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
- Shouxun LinSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works323 from public data
- 491
- Large Language Model Alignment: A Survey338
This survey endeavors to furnish an extensive exploration of alignment methodologies designed for LLMs, in conjunction with the extant capability research in this domain, and envision bridging the gap between the AI alignment research community and the researchers engrossed in the capability exploration of LLMs for both capable and safe LLMs.
- Evaluating Large Language Models: A Comprehensive Survey328
This survey endeavors to offer a panoramic perspective on the evaluation of LLMs, and categorizes the evaluated into three major groups: knowledge and capability evaluation, alignment evaluation and safety evaluation, which will channel their evolution into a direction that maximizes societal benefit while minimizing potential risks.
- Maximum entropy based phrase reordering model for statistical machine translation270
A novel reordering model for phrase-based statistical machine translation (SMT) that uses a maximum entropy (MaxEnt) model to predicate reorderings of neighbor blocks (phrase pairs) that obtains significant improvements in BLEU score on the NIST MT-05 and IWSLT-04 tasks.
- Variational Neural Machine Translation211
This paper builds a neural posterior approximator conditioned on both the source and the target sides, and equip it with a reparameterization technique to estimate the variational lower bound, and shows that the proposed variational neural machine translation achieves significant improvements over the vanilla neural machinetranslation baselines.
- 173
- Modeling Source Syntax for Neural Machine Translation150
This paper shows that source syntax can be explicitly incorporated into NMT effectively to provide further improvements, and linearize parse trees of source sentences to obtain structural label sequences.
- Semantic Similarity from Natural Language and Ontology Analysis135
This book proposes an in-depth characterization of existing proposals for semantic similarity estimation by discussing their features, the assumptions on which they are based and empirical results regarding their performance in particular applications, and provides a detailed discussion on the foundations of semantic measures.
- 134
- 131
- Accelerating Neural Transformer via an Average Attention Network129
The proposed average attention network is applied on the decoder part of the neural Transformer to replace the original target-side self-attention model and enables the neuralTransformer to decode sentences over four times faster than its original version with almost no loss in training time and translation performance.
- Shallow Convolutional Neural Network for Implicit Discourse Relation Recognition116
A Shallow Convolutional Neural Network (SCNN) is proposed, which contains only one hidden layer but is effective in relation recognition, which alleviates the overfitting problem, while the convolution and nonlinear operations help preserve the recognition and generalization ability of the model.
- Neural Machine Translation Advised by Statistical Machine Translation96
Experimental results on Chinese-English translation show that the proposed approach achieves significant and consistent improvements over state-of-the-art NMT and SMT systems on multiple NIST test sets.
- 92
- Variational Recurrent Neural Machine Translation84
A novel variational recurrent neural machine translation model that introduces a series of latent random variables to model the translation procedure of a sentence in a generative way, instead of a single latent variable.
- 82
- Large Language Model Safety: A Holistic Survey81
This survey provides a comprehensive overview of the current landscape of LLM safety, covering four major categories: value misalignment, robustness to adversarial attacks, misuse, and autonomous AI risks.
- 79
- 70
- Parsing the Penn Chinese Treebank with Semantic Knowledge70
Analysis of performance improvement indicates that semantic knowledge is helpful for nominal compounds, coordination, and N⋄V tagging disambiguation, as well as alleviating the sparseness of information available in treebank.
- Hierarchical Modeling of Global Context for Document-Level Neural Machine Translation68
This paper proposes a hierarchical model to learn the global context for document-level neural machine translation (NMT), which feedback the extracted global document context to each word in a top-down fashion to distinguish different translations of a word according to its specific surrounding context.
- Enhanced aspect-based sentiment analysis models with progressive self-supervised attention learning66
This work proposes a progressive self-supervised attention learning approach for attentional ABSA models that enhances the performance of all three models and integrates the proposed approach into three state-of-the-art neuralABSA models.
- Modeling the Translation of Predicate-Argument Structure for SMT66
Two discriminative, feature-based models to exploit predicate-argument structures for statistical machine translation are proposed: 1) a predicate translation model and 2) an argument reordering model.
- RiSAWOZ: A Large-Scale Multi-Domain Wizard-of-Oz Dataset with Rich Semantic Annotations for Task-Oriented Dialogue Modeling64
RiSAWOZ is a large-scale multi-domain Chinese Wizard-of-Oz dataset with Rich Semantic Annotations, which contains 11.2K human-to-human multi-turn semantically annotated dialogues, with more than 150K utterances spanning over 12 domains, which is larger than all previous annotated H2H conversational datasets.
- 63
- Error Detection for Statistical Machine Translation Using Linguistic Features63
This work uses a maximum entropy classifier to predict translation errors by integrating word posterior probability feature and linguistic features and shows that linguistic features alone outperform word posterior probability based confidence estimation in error detection.
- Translating Phrases in Neural Machine Translation58
This work proposes a method to translate phrases in NMT by integrating a phrase memory storing target phrases from a phrase-based statistical machine translation (SMT) system into the encoder-decoder architecture of NMT.
- Encoding Gated Translation Memory into Neural Machine Translation56
This paper proposes a novel method to combine the strengths of both TM and neural machine translation (NMT) for high-quality translation, which treats the target translation of a TM match as an additional reference input and encode it into NMT with an extra encoder.
- Lexical Chain Based Cohesion Models for Document-Level Statistical Machine Translation51
Experiments on large-scale training data show that the two proposed lexical chain based cohesion models can substantially improve translation quality in terms of BLEU and that the probability cohesion model outperforms previous models based on lexical cohesion devices.
- Efficient Object-Level Visual Context Modeling for Multimodal Machine Translation: Masking Irrelevant Objects Helps Grounding48
An Object-level Visual Context modeling framework (OVC) to efficiently capture and explore visual information for multimodal machine translation by encouraging MMT to ground translation on desirable visual objects by masking irrelevant objects in the visual modality.
