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
- Alex WaibelSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works252 from public data
- Qwen2.5 Technical Report5,408
The open-weight flagship Qwen2.5-72B-Instruct outperforms a number of open and proprietary models and demonstrates competitive performance to the state-of-the-art open-weight model, Llama-3-405B-Instruct, which is around 5 times larger.
- Qwen Technical Report4,443
Qwen is a comprehensive language model series that encompasses distinct models with varying parameter counts, and includes Qwen, the base pretrained language models, and Qwen-Chat, the chat models finetuned with human alignment techniques.
- Qwen2 Technical Report2,548
A comprehensive suite of foundational and instruction-tuned language models, encompassing a parameter range from 0.5 to 72 billion, featuring dense models and a Mixture-of-Experts model, which surpasses most prior open-weight models and demonstrates robust multilingual capabilities.
- 747
- Knowing What, How and Why: A Near Complete Solution for Aspect-Based Sentiment Analysis531
This paper introduces a new subtask under ABSA, named aspect sentiment triplet extraction (ASTE), and proposes a two-stage framework to address this task, which outperforms a few strong baselines adapted from state-of-the-art related methods.
- KnowPrompt: Knowledge-aware Prompt-tuning with Synergistic Optimization for Relation Extraction518
A Knowledge-aware Prompt-tuning approach with synergistic optimization (KnowPrompt) that injects latent knowledge contained in relation labels into prompt construction with learnable virtual type words and answer words.
- A Joint Neural Model for Information Extraction with Global Features481
A joint neural framework that aims to extract the globally optimal IE result as a graph from an input sentence and can be easily applied to new languages or trained in a multilingual manner, as OneIE does not use any language-specific feature.
- Large Language Models are Versatile Decomposers: Decomposing Evidence and Questions for Table-based Reasoning288
This work exploits large language models (LLMs) as decomposers for effective table-based reasoning, and proposes a novel ''parsing-execution-filling'' strategy to decompose a complex question into simper step-by-step sub-questions by generating intermediate SQL queries as a bridge.
- A Knowledge-Enhanced Pretraining Model for Commonsense Story Generation269
A knowledge-enhanced pretraining model to utilize commonsense knowledge from external knowledge bases to generate reasonable stories that can generate more reasonable stories than state-of-the-art baselines, particularly in terms of logic and global coherence.
- CBLUE: A Chinese Biomedical Language Understanding Evaluation Benchmark262
The first Chinese Biomedical Language Understanding Evaluation Evaluation (CBLUE) benchmark is presented: a collection of natural language understanding tasks including named entity recognition, information extraction, clinical diagnosis normalization, single-sentence/sentence-pair classification, and an associated online platform for model evaluation, comparison, and analysis.
- Hybrid Transformer with Multi-level Fusion for Multimodal Knowledge Graph Completion260
This paper proposes a hybrid transformer architecture with unified input-output for diverse multimodal knowledge graph completion tasks, and proposes multi-level fusion, which integrates visual and text representation via coarse-grained prefix-guided interaction and fine- grained correlation-aware fusion modules.
- Document-level Relation Extraction as Semantic Segmentation236
A Document U-shaped Network for document-level relation extraction is proposed, which leverages an encoder module to capture the context information of entities and a U- shaped segmentation module over the image-style feature map to capture global interdependency among triples.
- Raise a Child in Large Language Model: Towards Effective and Generalizable Fine-tuning235
A straightforward yet effective fine-tuning technique, Child-Tuning, which updates a subset of parameters (called child network) of large pretrained models via strategically masking out the gradients of the non-child network during the backward process is proposed.
- Automated Concatenation of Embeddings for Structured Prediction209
This paper proposes Automated Concatenation of Embeddings (ACE) to automate the process of finding better concatenations of embeddings for structured prediction tasks, based on a formulation inspired by recent progress on neural architecture search.
- GALAXY: A Generative Pre-trained Model for Task-Oriented Dialog with Semi-supervised Learning and Explicit Policy Injection187
GALAXY is a novel pre-trained dialog model that explicitly learns dialog policy from limited labeled dialogs and large-scale unlabeled dialog corpora via semi-supervised learning and has a stronger few-shot ability than existing models under various low-resource settings.
