diagnosis with Rich Linguistic Knowledge2
This work proposes an automatic diagnostic evaluation methodology, called Woodpecker, which enables multi-factored evaluation of MT systems based on linguistic categories and automatically constructed linguistic checkpoints, and can automatically monitor a MT system in translating various linguistic phenomena, thereby facilitating diagnostic evaluation.
- 2
- Self-Supervised Voice Denoising Network for Multi-Scenario Human–Robot Interaction1
This work introduces a method to enhance voice command-based HRI in noisy environments, leveraging synthetic data and a self-supervised denoising network to enhance its real-world applicability and enhance voice-guided robot control.
- 1
- Better Addressing Word Deletion for Statistical Machine Translation1
This paper proposes four effective models to handle undesired word deletion in statistical machine translation and develops an automatic evaluation metric that highly correlates with human judgement to evaluate word deletion problems.
- Learning Heterogeneous Coupling Relationships Between Non-IID Terms1
A novel term representation is introduced by involving the coupled relations from term to term that provides much richer information that enables us to create a coupled similarity metric for measuring document similarity, and a coupled document similarity based K-Nearest centroid classifier will be applied to the classification task.
- 1
- 1
- Collective Corpus Weighting and Phrase Scoring for SMT Using Graph-Based Random Walk1
An intuitive approach to improve translation modeling by collective corpus weighting and phrase scoring using the mutual reinforcement between the sentence pairs and the extracted phrase pairs, based on the observation that better sentence pairs often lead to better phrase extraction and vice versa.
- 1
- Function Word Generation in Statistical Machine Translation Systems1
A novel method is designed to separate the generation of target function words from target content words in SMT decoding, which improves the SMT performance significantly on ChineseEnglish translation task.
- 1
- –
- –
- –
- –
- SFNIC: Hybrid Spatial‐Frequency Information for Lightweight Neural Image Compression–
This work introduces a lightweight neural image compression framework designed to efficiently process both local and global information, and achieves a superior balance between rate‐distortion performance and complexity.
- –
- Character-interested binary-like image learning for text image demoiréing–
A novel network for text image demoiréing, where instead of restoring all pixels like those in natural images, a character attention module is conducted, allowing the network to pay more attention on the optical character-associated pixels and also achieving a consistent image style.
- –
- –
- –
- –
- Entity Translation with Collective Inference in Knowledge Graph–
This paper proposes a novel method to construct or enrich a knowledge base by entity translation with help of another KB but compiled in a different language, and shows the method can achieve very high translation precision compared with classical translation methods.
- A Coupled Similarity Kernel for Pairwise Support Vector Machine–
This work designs a novel coupled similarity metric between nominally attributed data and proposes an adapted SVM which can handle this and shows the proposed method outperforms the traditional SVM and other popular classification methods on various public data sets.
- Machine Translation with Real-Time Web Search–
Experimental results show that the web-based machine translation method demonstrates very promising performance in leveraging fresh translation knowledge and making translation decisions and significantly outperforms a state-of-the-art phrase-based statistical machine translation system.
- Design and implementation of Wavelet Transform using Adaptive Genetic Algorithms and Analog filters–
A low power and low-voltage analog switched-current filters implementation of wavelet transform (WT) for real-time requirements in signal processing using an adaptive genetic algorithm to calculate the transfer function of the filters, whose impulse response is the required wavelet base.
- –
- Study on Target Positioning Method Based on Color and Marginal Information Fusion–
The method of threshold segmentation based on color differentiation of the selected area as well as the strategy of target positioning based onColor and edge fusion, which could avoid the jump phenomenon are proposed.
- The MSR-NRC-SRI MT System for NIST Open Machine Translation 2008 Evaluation–
A new method to generate a better alignment between multiple MT hypotheses from different individual systems is developed, which is used to construct a high-quality confusion network.
- Comparing and Integrating Alignment Template and Standard Phrase-Based Statistical Machine Translation–
An improved model to integrate alignment template into standard phrase-based SMT as a new feature in a log-linear model is proposed and Experimental results show that the method outperforms the baseline method.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-14. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.