- 47
- A Topic Similarity Model for Hierarchical Phrase-based Translation46
This work proposes a topic similarity model to exploit topic information at the synchronous rule level for hierarchical phrase-based translation, and shows that this model significantly improves the translation performance over the baseline on NIST Chinese-to-English translation experiments.
- Multilingual Large Language Models: A Systematic Survey45
This paper presents a detailed taxonomy and roadmap covering the assessment of MLLMs' cross-lingual knowledge, reasoning, alignment with human values, safety, interpretability and specialized applications, and showcases how these models have driven innovation and improvements in these specialized fields.
- PEIT: Bridging the Modality Gap with Pre-trained Models for End-to-End Image Translation41
Experiments on the curated ECOIT benchmark dataset demonstrate that PEIT substantially outperforms both cascaded image translation systems (OCR+MT) and previous strong end-to-end image translation model, with fewer parameters and faster decoding speed.
- 40
- Cycle-Consistent Adversarial Autoencoders for Unsupervised Text Style Transfer40
This paper proposes a novel neural approach to unsupervised text style transfer which it refers to as Cycle-consistent Adversarial autoEncoders (CAE) trained from non-parallel data that enhances the capacity of the adversarial style transfer networks in content preservation.
- Learning Translation Boundaries for Phrase-Based Decoding40
It is shown that using constraints based on translation boundaries predicted by the classifiers achieves significant improvements over the baseline on large-scale Chinese-to-English translation experiments and the new constraints also significantly outperform constituent boundary based syntactic constrains.
- The Box is in the Pen: Evaluating Commonsense Reasoning in Neural Machine Translation39
It is demonstrated that neural machine translation performs poorly on commonsense reasoning of the three ambiguity types in terms of both reasoning accuracy and reasoning consistency.
- 39
- Efficient Cluster-Based k-Nearest-Neighbor Machine Translation38
A more efficient kNN-MT is explored and a cluster-based Compact Network for feature reduction in a contrastive learning manner to compress context features into 90+% lower dimensional vectors is proposed and shown good generalization on unseen domains.
- 36
- Attention Focusing for Neural Machine Translation by Bridging Source and Target Embeddings36
Experiments and analysis presented in this paper demonstrate that the proposed bridging models are able to significantly improve quality of both sentence translation, in general, and alignment and translation of individual source words with target words, in particular.
- A Topic-Based Coherence Model for Statistical Machine Translation36
This paper proposes a topic-based coherence model to produce coherence for document translation, in terms of the continuity of sentence topics in a text, and adopts a maximum entropy classifier to predict the target coherence chain that defines a linear topic structure for the target document.
- 35
- BiPaR: A Bilingual Parallel Dataset for Multilingual and Cross-lingual Reading Comprehension on Novels34
BiPaR is presented, a bilingual parallel novel-style machine reading comprehension (MRC) dataset, developed to support multilingual and cross-lingual reading comprehension, and finds that BiPaR offers good diversification in prefixes of questions, answer types and relationships between questions and passages.
- A Sense-Based Translation Model for Statistical Machine Translation34
A sense-based translation model to integrate word senses into statistical machine translation based on a nonparametric Bayesian topic model that automatically learns sense clusters for words in the source language is proposed.
- M3KE: A Massive Multi-Level Multi-Subject Knowledge Evaluation Benchmark for Chinese Large Language Models33
M3KE, a Massive Multi-Level Multi-Subject Knowledge Evaluation benchmark, is developed to measure knowledge acquired by Chinese large language models by testing their multitask accuracy in zero- and few-shot settings, guaranteeing a standardized and unified assessment process.
- Exploring Multilingual Concepts of Human Values in Large Language Models: Is Value Alignment Consistent, Transferable and Controllable across Languages?31
This paper empirically confirms the presence of value concepts within LLMs in a multilingual format and validates the feasibility of cross-lingual control over value alignment capabilities of LLMs, leveraging the dominant language as a source language.
- A neural generative autoencoder for bilingual word embeddings31
A neural generative bilingual autoencoder (NGBAE) which introduces a latent variable to explicitly induce the underlying semantics of bilingual text to obtain better BWEs from more robust bilingual semantics by modeling the semantic distributions ofilingual text.
- 30
- Towards Understanding Multi-Task Learning (Generalization) of LLMs via Detecting and Exploring Task-Specific Neurons29
This paper detects task-sensitive neurons in LLMs via gradient attribution on task-specific data, and finds that the overlap of task-specific neurons is strongly associated with generalization and specialization across tasks.
- Enhancing Language Models in Statistical Machine Translation with Backward N-grams and Mutual Information Triggers29
A backward language model that augments the conventional forward language model and a mutual information trigger model which captures long-distance dependencies that go beyond the scope of standard n-gram language models are presented.
- A syntax-driven bracketing model for phrase-based translation29
A new model is presented that automatically learns syntactic constraints, including but not limited to constituent matching/violation, from training corpus and achieves a substantial improvement over the baseline which is not syntactically informed.
- Do Large Language Models Mirror Cognitive Language Processing?28
Experimental results indicate that pre-training data size and model scaling are positively correlated with LLM-brain similarity, and alignment training can significantly improve LLM-brain similarity.
- A dependency treelet string correspondence model for statistical machine translation28
It is argued that the DTSC model proposed here is capable of lexicalization, generalization, and handling discontinuous phrases which are very desirable for machine translation.
- CogAlign: Learning to Align Textual Neural Representations to Cognitive Language Processing Signals27
In CogAlign, a shared encoder equipped with a modality discriminator is used to alternatively encode textual and cognitive inputs to capture their differences and commonalities, and a text-aware attention mechanism is proposed to detect task-related information and to avoid using noise in cognitive features.
- 27
- 27
- CToolEval: A Chinese Benchmark for LLM-Powered Agent Evaluation in Real-World API Interactions26
This work proposes CToolEval, a benchmark designed to evaluate LLMs in the context of Chinese societal applications, featuring 398 APIs across 27 widely-used Apps that cover 14 domains, and presents an evaluation framework that simulates real-life scenarios, to facilitate the assessment of tool invocation ability of LLMs for tool learning and task completion ability for user interation.
