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
- Chengqi ZhangSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works68 from public data
- DiSAN: Directional Self-Attention Network for RNN/CNN-Free Language Understanding794
A novel attention mechanism in which the attention between elements from input sequence(s) is directional and multi-dimensional (i.e., feature-wise) and a light-weight neural net is proposed, based solely on the proposed attention without any RNN/CNN structure, which outperforms complicated RNN models on both prediction quality and time efficiency.
- Multi-center federated learning: clients clustering for better personalization399
A novel multi-center aggregation mechanism to cluster clients using their models’ parameters is proposed that learns multiple global models from data as the cluster centers, and simultaneously derives the optimal matching between users and centers and outperforms several popular baseline methods.
- Federated unsupervised representation learning174
The federated contrastive averaging with dictionary and alignment (FedCA) algorithm, composed of two key modules: a dictionary module to aggregate the representations of samples from each client which can be shared with all clients for consistency of representation space, is adopted for local model training.
- Reinforced Self-Attention Network: a Hybrid of Hard and Soft Attention for Sequence Modeling148
An RNN/CNN-free sentence-encoding model, "reinforced self-attention network (ReSAN)", solely based on ReSA is proposed, which achieves state-of-the-art performance on both the Stanford Natural Language Inference and the Sentences Involving Compositional Knowledge datasets.
- Thread of Thought Unraveling Chaotic Contexts93
This work introduces the "Thread of Thought" strategy, which draws inspiration from human cognitive processes and systematically segments and analyzes extended contexts while adeptly selecting pertinent information within the chaotic context.
- Self-Attention Enhanced Selective Gate with Entity-Aware Embedding for Distantly Supervised Relation Extraction90
A brand-new light-weight neural framework to address the distantly supervised relation extraction problem and alleviate the defects in previous selective attention framework is proposed, which achieves a new state-of-the-art performance in terms of both AUC and top-n precision metrics.
- Multi-Task Learning for Conversational Question Answering over a Large-Scale Knowledge Base85
This work proposes an innovative multi-task learning framework where a pointer-equipped semantic parsing model is designed to resolve coreference in conversations, and naturally empower joint learning with a novel type-aware entity detection model.
- Federated Learning for Privacy-Preserving Open Innovation Future on Digital Health78
This chapter will discuss how federated learning can enable the development of an open health ecosystem with the support of AI.
- Re-Reading Improves Reasoning in Large Language Models74
RE2 demonstrates strong generality and compatibility with most thought-eliciting prompting methods, including CoT, and its adaptability is revealed, showing how it can be effectively integrated with different LLMs, thought-eliciting prompting, and ensemble strategies.
- 74
- Exploiting Structured Knowledge in Text via Graph-Guided Representation Learning74
This work proposes an entity masking scheme that exploits relational knowledge underlying the text, and uses knowledge graphs to obtain distractors for the masked entities, and proposes a novel distractor-suppressed ranking objective which is optimized jointly with masked language model.
- 70
- EventBERT: A Pre-Trained Model for Event Correlation Reasoning69
EventBERT is proposed, a pre-trained model to encapsulate eventuality knowledge from unlabeled text and outperforms existing pre- trained models by a large margin, e.g., 6.5 ∼ 23%, in zero-shot learning of these tasks.
- 66
- Modeling Event-Pair Relations in External Knowledge Graphs for Script Reasoning57
A knowledge model to learn the inferential relations between events from the whole eventuality KG and a neural script adapter to extend the knowledge model for inferring the associated relations be-tween an event chain and a subsequent event candidate is presented.
- Large Language Models are Strong Zero-Shot Retriever54
The Language language model as Retriever (LameR), is built upon no other neural models but an LLM, while breaking brute-force combinations of retrievers with LLMs and lifting the performance of zero-shot retrieval to be very competitive on benchmark datasets.
- BianCang: A Traditional Chinese Medicine Large Language Model50
BianCang is proposed, a TCM-specific LLM, using a two-stage training process that first injects domainspecific knowledge and then aligns it through targeted stimulation to enhance diagnostic and differentiation capabilities.
- PCL: Peer-Contrastive Learning with Diverse Augmentations for Unsupervised Sentence Embeddings50
A novel Peer-Contrastive Learning (PCL) with diverse augmentations is proposed, which performs peer-positive contrast as well as peer-network cooperation, which offers an inherent anti-bias ability and an effective way to learn from diverse augmentation.
- LexMAE: Lexicon-Bottlenecked Pretraining for Large-Scale Retrieval48
A brand-new pre-training framework, lexicon-bottlenecked masked autoencoder (LexMAE), to learn importance-aware lexicon representations, which shows state-of-the-art zero-shot transfer capability on BEIR benchmark with 12 datasets.