- Improving Named Entity Recognition by External Context Retrieving and Cooperative Learning183
This paper finds empirically that the contextual representations computed on the retrieval-based input view, constructed through the concatenation of a sentence and its external contexts, can achieve significantly improved performance compared to the original input view based only on the sentence.
- Hallucination Augmented Contrastive Learning for Multimodal Large Language Model179
This paper introduces contrastive learning into MLLMs and uses text with hallucination as hard negative examples, naturally bringing representations of non-hallucinative text and visual samples closer while pushing way representations of non-hallucinating and hallucinative text.
- StructuralLM: Structural Pre-training for Form Understanding153
This paper proposes a new pre-training approach, StructuralLM, to jointly leverage cell and layout information from scanned documents, with two new designs to make the most of the interactions of cell and layouts information.
- Contrastive Triple Extraction with Generative Transformer152
This paper introduces a novel model, contrastive triple extraction with a generative transformer module for encoder-decoder-based generation and proposes a novel triplet contrastive training object to generate faithful results.
- Mobile-Agent-E: Self-Evolving Mobile Assistant for Complex Tasks145
This work introduces Mobile-Agent-E, a hierarchical multi-agent framework capable of self-evolution through past experience capable of achieving 22% absolute improvement over previous state-of-the-art approaches across three foundation model backbones.
- FINDINGS OF THE IWSLT 2020 EVALUATION CAMPAIGN141
Each track’s goal, data and evaluation metrics are introduced, and the results of the received submissions are reported.
- Relational Learning with Gated and Attentive Neighbor Aggregator for Few-Shot Knowledge Graph Completion136
A novel gated and attentive neighbor aggregator is built for accurately integrating the semantics of a few-shot relation's neighborhood, which helps filtering the noise neighbors even if a KG contains extremely sparse neighborhoods.
- Boundary Enhanced Neural Span Classification for Nested Named Entity Recognition135
This work proposes a boundary enhanced neural span classification model that has the ability to generate high-quality candidate spans and greatly reduces the time complexity during inference, and incorporates an additional boundary detection task to predict those words that are boundaries of entities.
- 134
- E2E-VLP: End-to-End Vision-Language Pre-training Enhanced by Visual Learning132
This paper proposes the first end-to-end vision-language pre-trained model for both V+L understanding and generation, namely E2E-VLP, where a unified Transformer framework is built to jointly learn visual representation, and semantic alignments between image and text.
- Goodness: A Method for Measuring Machine Translation Confidence112
A novel framework to predict word-level and sentence-level MT errors with a large number of novel features and a visualization prototype of MT errors at the word and sentence levels with the objective to improve post-editor productivity.
- 111
- Improving Biomedical Pretrained Language Models with Knowledge105
KeBioLM is proposed, a biomedical pretrained language model that explicitly leverages knowledge from the UMLS knowledge bases and has better ability to model medical knowledge.
- Distributional representations for handling sparsity in supervised sequence-labeling100
It is demonstrated that distributional representations of word types, trained on unannotated text, can be used to improve performance on rare words and reduces the sample complexity of sequence labeling.
- 87
- PALM: Pre-training an Autoencoding&Autoregressive Language Model for Context-conditioned Generation79
PALM is presented with a novel scheme that jointly pre-trains an autoencoding and autoregressive language model on a large unlabeled corpus, specifically designed for generating new text conditioned on context.
- Mining key phrase translations from web corpora79
A new framework to mine key phrase translations from web corpora is proposed, which achieves 46% phrase translation accuracy when using top 10 returned snippets, and 80% accuracy with 165 snippets, significantly better than several existing methods.
- 74
- 73
- Unsupervised Multi-Modal Neural Machine Translation72
This work investigates the possibility of utilizing images for disambiguation to improve the performance of UNMT, and proposes an unsupervised multi-modal machine translation (UMNMT) framework based on the language translation cycle consistency loss conditional on the image, targeting to learn the bidirectional multi- modal translation simultaneously.
- 71
- Automatic extraction of named entity translingual equivalence based on multi-feature cost minimization71
An integrated approach to extract NE translingual equivalence from a parallel Chinese-English corpus using an NE transliteration model and a greedy search algorithm to minimize the overall multi-feature alignment cost.