- Early detection of hepatocellular carcinoma via no end-repair enzymatic methylation sequencing of cell-free DNA and pre-trained neural network26
By combining high-fidelity methylation data from NEEM-seq with the DeepTrace model, the method has great potential for HCC early detection with high sensitivity and specificity, making it potentially suitable for clinical applications.
- Lipschitz Constrained Parameter Initialization for Deep Transformers26
This paper empirically demonstrate that a simple modification made in the official implementation, which changes the computation order of residual connection and layer normalization, can significantly ease the optimization of deep Transformers and presents a parameter initialization method that leverages the Lipschitz constraint on the initialization of Transformer parameters that effectively ensures training convergence.
- Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data26
A radical enhanced Chinese word embedding model that uses conversion and radical escaping mechanisms to extract the intrinsic information in Chinese corpus and has effectively improved the accuracy of word vector expression and the direct relevance of similar words.
- 25
- TED-CDB: A Large-Scale Chinese Discourse Relation Dataset on TED Talks25
The TED-CDB comprises a large set of TED talks in Chinese that have been manually annotated according to the goals and principles of Penn Discourse Treebank, but adapted to features that are not present in English.
- 24
- Learning Source Phrase Representations for Neural Machine Translation23
This paper proposes an attentive phrase representation generation mechanism which is able to generate phrase representations from corresponding token representations and incorporates the generated phrase representations into the Transformer translation model to enhance its ability to capture long-distance relationships.
- Bilingual Correspondence Recursive Autoencoder for Statistical Machine Translation23
This paper proposes a new neural network model called Bilingual Correspondence Recursive Autoencoder (BCorrRAE) to model bilingual phrases in translation and incorporates word alignments into BCorrRAE to allow it freely access bilingual constraints at different levels.
- Bilingual Correspondence Recursive Autoencoders for Statistical Machine Translation23
This paper proposes a new neural network model called Bilingual Correspondence Recursive Autoencoder (BCorrRAE) to model bilingual phrases in translation and incorporates word alignments into BCorrRAE to allow it freely access bilingual constraints at different levels.
- 23
- 22
- Adaptive Differential Privacy for Language Model Training22
This paper proposes an Adaptive Differential Privacy (ADP) framework for language modeling without resorting to prior privacy information, and estimates the probability that a linguistic item contains privacy based on a language model.
- CogTaskonomy: Cognitively Inspired Task Taxonomy Is Beneficial to Transfer Learning in NLP22
Experiments on 12 NLP tasks demonstrate that the proposed CogTaxonomy is able to guide transfer learning, achieving performance competitive to the Analytic Hierarchy Process used in visual Taskonomy but without requiring exhaustive pairwise O(m^2) task transferring.
- Modeling Long Context for Task-Oriented Dialogue State Generation22
A multi-task learning model with a simple yet effective utterance tagging technique and a bidirectional language model as an auxiliary task for task-oriented dialogue state generation to solve the problem that the performance of the baseline significantly drops when the input dialogue context sequence is long.
- 21
- Variational Neural Discourse Relation Recognizer21
This paper introduces neural discourse relation models to approximate the prior and posterior distributions of the latent variable, and employs these approximated distributions to optimize a reparameterized variational lower bound that allows VarNDRR to be trained with standard stochastic gradient methods.
- BattRAE: Bidimensional Attention-Based Recursive Autoencoders for Learning Bilingual Phrase Embeddings20
A bidimensional attention based recursive autoencoder (BattRAE) is proposed to integrate clues and sourargetinteractions at multiple levels of granularity into bilingualphrase representations to evaluate the effectiveness of this state-of-the-art SMT system.
- Effective Data Augmentation Approaches to End-to-End Task-Oriented Dialogue19
Experimental results on the CamRest676 and KVRET datasets demonstrate that each of the four data augmentation approaches is able to obtain a significant improvement over a strong baseline in terms of Success F1 score and that the ensemble of theFour approaches achieves the state-of-the-art results in the two datasets.
- Topic-Based Dissimilarity and Sensitivity Models for Translation Rule Selection19
This work introduces two topic-based models for translation rule selection which incorporates global topic information into translation disambiguation and integrates the proposed topic dissimilarity and sensitivity model into hierarchical phrase-based machine translation for synchronoustranslation rule selection.
- Mitigating Privacy Seesaw in Large Language Models: Augmented Privacy Neuron Editing via Activation Patching18
Experimental results show that the proposed APNEAP is capable of alleviating the privacy seesaw phenomenon and offers a more stable and reliable approach to privacy protection in LLMs than previous methods.
- Efficient Context-Aware Neural Machine Translation with Layer-Wise Weighting and Input-Aware Gating18
This paper finds that representations from shallow layers of a pre-trained sentence-level encoder play a vital role in source context encoding, and proposes to perform source context encoding upon weighted combinations of pre-trained encoder layers' outputs.
- Topic-based term translation models for statistical machine translation18
This article investigates three issues of term translation in the context of statistical machine translation and proposes three corresponding models: a term translation disambiguation model which selects desirable translations for terms in the source language with domain information, a termtranslation consistency model that encourages consistent translations forterms with a high strength of translation consistency throughout a document, and a term unithood model that rewards translation hypotheses where source terms are translated into target strings as a whole unit.
- 18
- CMoralEval: A Moral Evaluation Benchmark for Chinese Large Language Models17
This paper develops a morality taxonomy and a set of fundamental moral principles that are not only rooted in traditional Chinese culture but also consistent with contemporary societal norms and establishes a platform with AI-assisted instance generation to streamline the annotation process.
- KaFSP: Knowledge-Aware Fuzzy Semantic Parsing for Conversational Question Answering over a Large-Scale Knowledge Base17
A knowledge-aware fuzzy semantic parsing framework that defines fuzzy comparison operations in the grammar system for uncertain reasoning based on the fuzzy set theory and proposes a multi-label classification framework to not only capture correlations between entity types and relations but also detect knowledge base information relevant to the current utterance.