- 42
- UnifieR: A Unified Retriever for Large-Scale Retrieval37
This work proposes a new learning framework, Unifier, which unifies dense-vector and lexicon-based retrieval in one model with a dual-representing capability, and experiments on passage retrieval benchmarks verify its effectiveness in both paradigms.
- 37
- LexLIP: Lexicon-Bottlenecked Language-Image Pre-Training for Large-Scale Image-Text Sparse Retrieval36
- Sequential Diagnosis Prediction with Transformer and Ontological Representation36
An end-to-end robust transformer-based model called SETOR is proposed, which exploits neural ordinary differential equation to handle both irregular intervals between a patient’s visits with admitted timestamps and length of stay in each visit, and to alleviate the limitation of insufficient data by integrating medical ontology.
- Temporal Self-Attention Network for Medical Concept Embedding35
This work proposes a novel attention mechanism which captures the contextual information and temporal relationships between medical concepts and is the first to exploit temporal self-attentive relations between medical events.
- 32
- MS MARCO Web Search: A Large-scale Information-rich Web Dataset with Millions of Real Click Labels31
This paper introduces MS MARCO Web Search, the first large-scale information-rich web dataset, featuring millions of real clicked query-document labels, which paves the way for future advancements in AI and system research.
- Eliminating Sentiment Bias for Aspect-Level Sentiment Classification with Unsupervised Opinion Extraction30
A span- based anti-bias aspect representation learning framework that eliminates the sentiment bias in the aspect embedding by adversarial learning against aspects' prior sentiment and aligns the distilled opinion candidates with the aspect by span-based dependency modeling to highlight the interpretable opinion terms.
- LED: Lexicon-Enlightened Dense Retriever for Large-Scale Retrieval29
The proposed dense retriever can bring consistent and significant improvements, and even outdo its teacher, and its lexicon-aware distillation strategies are compatible with the standard ranker distillation, which can further lift state-of-the-art performance.
- Domain-specific meta-embedding with latent semantic structures28
Experiments show that the proposed unsupervised meta-embedding method outputs more accurate word meta- embeddings for the task domain, compared to other state-of-the-art competitors.
- Self-attention Enhanced Patient Journey Understanding in Healthcare System27
A novel self-attention mechanism that can simultaneously capture the contextual and temporal relationships hidden in patient journeys is proposed, specifically designed to learn the representations of patient journeys that is used to be a long sequence of activities.
- Improving Long-Tail Relation Extraction with Collaborating Relation-Augmented Attention26
A novel neural network, Collaborating Relation-augmented Attention (CoRA), is proposed to handle both the wrong labeling and long-tail relations and improves the prior state-of-the-art performance by a large margin in terms of Precision@N, AUC and Hits@K.
- SK2: Integrating Implicit Sentiment Knowledge and Explicit Syntax Knowledge for Aspect-Based Sentiment Analysis21
A brand-new unified framework for ABSA is proposed in this work, which incorporates both implicit sentiment knowledge and explicit syntax knowledge to better complete all ABSA tasks.
- BiteNet: Bidirectional Temporal Encoder Network to Predict Medical Outcomes18
A novel self-attention mechanism that captures the contextual dependency and temporal relationships within a patient's healthcare journey is proposed and an end-to-end bidirectional temporal encoder network (BiteNet) then learns representations of the patient's journeys, based solely on the proposed attention mechanism.
- Tensorized Self-Attention: Efficiently Modeling Pairwise and Global Dependencies Together17
A novel attention mechanism called “Multi-mask Tensorized Self-Attention” (MTSA), which is as fast and as memory-efficient as a CNN, but significantly outperforms previous CNN-/RNN-/attention-based models.
- 16
- 14
- Good Questions Help Zero-Shot Image Reasoning12
Question-Driven Visual Exploration (QVix) is introduced, a novel prompting strategy that enhances the exploratory capabilities of LVLMs in zero-shot reasoning tasks and significantly outperforms existing methods, highlighting its effectiveness in bridging the gap between complex visual data and LVL Ms' exploratory abilities.
- RatE: Relation-Adaptive Translating Embedding for Knowledge Graph Completion11
A relation-adaptive translation function built upon a novel weighted product in complex space, where the weights are learnable, relation-specific and independent to embedding size, which improves expressive power and alleviates embedding ambiguity problem.
- 10
- Rethinking Task-Specific Knowledge Distillation: Contextualized Corpus as Better Textbook9
Surprisingly, contextualized corpus enables task-specific distillation from scratch without general distillation while maintaining comparable performance, making it more flexible to customize the student model with desired model size under various computation constraints.
- Effective Search of Logical Forms for Weakly Supervised Knowledge-Based Question Answering9
This work proposes an effective search method for weakly supervised KBQA based on operator prediction for questions with search space constrained by predicted operators, which means a larger proportion of questions in a weakly supervision training set are equipped with logical forms, and fewer spurious logical forms are generated.