- Fusing Heterogeneous Factors with Triaffine Mechanism for Nested Named Entity Recognition70
Experiments show that the proposed triaffine mechanism outperforms previous span-based methods, achieves the state-of-the-art F_1 scores on nested NER datasets GENIA and KBP2017, and shows comparable results on ACE2004 and ACE2005.
- Nested Named Entity Recognition with Partially-Observed TreeCRFs70
To compute the probability of partial trees with partial marginalization, a variant of the Inside algorithm, the Masked Inside algorithm is proposed that supports different inference operations for different nodes (evaluation for the observed, marginalization for the latent, and rejection for nodes incompatible with the observed) with efficient parallelized implementation, thus significantly speeding up training and inference.
- Unified Dialog Model Pre-training for Task-Oriented Dialog Understanding and Generation69
SPACE is proposed, a novel unified pre-trained dialog model learning from large-scale dialog corpora with limited annotations, which can be effectively fine-tuned on a wide range of downstream dialog tasks.
- RAMM: Retrieval-augmented Biomedical Visual Question Answering with Multi-modal Pre-training66
This paper collects a new biomedical dataset named PMCPM which offers patient-based image-text pairs containing diverse patient situations from PubMed and proposes a retrieval-augmented pretrain-and-finetune paradigm named RAMM for biomedical VQA to overcome the data limitation issue.
- Aspect Sentiment Classification with Aspect-Specific Opinion Spans66
A neat and effective structured attention model by aggregating multiple linear-chain CRFs that allows the model to extract aspect-specific opinion spans and then evaluate sentiment polarity by exploiting the extracted opinion features.
- 64
- OntoED: Low-resource Event Detection with Ontology Embedding62
Based on the event ontology, OntoED can leverage and propagate correlation knowledge, particularly from data-rich to data-poor event types and can be applied to new unseen event types, by establishing linkages to existing ones.
- Learning Representations for Weakly Supervised Natural Language Processing Tasks62
Novel techniques for extracting features from n-gram models, Hidden Markov Models, and other statistical language models are investigated, including a novel Partial Lattice Markov Random Field model.
- Joint bilingual name tagging for parallel corpora58
Two novel approaches to jointly and consistently extract names from parallel corpora using standard linear-chain Conditional Random Fields as the learning framework, incorporating cross-lingual features propagated between two languages are proposed.
- Proton: Probing Schema Linking Information from Pre-trained Language Models for Text-to-SQL Parsing56
A novel framework to elicit relational structures from large-scale pre-trained language models (PLMs) via a probing procedure based on Poincaré distance metric, and use the induced relations to augment current graph-based parsers for better schema linking is proposed.
- VECO: Variable and Flexible Cross-lingual Pre-training for Language Understanding and Generation56
A cross-attention module is plugged into the Transformer encoder to explicitly build the interdependence between languages and outperforms all existing cross-lingual models and state-of-the-art Transformer variants on WMT14 English-to-German and English- to-French translation datasets.
- Preview, Attend and Review: Schema-Aware Curriculum Learning for Multi-Domain Dialogue State Tracking56
This paper proposes a model-agnostic framework called Schema-aware Curriculum Learning for Dialog State Tracking (SaCLog), which consists of a preview module that pre-trains a DST model with schema information, a curriculum module that optimizes the model with CL, and a review module that augments mispredicted data to reinforce the CL training.
- Dynamic Hybrid Relation Exploration Network for Cross-Domain Context-Dependent Semantic Parsing52
This paper presents a dynamic graph framework that is capable of effectively modelling contextual utterances, tokens, database schemas, and their complicated interaction as the conversation proceeds, and employs a dynamic memory decay mechanism that incorporates inductive bias to integrate enriched contextual relation representation.
- 52
- Relation Extraction as Open-book Examination50
This work regards RE as an open-book examination and proposes a new semiparametric paradigm of retrieval-enhanced prompt tuning for relation extraction, which can achieve state-of-the-art in both standard supervised and few-shot settings.