- Learning better discourse representation for implicit discourse relation recognition via attention networks17
A full attention model that combines their strengths into a unified framework is proposed that significantly benefits from highlighting relation-relevant words and yields competitive and even better results against several state-of-the-art systems.
- A Context-Aware Topic Model for Statistical Machine Translation17
A context-aware topic model is proposed for lexical selection that not only models local contexts and global topics but also captures their correlations, and a Gibbs sampling algorithm is introduced for statistical estimation and inference.
- ChatSOP: An SOP-Guided MCTS Planning Framework for Controllable LLM Dialogue Agents16
This work proposes **ChatSOP**, a novel SOP-guided Monte Carlo Tree Search (MCTS) planning framework designed to enhance the controllability of LLM-driven dialogue agents and proposes a novel method that integrates Chain of Thought reasoning with supervised fine-tuning for SOP prediction and utilizes SOP-guided Monte Carlo Tree Search for optimal action planning during dialogues.
- 16
- 16
- 16
- Multi-Head Highly Parallelized LSTM Decoder for Neural Machine Translation16
This work approximates full LSTM context modelling by computing hidden states and gates with the current input and a simple bag-of-words representation of the preceding tokens context, and then connects the outputs of each parallel step with computationally cheap element-wise computations to enable sequence-level parallelization of LSTMs.
- Sentence Weighting for Neural Machine Translation Domain Adaptation16
A domain similarity metric is introduced to evaluate the relevance between a sentence and an available entire domain dataset and it is shown that the sentence weighting method is able to achieve an significant improvement over strong baselines.
- 15
- JoinER-BART: Joint Entity and Relation Extraction With Constrained Decoding, Representation Reuse and Fusion15
This article presents a new generative JERE framework which employs a constrained classifier which only predicts either a token of the input sentence or a relation in each decoding step, and reuse representations from the pre-trained BART encoder in the classifier instead of a newly trained weight matrix, which empirically leads to better performance.
- 15
- Re-embedding Difficult Samples via Mutual Information Constrained Semantically Oversampling for Imbalanced Text Classification15
A Mutual Information constrained Semantically Oversampling framework (MISO) that can generate anchor instances to help the backbone network determine the re-embedding position of a non-overlapping representation for each difficult sample.
- Simplifying Neural Machine Translation with Addition-Subtraction Twin-Gated Recurrent Networks15
Experiments on WMT14 translation tasks demonstrate that ATR-based neural machine translation can yield competitive performance on English-German and English-French language pairs in terms of both translation quality and speed.
- CKDST: Comprehensively and Effectively Distill Knowledge from Machine Translation to End-to-End Speech Translation14
This paper proposes a comprehensive knowledge distillation framework for speech translation, CKDST, which is capable of comprehensively and effectively distilling knowledge from machine translation to speech translation from two perspectives: cross-modal contrastive representation distillation and simultaneous de-coupled knowledge distillation.
- Mining parallel sentences from internet with multi-view knowledge distillation for low-resource language pairs14
The multi-view knowledge distillation model (MvKD) that use the knowledge of high-resource language pairs transfer into low-resource languages by leveraging internal language invariant in different languages is proposed.
- A Learning-Exploring Method to Generate Diverse Paraphrases with Multi-Objective Deep Reinforcement Learning14
This work proposes a learning-exploring method to generate sentences as learning objectives from the learned data distribution, and employs reinforcement learning to combine these new learning objectives for model training.
- Cache-based Document-level Neural Machine Translation.14
A cache-based approach to document-level neural machine translation by capturing contextual information either from recently translated sentences or the entire document, which achieves substantial improvements over several state-of-the-art SMT and NMT baselines.
- Linguistically annotated BTG for statistical machine translation14
This paper proposes a Linguistically Annotated BTG (LABTG) that conveys linguistic knowledge of source-side syntax structures to BTG hierarchical structures through linguistic annotation and presents an annotation algorithm that captures syntactic information for BTG nodes.
- Towards Robust In-Context Learning for Machine Translation with Large Language Models13
A robust method to enable LLMs to achieve robust translation with ICL by considering both sentence- and word-level similarity, which provides fine-grained demonstrations that effectively prompt the translation of LLMs.
- Tab-CQA: A Tabular Conversational Question Answering Dataset on Financial Reports13
Tab-CQA, a tabular CQA dataset created from Chinese financial reports that are extracted from listed companies in a wide range of different sectors in the past 30 years, is proposed and Experiment results show that Tab-CZA is a very challenging dataset, where a huge performance gap exists between human and neural models.
- Probing Word Translations in the Transformer and Trading Decoder for Encoder Layers13
This work shows that translation already happens progressively in encoder layers and even in the input embeddings, and suggests a Transformer configuration change that can increase speed by up to a factor 2.3 with small gains in translation quality, while an 18-4 deep encoder configuration boosts translation quality by +1.4.
- Autocorrect in the Process of Translation — Multi-task Learning Improves Dialogue Machine Translation13
A joint learning method to identify omission and typo, and utilize context to translate dialogue utterances, and shows that the proposed method improves translation quality by 3.2 BLEU over the baselines.
- Dynamically Adjusting Transformer Batch Size by Monitoring Gradient Direction Change13
This paper analyzes how increasing batch size affects gradient direction, and proposes to automatically and dynamically determine batch sizes by accumulating gradients of mini-batches and performing an optimization step at just the time when the direction of gradients starts to fluctuate.
- Refinements in BTG-based Statistical Machine Translation13
Two refinements for BTG-based SMT to achieve better reordering and higher-speed decoding are introduced, which include (1) reordering heuristics to prevent incorrect swapping and reduce search space, and (2) special phrases with tags to indicate sentence beginning and ending.
- A linguistically annotated reordering model for BTG-based statistical machine translation13
A linguistically annotated reordering model for BTG-based statistical machine translation that incorporates linguistic knowledge to predict orders for both syntactic and non-syntactic phrases is proposed.
