Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Suggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works85 from public data
- ProphetNet: Predicting Future N-gram for Sequence-to-SequencePre-training512
A new sequence-to-sequence pre-training model called ProphetNet is presented, which introduces a novel self-supervised objective named future n-gram prediction and the proposed n-stream self-attention mechanism that predicts the next n tokens simultaneously based on previous context tokens at each time step.
- XGLUE: A New Benchmark Dataset for Cross-lingual Pre-training, Understanding and Generation404
A recent cross-lingual pre-trained model Unicoder is extended to cover both understanding and generation tasks, which is evaluated on XGLUE as a strong baseline and the base versions of Multilingual BERT, XLM and XLM-R are evaluated for comparison.
- Learning to Rank by Optimizing NDCG Measure229
This work proposes a probabilistic framework that addresses the challenge of evaluating ranking algorithms using the expectation of NDCG over all the possible permutations of documents, and proposes an algorithm that outperforms state-of-the-art ranking algorithms on several benchmark data sets.
- HittER: Hierarchical Transformers for Knowledge Graph Embeddings155
This paper proposes HittER, a Hierarchical Transformer model to jointly learn Entity-relation composition and Relational contextualization based on a source entity’s neighborhood and designs a masked entity prediction task to balance information from the relational context and the source entity itself.
- 138
- DeepIntent121
This paper proposes a novel attention network, which learns to assign attention scores to different word locations according to their intent importance, and shows that using the learned attention scores, one is able to produce sub-queries that are of better qualities than those of the state-of-the-art methods.
- TwinBERT114
TwinBERT model, a compressed multi-layer transformer encoder is proposed with special training strategies as a substitution of the original complex BERT encoder, which has met the real-time latency requirement and achieve close or on-par performance to BERT-Base model.
- Bid landscape forecasting in online ad exchange marketplace112
A divide-and-conquer approach that breaks down the campaign-level forecasting problem and can accurately forecast the bid distributions for various campaigns running on the world's largest NGD advertising exchange system, outperforming two baseline methods in term of forecasting errors.
- Exploitation and exploration in a performance based contextual advertising system112
This paper develops two novel EE strategies for online advertising that can adaptively balance the two aspects of EE by automatically learning the optimal tradeoff and incorporating confidence metrics of historical performance.
- Mask Attention Networks: Rethinking and Strengthen Transformer83
A novel understanding of SAN and FFN are presented as Mask Attention Networks (MANs) and it is shown that they are two special cases of MANs with static mask matrices, which limit the capability for localness modeling in text representation learning.
- GLGE: A New General Language Generation Evaluation Benchmark75
The General Language Generation Evaluation (GLGE), a new multi-task benchmark for evaluating the generalization capabilities of NLG models across eight language generation tasks, is presented and a leaderboard with strong baselines including MASS, BART, and ProphetNet is built.
- Multimedia features for click prediction of new ads in display advertising73
It is demonstrated that adding multimedia features can significantly improve the accuracy of click prediction for new ads, compared to a state-of-the-art baseline model.
- 67
- 64
- ProphetNet-X: Large-Scale Pre-training Models for English, Chinese, Multi-lingual, Dialog, and Code Generation56
The ProphetNet family pre-training models, named ProphetNet-X, where X can be English, Chinese, Multi-lingual, and so on, are presented, and achieve new state-of-the-art performance on 10 benchmarks.
- 53
- Generating Better Search Engine Text Advertisements with Deep Reinforcement Learning51
This work jointly train a model to minimize cross-entropy on an existing corpus of Landing Page/Text Ad pairs using typical sequence to sequence training techniques while also optimizing the expected click-through rate as predicted by an existing oracle model using SCST.
- GalaXC: Graph Neural Networks with Labelwise Attention for Extreme Classification50
The GalaXC algorithm for Extreme Classification is developed, where the task is to annotate a document with the most relevant subset of labels from an extremely large label set, and was found to be up to 25% more accurate than extreme classification algorithms specifically designed for warm start settings.
- Click-Through Rate Estimation for Rare Events in Online Advertising44
This chapter develops models and methods to smoothen CTR estimation by taking advantage of the data hierarchy in nature or by clustering and data continuity in time to leverage information from data close to the events of interest.
- The impact of visual appearance on user response in online display advertising43
This paper quantitatively study the relationship between the visual appearance and performance of creatives using large scale data in the world's largest display ads exchange system, RightMedia, and presents recommendations of visual features that have the most important impact on CTR to the professional designers in order to optimize their creative design.
- Nanozymes and Their Potential Roles in the Origin of Life42
This study proposes that nanozymes, nanomaterials with enzyme‐like activities, act as efficient prebiotic catalysts in the origin of life and hopes the research on nanozymes and the origin of life will bridge the gap between inorganic precursors and biomolecules under primitive environments.