- Mining translations of OOV terms from the web through cross-lingual query expansion47
This paper proposes a novel approach to automatically translate OOV terms on the fly through cross-lingual query expansion that has achieved an inclusion rate of 95% and overall translation accuracy of 90%, outperforming state-of-the-art OOV translation techniques.
- DialogueCSE: Dialogue-based Contrastive Learning of Sentence Embeddings45
The proposed DialogueCSE introduces a novel matching-guided embedding mechanism, which generates a context-aware embedding for each candidate response embedding according to the guidance of the multi-turn context-response matching matrices and minimizes the contrastive loss across all pairs.
- Noisy BiLSTM-Based Models for Disfluency Detection45
It is demonstrated that strong performance can be achieved without extensively use of very large training data and that it is possible to be robust across data sets with noisy training approach in which insertion is the most useful noise for augmenting training data.
- Duplex Conversation: Towards Human-like Interaction in Spoken Dialogue Systems44
The concept of full-duplex in telecommunication is used to demonstrate what a human-like interactive experience should be and how to achieve smooth turn-taking through three subtasks: user state detection, backchannel selection, and barge-in detection.
- DAMO-NLP at SemEval-2022 Task 11: A Knowledge-based System for Multilingual Named Entity Recognition43
- Structure-Level Knowledge Distillation For Multilingual Sequence Labeling42
This paper proposes two novel KD methods based on structure-level information that approximately minimizes the distance between the student’s and the teachers’ structure- level probability distributions, and aggregates theructure-level knowledge to local distributions and minimizesThe distance between two local probability distributions.
- Improved Arabic Dialect Classification with Social Media Data42
This paper proposes an approach to improving Arabic dialect classification with semi-supervised learning: multiple classifiers are trained with weakly supervised, strongly supervised, and unsupervised data, which yields significant and consistent improvement on two different test sets.
- Hierarchical System Combination for Machine Translation42
A hierarchical system combination framework for machine translation that integrates multiple MT systems’ output at the word-, phrase- and sentence- levels achieves better translation quality with much less redecoding time.
- 41
- Extracting Action and Event Semantics from Web Text41
A novel system, PREPOST, is described that tackles the problem of extracting the preconditions and effects of actions and events, two important kinds of knowledge for connecting world state and the actions that affect it.
- 39
- Exploring Representation-Learning Approaches to Domain Adaptation39
This work investigates unsupervised techniques for representation learning that provide new features which are stable across domains, in that they are predictive in both the training and out-of-domain test data.
- 39
- Comparative Effectiveness of Digital Cognitive Behavioral Therapy vs Medication Therapy Among Patients With Insomnia38
It is suggested that dCBT-I needs to be thoughtfully combined with medication for long-term benefits in the treatment of insomnia and that further research into the design, implementation, and delivery of dC BT-I in terms of engagement and stability is required.
- Deep cascaded cross-modal correlation learning for fine-grained sketch-based image retrieval38
This paper forms FG-SBIR as a coarse-to-fine process, and proposes a Deep Cascaded Cross-modal Ranking Model (DCCRM) that can exploit all the beneficial multimodal information in sketches and annotated images and improve both the retrieval efficiency and the top-K ranked effectiveness.
- 34
- Monte Carlo Ensemble Neural Network for the diagnosis of Alzheimer’s disease34
The experiments show that the AD classification accuracy of the Monte Carlo Ensemble Neural Network model is comparable when using high- and low-resolution brain images, suggesting the insensitivity of the MCENN to image resolution.
- Semi-supervised Convolutional Networks for Translation Adaptation with Tiny Amount of In-domain Data34
A method which uses semi-supervised convolutional neural networks (CNNs) to select in-domain training data for statistical machine translation and can improve the performance up to 3.1 BLEU, which is significant better than three state-of-the-art language model based data selection methods.
- 33
- Sketch-based image retrieval with deep visual semantic descriptor32
The main contribution of this work is that it effectively applies the deep visual semantic descriptor to enable deep sketch-image matching, which has provided a more reasonable base for us to fuse local low-level visual features with high-level semantic features by determining an optimal correlated mapping.
- 30
- 30
- 29
- 27
- A Unified Span-Based Approach for Opinion Mining with Syntactic Constituents27
The experimental results show that the proposed unified span-based approach achieves significant improvements over previous works in the exact F1 score and reduces the number of wrongly-predicted opinion expressions and roles, showing the effectiveness of the method.