- 12
- Short text matching model with multiway semantic interaction based on multi-granularity semantic embedding12
A sentence matching model with multiway semantic interaction based on multi-granularity semantic embedding(MSIM) to dispose of the problem of Chinese short text matching, outperforming other methods.
- Exploring Bilingual Parallel Corpora for Syntactically Controllable Paraphrase Generation12
A novel end-to-end framework to leverage existing large-scale bilingual parallel corpora to generate paraphrases under the control of syntactic exemplars and employ a latent variable to model the syntactic style of a given exemplar in order to guide the two decoders for generation.
- Bilingual Lexical Cohesion Trigger Model for Document-Level Machine Translation12
A bilingual lexical cohesion trigger model is integrated into hierarchical phrase-based machine translation and achieves an absolute improvement of 0.85 BLEU points on average over the baseline on NIST Chinese-English test sets.
- KnowRL: Boosting LLM Reasoning via Reinforcement Learning with Minimal-Sufficient Knowledge Guidance11
This work proposes \textbf{KnowRL} (Knowledge-Guided Reinforcement Learning), an RL training framework that treats hint design as a minimal-sufficient guidance problem and identifies a pruning interaction paradox and explicitly optimize for robust subset curation under this dependency structure.
- Rewiring the Transformer with Depth-Wise LSTMs11
It is shown that layer normalization and feed-forward computation within a Transformer layer can be absorbed into depth-wise LSTMs connecting pure Transformer attention layers, and the effectiveness of depth-wise LSTM on the convergence and performance of deep Transformers.
- FollowEval: A Multi-Dimensional Benchmark for Assessing the Instruction-Following Capability of Large Language Models11
This paper has evaluated various LLMs using the FollowEval benchmark and found that their performance significantly lags behind that of humans, highlighting the considerable room for improvement in the instruction-following ability of these models.
- Domain-Aware Self-Attention for Multi-Domain Neural Machine Translation11
A domain-aware self-attention mechanism that jointly learns domain representations with the single NMT model that out-performs a strong baseline built on the Transformer and other previous multi-domain NMT approaches.
- An Empirical Study on Adversarial Attack on NMT: Languages and Positions Matter11
This paper exploits the attention distribution of the victim model to attack source sentences at positions that have a strong association with front target words and demonstrates that the attention-based adversarial attack is more effective than adversarial attacks by sampling positions randomly or according to gradients.
- The Direct Path May Not Be The Best: Portuguese-Chinese Neural Machine Translation11
This paper addresses the feasibility of creating a MT system for Portuguese-Chinese, using only freely available resources, by experimenting with various approaches to pairing source and target parallel data during training, and finds approaches whose performance is higher than that of the strong baseline consisting of an MT service provided by an IT industry giant.
- Recurrent Neural Machine Translation11
A novel recurrent neural machine translation (RNMT), which not only preserves the ability to model varying-length source annotations but also better captures long-term dependencies, and can implicitly reflect the alignment to source sentence.
- 10
- 10
- 10
- Syntactically-Informed Unsupervised Paraphrasing with Non-Parallel Data10
This paper proposes a syntactically-informed unsupervised paraphrasing model based on conditional variational auto-encoder (VAE) which can generate texts in a specified syntactic structure and designs a two-stage learning method to effectively train the model using non-parallel data.
- TGEA: An Error-Annotated Dataset and Benchmark Tasks for TextGeneration from Pretrained Language Models10
TGEA, an error-annotated dataset with multiple benchmark tasks for text generation from pretrained language models (PLMs), is proposed, which is the first dataset with comprehensive annotations for PLM-generated texts, which facilitates the diagnostic evaluation ofPLM-based text generation.
- Modeling Task-Aware MIMO Cardinality for Efficient Multilingual Neural Machine Translation10
A Multi-Input-Multi-Output (MIMO) architecture that allows each transformation of the block to have its own input and a task-aware attention mechanism to learn to selectively utilize individual transformations from a set of transformations for different translation directions is presented.
- Generating Highly Relevant Questions10
Two methods to address the issue of generic and undiversified questions that are poorly relevant to the given passage and target answer are proposed.
- 9
- FineMath: A Fine-Grained Mathematical Evaluation Benchmark for Chinese Large Language Models9
FineMath is proposed, a fine-grained mathematical evaluation benchmark dataset for assessing Chinese LLMs, created to cover the major key mathematical concepts taught in elementary school math, enabling in-depth analysis of mathematical reasoning abilities of LLMs.
- MoE-CT: A Novel Approach For Large Language Models Training With Resistance To Catastrophic Forgetting9
By preserving the base model's integrity and focusing on strategic parameter expansion, the methodology advances multilingual language modeling and represents a significant step forward for low-resource language inclusion in LLMs, indicating a fruitful direction for future research in language technologies.
- 9
- Learning Structural Information for Syntax-Controlled Paraphrase Generation9
A syntax encoder based on tree-transformer to capture parent-child and sibling relations is designed and validated, and the effectiveness of SI-SCP to generate diverse paraphrases with retrieved syntactic structures is demonstrated.
- Balanced Joint Adversarial Training for Robust Intent Detection and Slot Filling9
A Balanced Joint Adversarial Training (BJAT) model that applies a balance factor as a regularization term to the final loss function, which yields a stable training procedure that achieves significantly higher scores and substantially improve the robustness of both intent detection and slot filling.
- Neural Machine Translation with Decoding History Enhanced Attention9
A DHEA mechanism is reformulated to render NMT model better at selecting both source-side and target-side information, and achieves comparable results with the state-of-the-art.
- Automatic Long Sentence Segmentation for Neural Machine Translation9
A split and reordering model is introduced to collectively detect the optimal sequence of segmentation points for a long source sentence and achieves a substantial improvement on NIST Chinese-English translation tasks.
- Tagging Complex NEs with MaxEnt Models: Layered Structures Versus Extended Tagset9
It is empirically prove that the {start, continue, end, unique, other} tag set is the best tag set for NE recognition with MaxEnt models.