- Synergistic Interplay between Search and Large Language Models for Information Retrieval7
InteR is proposed, a novel framework that facilitates knowledge refinement through interaction between SEs and LLMs, which allows SEs to expand knowledge in queries using LLM-generated knowledge collections and enables LLMs to enhance prompt formulation using SE-retrieved documents.
- MIPO: Mutual Integration of Patient Journey and Medical Ontology for Healthcare Representation Learning6
MIPO (Mutual Integration of Patient Journey and Medical Ontology) is proposed, a robust end-to-end framework that employs a Transformer-based architecture for representation learning and consistently outperforms baseline methods under both sufficient and limited data conditions.
- RetriEVAL: Evaluating Text Generation with Contextualized Lexical Match6
The proposed RetriEVAL is an unsupervised automatic evaluation metric inspired by the concept of lexical match in information retrieval that leverages contextualized lexical matching to measure exact matches between identical tokens and dynamically matches different tokens based on their contextualized representations.
- Advancing Chinese Conversation-based Patient Guidance with a Benchmark and Knowledge-Evolvable Assistant6
The Knowledge-Evolvable Assistant (KEA) is introduced, a novel framework that augments any LLM with an experience bank of validated, successful CCPG cases for analogy-based reasoning, and a reflection bank that records previously misclassified cases together with their corrections and self-summarized error analyses.
- FAA: Fine-grained Attention Alignment for Cascade Document Ranking5
This work utilizes the attention activations over the passages from the ranker as fine-grained attention feedback to optimize the selector and fuse the relevance scores from the passage selector into theRanker to assist in calculating the cooperative matching representation.
- CORE: Cooperative Training of Retriever-Reranker for Effective Dialogue Response Selection5
This work presents a cooperative training of the response retriever and the reranker whose parameters are dynamically optimized by the ground-truth labels as well as list-wise supervision signals from each other, so that the two modules can learn from eachother and evolve together throughout the training.
- 5
- KnowDA: All-in-One Knowledge Mixture Model for Data Augmentation in Low-Resource NLP5
It is demonstrated that the proposed Knowledge Mixture enables pre-trained language models the capability of generating proper synthetic instances from scratch for complicated tasks (i.e., the data sample has long sequences or multiple sentences).
- Semantic Triple Encoder for Fast Open-Set Link Prediction.5
A novel context-based encoder that separately maps each part and its context into a latent semantic space and covers the contextualized semantics of the triples in the encoder so it can handle unseen entities during inference, which promisingly improves the generalization ability.
- UniEvent: Unified Generative Model with Multi-Dimensional Prefix for Zero-Shot Event-Relational Reasoning4
This work proposes a novel unified framework, called UNIEVENT, that organizes event relational reasoning tasks into a coordinate system with multiple axes, representing inter-event relations and reasoning formulations and achieves state-of-the-art or competitive performance on both zero-shot and supervised reasoning tasks.
- 3
- 3
- CCPrefix: Counterfactual Contrastive Prefix-Tuning for Many-Class Classification3
This work proposes a brand-new prefix-tuning method, Counterfactual Contrastive Prefix-tuning (CCPrefix), where an instance-dependent soft prefix, derived from fact-counterfactual pairs in the label space, is leveraged to complement the language verbalizers in many-class classification.
- LexLIP: Lexicon-Bottlenecked Language-Image Pre-Training for Large-Scale Image-Text Retrieval3
This work proposes the lexicon-weighting paradigm, where sparse representations in vocabulary space are learned for images and texts to take advantage of the bag-of-words models and efficient inverted indexes, resulting in significantly reduced retrieval latency.
- 3
- 3
- 2
- 2
- 2
- 2
- Hierarchical Relation-Guided Type-Sentence Alignment for Long-Tail Relation Extraction with Distant Supervision1
This work proposes a novel model to enrich distantly-supervised sentences with entity types that achieves new state-of-the-art results in overall and long-tail performance on benchmarks.
- MedShort: Medicine Shortage Detection on Social Media with Privacy Preservation–
A novel framework that leverages large language models to monitor, analyze, and detect Medicine Shortages from social media data while preserving user privacy is introduced, enabling near real-time and scalable monitoring of medicine shortages.
- Dimension-Prompts Boost Commonsense Consolidation–
A brand-new dimension-disentangled knowledge model (D2KM) learning paradigm with multiple sources is proposed that is, a generative language model with dimension-specific soft prompts is trained to disentangle knowledge acquisitions along with different dimensions and facilitate potential intra-dimension consolidation across CS-KG sources.
- Improving Open-Domain Answer Sentence Selection by Distributed Clients with Privacy Preservation–
A brand-015 new learning framework for P rivacy-preserving O D-AS2, dubbed PDD-AS2, consists of a client-018 customized query encoding for better personal-019 ization and a cross-client negative sampling for 020 learning effectiveness.
- –
- –
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.