- Rare Query Expansion Through Generative Adversarial Networks in Search Advertising39
Using GAN to generate bid keywords directly from query in sponsored search ads selection, especially for rare queries, shows big potential to bring extra revenue improvement and improve the effectiveness and efficiency of query-keyword matching based ads selection in search advertising.
- A stochastic learning-to-rank algorithm and its application to contextual advertising39
This paper proposes a simple yet effective stochastic optimization algorithm to directly minimize any loss function, which can be defined on NDCG or MAP for the learning-to-rank problem, and employs Simulated Annealing along with Simplex method for its parameter search.
- Multimedia Data Mining38
This novel, self-contained book examines how the merging of multimedia and data mining research can promote the understanding and advance the development of knowledge discovery in multimedia data.
- Large-Scale Multi-Label Learning with Incomplete Label Assignments36
This paper proposes an approach, called Mpu, based upon positive and unlabeled stochastic gradient descent and stacked models that can effectively and efficiently consider missing labels and label correlations simultaneously, and is very scalable, that has linear time complexities over the size of the data.
- 35
- An Enhanced Knowledge Injection Model for Commonsense Generation32
This work integrates two additional modules into the pretrained encoder-decoder model for prototype modeling to enhance the knowledge injection procedure and shows results that significantly improves the performance on all the metrics.
- BALAS: Empirical Bayesian learning in the relevance feedback for image retrieval32
The experimental evaluations have confirmed the rationale of the proposed ranking scheme, and have demonstrated that BALAS is superior to an existing relevance feedback method in the current literature in capturing the overall retrieval semantics.
- 27
- 26
- Addressing CBIR efficiency, effectiveness, and retrieval subjectivity simultaneously26
Theoretical analysis and experimental evaluations show that FAST methodology holds a great promise in delivering fast and semantics-tailored image retrieval in CBIR.
- Visual appearance of display ads and its effect on click through rate25
This paper quantitatively study the relationship between the visual appearance and performance of creatives using large scale data in the world's largest display ads exchange system, RightMedia, and designs a set of 43 visual features, some of which are novel and others are inspired by related work.
- 24
- 24
- SegFormer: A Topic Segmentation Model with Controllable Range of Attention21
A new text segmentation model SegFormer with unidirectional attention blocks to better model sentence representations and a novel additional context aggregator and a topic classification loss to guide the model to aggregate the information within the appropriate range.
- 19
- FastSeq: Make Sequence Generation Faster18
The proposed optimization techniques include an attention cache optimization, an efficient algorithm for detecting repeated n-grams, and an asynchronous generation pipeline with parallel I/O that are general enough to be applicable to Transformer-based models.
- Joint categorization of queries and clips for web-based video search17
Evaluation of the developed system and user study has indicated that the joint categorization of queries and video data boosts the video search relevance and user search experience.
- Stretching Bayesian Learning in the Relevance Feedback of Image Retrieval17
An integrated ranking scheme in BALAS is developed which complementarily combines the subjective relevancy confidence and the objective feature-based distance measure to capture the overall retrieval semantics.
- ProphetNet-Ads: A Looking Ahead Strategy for Generative Retrieval Models in Sponsored Search Engine15
Proposing a looking ahead strategy for generative retrieval models named ProphetNet-Ads, which improves the retrieval ability by directly optimizing the Trie-constrained searching space and improving the recall by 15.58\% and 18.8\% respectively.
- A Robust Color Object Analysis Approach to Efficient Image Retrieval15
A novel indexing and retrieval methodology integrating color, texture, and shape information for content-based image retrieval in image databases called CLEAR, which applies unsupervised image segmentation to partition an image into a set of objects.
- 13
- Automatic keywords generation for contextual advertising13
Experimental results over a very large real-world data set shows that the proposed approach can outperform the state-of-the-art system in both relevance and monetization metrics.
- Enhancing Self-Attention with Knowledge-Assisted Attention Maps11
This paper proposes a novel and generic solution, KAM-BERT, which directly incorporates knowledge-generated attention maps into the self-attention mechanism and requires only a few extra parameters and supports efficient fine-tuning once new knowledge is added.
- Ads Keyword Rewriting Using Search Engine Results11
This paper proposes an approach to rewrite the rare BKs to more commonly searched keywords, without compromising the original BKS intent, which increases the coverage and depth of PS ads and thus it delivers higher monetization power.
- A Lipschitz Exploration-Exploitation Scheme for Bayesian Optimization11
An algorithm with a distinct exploration phase followed by an exploitation phase that significantly outperforms the baseline algorithms and leverages the Lipschitz property.
- FAST: Toward more effective and efficient image retrieval11
Theoretical analysis and experimental evaluations show that FAST methodology holds great promise in delivering fast and semantics-tailored image retrieval in CBIR.
- 10
- 10
- A data mining approach to modeling relationships among categories in image collection10
The experimental evaluations have demonstrated that the proposed approach models the semantic relationships effectively and the image retrieval prototype system utilizing the derived model is promising both in effectiveness and efficiency.