- Praetor: A Fine-Grained Generative LLM Evaluator with Instance-Level Customizable Evaluation Criteria8
Praiseor, a fine-grained generative LLM evaluator with instance-level customaz-able evaluation criteria, is proposed, which outperforms previous LLM evaluators and instruction-tuned LLMs on multiple benchmarks, setting new SOTA results.
- 8
- 8
- TJUNLP:System Description for the WMT23 Literary Task in Chinese to English Translation Direction8
Experimental results show that a data augmentation technique employed in the WMT23 machine translation evaluation task can effectively improve neural machine translation performance.
- 8
- Informative Language Representation Learning for Massively Multilingual Neural Machine Translation8
Two methods are proposed, language embedding embodiment and language-aware multi-head attention, to learn informative language representations to channel translation into right directions, and matrix-based language representations learned by these methods are capable of capturing rich linguistic typology features.
- FewFedWeight: Few-shot Federated Learning Framework across Multiple NLP Tasks8
Experiments show that FewFedWeight can significantly improve the performance of client models on 61% tasks with an average performance improvement rate of 30.5% over the baseline and substantially outperform FedAvg and other decentralized learning methods.
- Unsupervised and few-shot parsing from pretrained language models8
UPOA, an Unsupervised constituent Parsing model that calculates an Out Association score solely based on the self-attention weight matrix learned in a pretrained language model as the syntactic distance for span segmentation is proposed.
- Chinese WPLC: A Chinese Dataset for Evaluating Pretrained Language Models on Word Prediction Given Long-Range Context8
This paper presents a Chinese dataset for evaluating pretrained language models on Word Prediction given Long-term Context (Chinese WPLC), and proposes both automatic and manual selection strategies tailored to Chinese to guarantee that target words in passages collected from over 69K novels can only be predicted with long-term context beyond the scope of sentences containing the target words.
- I 2 R Multi-Pass Machine Translation System for IWSLT 20088
The design philosophy, overall architecture, each individual system and various system combination methods that the I 2 R system has explored, and the performance on development and test sets are reported in detail.
- Lexicalized Beam Thresholding Parsing with Prior and Boundary Estimates8
The use prior and boundary estimates as the approximation of outside probability and beam thresholding strategies based on these estimates are established and differentiated completed edges from incomplete edges paves the way for using boundary estimates in the edge-based beam chart parsing.
- Meta-photonics: A bridge between physical association and digital models in photonics7
This work is armed with different deep learning models to build intrinsic dimensions among different physical elements and bring together amplitude, phase, polarization conversion ratio, extinction ratio, unwrapping phase, surface electric field, and connected domain, most of which have never been reported, based on the results of the forward and inverse models.
- Recovering Gold from Black Sand: Multilingual Dense Passage Retrieval with Hard and False Negative Samples7
Experimental results show that mHFN outperforms strong sparse, dense and hybrid baselines and achieves new state-of-the-art performance on all languages.
- Modeling Homophone Noise for Robust Neural Machine Translation7
A robust neural machine translation (NMT) framework to deal with homophone errors is proposed and extensive experiments on Chinese→English translation demonstrate that the proposed method not only significantly outperforms baselines on noisy test sets withhomophone noise, but also achieves substantial improvements over them on clean texts.
- Alignment-consistent recursive neural networks for bilingual phrase embeddings7
This paper considers word alignments as a semantic bridge between the source and target phrases, and proposes two neural networks based on the conventional recursive autocoder, which exploitword alignments to generate alignment-consistent bilingual phrase structures.
- Graph-Based Collective Lexical Selection for Statistical Machine Translation7
A graph-based framework for collective lexical selection is established on a translation graph that captures not only local associations between source-side content words and their target translations but also targetside global dependencies in terms of relatedness among target items.
- Modeling Term Translation for Document-informed Machine Translation7
This paper investigates three issues of term translation in the context of documentinformed SMT and proposes three corresponding models that can achieve significant improvements over the baseline and evaluates their effectiveness on NIST ChineseEnglish translation tasks with large-scale training data.
- ICT System Description for the 2006 TC-STAR Run #2 SLT Evaluation7
Three systems were developed based on different techniques: system Confucius based on phrase, system Lynx based on tree-to-string alignment template and system Bruin based on BTG (Bracketing Transduction Grammar), which share the same phrase-based translation model and language model.
- 6
- 6
- Enhancing Chinese Word Segmentation via Pseudo Labels for Practicability6
Experiments show that the proposed semi-supervised neural method can not only keep the practi-cability of the lightweight student model but also improve the performance of segmentation effectively and is reliable and practical as a pre-processing step of the downstream NLP tasks at the minimum cost.
- 6
- 6
- A Test Suite for Evaluating Discourse Phenomena in Document-level Neural Machine Translation6
A test suite to evaluate three common discourse phenomena in English-Chinese translation: pronoun, discourse connective and ellipsis where discourse divergences lie across the two languages and suggests that the test suite can be used as a challenging benchmark test bed for evaluating document-level NMT.
- 6
- Convolution-Enhanced Bilingual Recursive Neural Network for Bilingual Semantic Modeling.6
A convolution-enhanced bilingual recursive neural network (ConvBRNN) is proposed, which not only exploits word alignments to guide the generation of phrase structures but also integrates multiple-level information of the generated phrase structures into bilingual semantic modeling.
- Unsupervised Discriminative Induction of Synchronous Grammar for Machine Translation6
A global log-linear model for synchronous grammar induction, which is capable of incorporating arbitrary features, is presented and a novel and efficient cube pruning based synchronous parsing algo- rithm is proposed.
- HTRDP evaluations on Chinese information processing and intelligent human-machine interface6
From 1991 to 2005, China’s High Technology Research and Development Program (HTRDP) sponsored a series of technology evaluations on Chinese information processing and intelligent human-machine interface, which is called HTRDP evaluations, or “863” evaluations in brief.