- VECO: Variable Encoder-decoder Pre-training for Cross-lingual Understanding and Generation27
A variable encoder-decoder (VECO) pre-training approach to unify the two mainstreams in both model architectures and pre- training tasks, which delivers new state-of-the-art results on various cross-lingual understanding tasks of the XTREME benchmark.
- 27
- 26
- 25
- 25
- ARAML: A Stable Adversarial Training Framework for Text Generation24
Experiments show that the proposed novel framework, Adversarial Reward Augmented Maximum Likelihood (ARAML), can outperform state-of-the-art text GANs with a more stable training process.
- 24
- 24
- 24
- 23
- 22
- 21
- 21
- NAST: A Non-Autoregressive Generator with Word Alignment for Unsupervised Text Style Transfer20
A non-autoregressive generator for unsupervised text Style Transfer (NAST), which focuses on the connections between aligned words, which learns the word-level transfer between styles and can significantly improve the overall performance and provide explainable word alignments.
- 19
- MuVER: Improving First-Stage Entity Retrieval with Multi-View Entity Representations18
This work proposes Multi-View Entity Representations (MuVER), a novel approach for entity retrieval that constructs multi-view representations for entity descriptions and approximates the optimal view for mentions via a heuristic searching method.
- 18
- 17
- 17
- Unsupervised Learning of Deterministic Dialogue Structure with Edge-Enhanced Graph Auto-Encoder17
A conversational graph (CG) is proposed to represent deterministic dialogue structure where nodes and edges represent the utterance and context information respectively and a self-supervised objective is introduced with the response selection task to guide the unsupervised learning of the dialogue structure.
- 16
- 16
- 16
- 16
- 16
- Low-resource extraction with knowledge-aware pairwise prototype learning16
K-HPN is able to recognize inherent correlation among classes, where each class is represented as a prototype on the hypersphere and outperforms previous methods of KE, particularly with low-resource training data regimes.
- 16
- 16
- 16
- A Semantic-based Method for Unsupervised Commonsense Question Answering15
This paper presents a novel SEmantic-based Question Answering method (SEQA) for unsupervised commonsense question answering that first generates a set of plausible answers with generative models, and then uses these plausible answers to select the correct choice by considering the semantic similarity between each plausible answer and each choice.
- 14
- 14
- 13
- Knowledge-aware Named Entity Recognition with Alleviating Heterogeneity13
A new KB-aware NER framework (KaNa), which utilizes type-heterogeneous knowledge to improve NER, and demonstrates KaNa's state-of-the-art performance on five public benchmark datasets from different domains.
- 13
- 12
- 12
- 12
- 12
- 12
- 12
- 12
- 11
- 11
- Learning to Ask for Data-Efficient Event Argument Extraction (Student Abstract)11
This study cast EAE as a question-based cloze task and empirically analyze fixed discrete token template performance, and proposes a novel approach called “Learning to Ask,” which can learn optimized question templates for EAE without human annotations.
- 10
- 10
- 10
- COPA : Efficient Vision-Language Pre-training through Collaborative Object- and Patch-Text Alignment10
- 10
- 8
- 8
- 8
- 8
- 8
- 8
- 8
- 8
- Bridging the Gap between Synthetic and Natural Questions via Sentence Decomposition for Semantic Parsing7
This paper proposes a decomposition-based method to unify the sentence structures of questions, which benefits the generalization to natural questions and extends to the task of sentence embedding learning, and observes substantial improvements on sentence retrieval.
- 7
- 7
- 7
- 6
- 6
- 6
- 6
- 6
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- Knowledgeable Salient Span Mask for Enhancing Language Models as Knowledge Base3
This work develops two solutions to help the model learn more knowledge from unstructured text in a fully self-supervised manner and is the first to explore fullySelf Supervised learning of knowledge in continual pre-training.
- 3
- 3
- 3
- 3
- 3
- 3
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- Extend Model Merging from Fine-Tuned to Pre-Trained Large Language Models via Weight Disentanglement1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
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 2026-10-13. 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.