- 8
- 8
- AdsCVLR: Commercial Visual-Linguistic Representation Modeling in Sponsored Search7
This paper proposes a transformer architecture with Ads data on Commercial Visual-Linguistic Representation (AdsCVLR) with contrastive learning that naturally extends the transformer encoder with the complementary multi-modal inputs, serving as a strong aggregator of image-text features.
- Dynamic ad layout revenue optimization for display advertising7
A dynamic ad layout revenue optimization framework is developed for display advertising, in terms of both guaranteed and non-guaranteed advertising, and two objectives are explored for the forecasting algorithms of ad layout optimization, the expected revenue and actual revenue.
- Maximum Margin Active Learning for Sequence Labeling with Different Length7
An active learning technique to select the most informative subset of unlabeled sequences for annotation by choosing sequences that have largest uncertainty in their prediction is proposed.
- 7
- PAIR: Pre-denosing Augmented Image Retrieval Model for Defending Adversarial Patches6
The Pre-denosing Augmented Image Retrieval (PAIR) model is proposed, a new approach designed to protect image retrieval systems against adversarial patch attacks and a novel training strategy that incorporates a semantic discriminator.
- SwiftPruner: Reinforced Evolutionary Pruning for Efficient Ad Relevance6
This work aims to design a new, low-latency BERT via structured pruning to empower real-time online inference for cold start ads relevance on a CPU platform and proposes SwiftPruner - an efficient framework that leverages evolution-based search to automatically find the best-performing layer-wise sparse BERT model under the desired latency constraint.
- 6
- AutoADR4
This paper proposes AutoADR (Automatic model design for AD Relevance) --- a novel end-to-end framework to address the challenge of pre-trained language models deployment into industrial online systems, and shares the experience to ship these cutting-edge techniques into online Ad Relevantance system at Microsoft Bing.
- Contextual Query Intent Extraction for Paid Search Selection4
Experimental results over a very large real-world data set demonstrate the superb performance of proposed approach in optimizing both relevance and monetization metrics compared with one of the existing successful algorithms in the system.
- Multiview hierarchical bayesian regression model andapplication to online advertising4
A novel hierarchical Bayesian mixture regression model is proposed, which discovers and then exploits the relationships among multiple views of the data to perform various machine learning tasks, and a stochastic EM inference and learning algorithm is derived.
- A Self-Paced Mixed Distillation Method for Non-Autoregressive Generation3
A novel self-paced mixed distillation method based on the AR stream knowledge that improves the generation quality and has no influence on the inference latency of BANG and achieves more than 7x speedup.
- 3
- 3
- 3
- Exploring Intrinsic Alignments Within Text Corpus2
It is demonstrated that ISAAC can provide preliminary alignment to LLMs, thereby reducing the construction costs of existing human-feedback-based methods and improving the safety of these models under specific safety domains.
- MutationGuard: A Graph and Temporal-Spatial Neural Method for Detecting Mutation Telecommunication Fraud2
This research introduces MutationGuard, which leverages Graph Neural Networks (GNN) to capture changes in user communication graphs and temporal-spatial Voice of Call (VOC) features and validate the effectiveness of the method on the publicly available Sichuan Telecommunication Fraud dataset.
- CULG: Commercial Universal Language Generation2
This work proposes CULG, a large-scale commercial universal language generation model which is pre-trained on a corpus drawn from 10 markets across 7 languages, and demonstrates that the proposed strategy yields performance improvements on three generation tasks as compared to single-stage pre-training.
- 2
- 2
- Internet multimedia advertising2
This tutorial aims at bringing together recent insights from the research on multimedia advertising that addresses the theoretical fundamentals, solution concepts, and the issues related to the development of modern multimedia advertising schemes.
- 2
- 1
- Healing Unsafe Dialogue Responses with Weak Supervision Signals1
An unsupervised pseudo-label sampling method, TEMP, that can automatically assign potential safe responses to agents in large-scale dialogue systems and outperforms state-of-the-art models with weak supervision signals and obtains comparable results under unsuper supervised learning settings.
- 1
- 1
- Solving the Small and Asymmetric Sampling Problem in the Context of Image Retrieval1
The experimental evaluations have confirmed the rationale of the proposed ranking scheme, and have demonstrated that BALAS is superior to an existing relevance feedback method in the current literature in capturing the overall retrieval semantics.
- 1
- 1
- Region based /spl alpha/-semantics graph driven image retrieval1
This work is about content based image database retrieval, focusing on developing a classification based methodology to address semantics-intensive image retrieval, and defines a structure, called /spl alpha/-semantics graph, to discover the hidden semantic relationships among the semantic repositories embodied in the image database.
- 1
- Using Region Analysis Technique to Efficiently Retrieve Online Image Databases1
This research investigates the image retrieval problem for online image databases by proposing a novel solution, called CLEAR, that incorporates multi-features including color, texture, shape, as well as conventional geometric information, which allows the features to be more descriptive in indexing the objects of an image.
- –
- –
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 2026-10-15. 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.