- 5
- Evaluating Chinese Large Language Models on Discipline Knowledge Acquisition via Memorization and Robustness Assessment5
Experimental findings indicate a positive correlation between the release time of LLMs and their memorization capabilities, but they struggle with variations in original question-options pairs, and findings suggest that question descriptions have a more significant impact on LLMs’ performance.
- 5
- 5
- 5
- 5
- Long Text Generation with Topic-aware Discrete Latent Variable Model5
This work builds a topic-aware latent code-guided text generation model that can generate more topic-relevant and coherent texts and proposes a span-level bag-of-words training objective for the model.
- Bridging between Cognitive Processing Signals and Linguistic Features via a Unified Attentional Network5
A unified attentional framework is presented that is composed of embedding, attention, encoding and predicting layers to selectively map cognitive processing signals to linguistic features and can be used to detect a wide range of linguistic features with a single cognitive dataset.
- Swing Distillation: A Privacy-Preserving Knowledge Distillation Framework5
A novel knowledge distillation method, swing distillation, is proposed, which can effectively protect the private information of the teacher model from flowing to the student model and is robust against the increasing privacy budget.
- Integrating Pre-trained Model into Rule-based Dialogue Management5
This paper introduces the DM of Carina Dialog System and proposes the "model-trigger" design to make the DM trainable thus scalable to scenario changes, and integrates pre-trained models and empower the DM with few-shot capability.
- Secoco: Self-Correcting Encoding for Neural Machine Translation5
Self-correcting Encoding (Secoco) is a framework that effectively deals with input noise for robust neural machine translation by introducing self- correcting predictors and is able to achieve significant improvements over strong baselines on two real-world test sets and a benchmark WMT dataset with good interpretability.
- Detecting and Translating Dropped Pronouns in Neural Machine Translation5
This work proposes a method to both automatically detect the dropped pronouns (DPs) and recover their translation equivalences rather than their original forms in source sentences, simultaneously performed as a sequence labeling task on source sentences.
- Improving Statistical Machine Translation with Selectional Preferences5
This paper incorporates preferences that verbs impose on subjects and objects into translation into machine translation and demonstrates that statistical machine translation using verbal selectional preferences can achieve statistically significant improvements over a state-of-the-art baseline.
- Learning Semantic Representations for Nonterminals in Hierarchical Phrase-Based Translation5
A framework to refine nonterminals in hierarchical translation rules with real-valued semantic representations is proposed and results show that the proposed model significantly improves translation quality on NIST test sets.
- Max-Margin Synchronous Grammar Induction for Machine Translation5
A max-margin estimation approach to discriminatively inducing synchronous grammars for machine translation, which directly optimizes translation quality measured by BLEU, and test the effectiveness of the framework on a competitive hierarchical phrase-based system.
- I 2 R's Machine Translation System for IWSLT 20095
The system and approach used by the Institute for Infocomm Research for the IWSLT 2009 spoken language translation evaluation campaign is described, and the system combination framework shows better performance in Chinese-English BTEC task.
- From Curated Data to Scalable Models: Continual Pre-training of Dense and MoE Large Language Models for Tibetan4
This work constructs a 72 GB high-quality Tibetan corpus, the largest to date, and adapt Qwen2.5-7B through balanced multilingual continual pre-training with Tibetan, Chinese, and English, followed by multilingual instruction tuning.
- 4
- LANDeRMT: Detecting and Routing Language-Aware Neurons for Selectively Finetuning LLMs to Machine Translation4
Experimental results demonstrate that the proposed LANDeRMT is very effective in learning translation knowledge, significantly improving translation quality over various strong baselines for multiple language pairs.
- 4
- 4
- Can Large Language Models Learn Translation Robustness from Noisy-Source In-context Demonstrations?4
This work empirically demonstrate that LLMs can learn how to handle noise and translation methods from noisy-source demonstration examples, thereby improving their translation performance on noisy sentences, and finds that increasing the noise ratio appropriately for the noisy-source demonstration examples can enhance the translation robustness of LLMs.
- CBBQ: A Chinese Bias Benchmark Dataset Curated with Human-AI Collaboration for Large Language Models4
- HuaSLIM: Human Attention Motivated Shortcut Learning Identification and Mitigation for Large Language models4
This paper proposes a human attention guided approach to identifying and mitigating short-cut learning, which encourages the LLM-based target model to learn relevant features and utilizes human attention as a supervisory signal to constrain large language models to pay more attention to relevant tokens.
- AAN+: Generalized Average Attention Network for Accelerating Neural Transformer4
A generalized average attention network (AAN+) aiming at speeding up decoding by reducing the dependency from O(m) to O(1) is proposed, finding that the learned self-attention weights in the decoder follow some patterns which can be approximated via a dynamic structure.
- Knowledge-enhanced graph convolutional network for recommendation4
To capture relatedness of inter-item effectively, the attention mechanism is improved to better probe into the correlation of the item’s neighbors, and a new aggregator to enhance feature aggregation is designed.
- AdaST: Dynamically Adapting Encoder States in the Decoder for End-to-End Speech-to-Text Translation4
The benefits of varying acoustic states according to decoder hidden states are shown and an adaptive speech-to-text translation model that is able to dynamically adapt acoustic states in the decoder is proposed that significantly outperforms state-of-the-art neural speech translation models.
- 4
- 4
- Cross-lingual implicit discourse relation recognition with co-training4
A cross-lingual implicit DRR framework that exploits an available English corpus for the Chinese DRR task, using machine translation to generate Chinese instances from a labeled English discourse corpus and training two classifiers in Chinese and English in a co-training way.
- 4
- 4
- 4
- 4
- 4
- Linguistically Annotated Reordering: Evaluation and Analysis4
A comparative analysis is conducted that not only provides the insight into how linguistic knowledge affects phrase movement but also reveals new challenges in phrase reordering.
- EvoSci: A Bio-Inspired Multi-Agent Framework for the Evolution of Scientific Discovery3
Experiments demonstrate that EvoSci significantly outperforms strong baselines in LLM-based structured peer-review and comparative ranking evaluations, achieving the highest overall peer-review score and top ranking, suggesting its superiority in both scientific idea generation and continuous discovery.
- 3
- 3
- IT2ACL Learning Easy-to-Hard Instructions via 2-Phase Automated Curriculum Learning for Large Language Models3
A 2-phase automated curriculum learning guided instruction tuning framework, IT2ACL that learns easy-to-hard instructions for LLMs in a self-adjusting dynamic manner and proposes a loss-driven progress signal for two-phase strategies: instruction prediction gain that decides the instruction level syllabus.
- 3
- Interdisciplinary analysis and optimization of digital photonic devices for meta-photonics3
The proposed concept of meta-photonics transcends the conventional boundaries of disciplinary silos, demonstrating the transformative potential of interdisciplinary fusion.
- 3
- Analyzing Word Translation of Transformer Layers3
This paper proposes approaches to analyze the translation performed in encoder / decoder layers of the Transformer and reveals that the translation starts at the very beginning of the "encoding" (specifically at the source word embedding layer), and shows how translation evolves during the forward computation of layers.
- Improving Chinese-English Neural Machine Translation with Detected Usages of Function Words3
Classifiers to automatically detect usages of common Chinese function words based on Chinese Function usage Knowledge Base (CFKB) are developed and a function word usage embedding model is proposed to incorporate detection results into neural machine translation (NMT).
- Context-Aware Phrase Representation for Statistical Machine Translation3
This work proposes a context-aware phrase representation learning framework, which extends the bilingually-constrained recursive autoencoder with context modeling component and introduces the bipartite network embedding method to learn better word and topic embedding, which further improve the quality of the phrase representation.
- Learning Event Expressions via Bilingual Structure Projection3
A bilingual structure projection algorithm that explores linguistic divergences between two languages (Chinese and English) and mines new phrases with new syntactic structures, which have been ignored in the previous work are proposed.
- Discriminative reordering model adaptation via structural learning3
This paper proposes a novel adaptive discriminative reordering model (DRM) based on structural learning, which can capture correspondences among reordering features from two different domains, and learns a shared feature representation for cross-domain phrase reordering.
- A Topic-Based Reordering Model for Statistical Machine Translation3
A topic-based reordering model to predict orders for neighboring blocks by capturing topic-sensitive reordering patterns is proposed to predict orders for neighboring blocks by capturing topic-sensitive reordering patterns.
- Backward and trigger-based language models for statistical machine translation3
This article presents a backward language model that augments the conventional forward language model, and a mutual information trigger model which captures long-distance dependencies that go beyond the scope of standard n-gram language models.
- The ICT Statistical Machine Translation Systems for IWSLT 20073
An overview of the ICT statistical machine translation systems for the evaluation campaign of the International Workshop on Spoken Language Translation (IWSLT) 2007 is given, which ranks 2 among the 15 primary results according to the official evaluation results.
- An EBMT system based on word alignment.3
This system is an experiment of examples based approach that uses word alignment information to find and combine fragments in a corpus containing 220 thousand sentence pairs with word alignment.
- 2
- Dense retrieval and reranking for referenced provisions in electric power audit systems2
Experiments are conducted on a Chinese electric power audit text dataset, demonstrating that the proposed retrieval–reranking design provides an effective and practical solution for accurate referenced provision retrieval.
- 2
- Data Mixing for Large Language Models Pretraining: A Survey and Outlook2
This paper formalizes data mixture optimization as a bilevel problem on the probability simplex and clarify the role of data mixing in the pretraining pipeline, and introduces a fine-grained taxonomy that organizes existing methods along two dimensions: static versus dynamic mixing.
- Optimize Quantization for Large Language Models via Progressive Training2
This paper dives into the challenges associated with data selection and training approach, and proposes an iterative training approach that enhances the stability of model quantization at extremely low-bit configuration.
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- Neural Machine Translation with Phrasal Attention2
The proposed phrasal attention framework is simple yet effective, keeping the strength of phrase-based statistical machine translation (SMT) on the source side and able to statistically improve word-level attention-based NMT.
- Machine Translation2
This book will not make you feel bored when reading, so you may not need to get it at once in a day, and machine translation really offers what everybody wants.
- Bilingual Autoencoders with Global Descriptors for Modeling Parallel Sentences2
This paper explores a bilingual autoencoder approach to model parallel sentences by extracting sentence-level global descriptors from word embeddings, and constructing two monolingual autoencoders over these descriptors on the source and target language.
- Semantics, Discourse and Statistical Machine Translation2
This tutorial aims at providing a timely and combined introduction of recent work on SMT with discourse, including a brief review on discourse studies from linguistics and computational viewpoints, discourse research from monolingual to multilingual, discourse-based SMT and a few advanced topics.
- 1
- DVMap: Fine-Grained Pluralistic Value Alignment via High-Consensus Demographic-Value Mapping1
DVMap (High-Consensus Demographic-Value Mapping), a framework for fine-grained pluralistic value alignment, is proposed, demonstrating strong generalization and robustness.
- Beyond Value Benchmarks: Measuring Value-Structure Alignment in Large Language Models via Symmetric Q-Sorts1
This work proposes a symmetric human-LLM evaluation framework, grounded in Q methodology, to measure value-structure alignment, and reveals significant cross-family heterogeneity, model-specific sensitivity to generation stochasticity and localized misalignment.
- 1
- Mix-MoE: Improving Multilingual Machine Translation of Large Language Models Through Mixed MoEs1
Mix-MoE, a mixed Mixture-of-Experts framework designed to train LLMs for multilingual MT, excels in multilingual MT, significantly outperforming existing baselines and showing notable progress in mitigating parameter interference.
- 1
- 1
- FuxiMT: Sparsifying Large Language Models for Chinese-Centric Multilingual Machine Translation1
Experimental results demonstrate that FuxiMT significantly outperforms strong baselines, including state-of-the-art LLMs and machine translation models, particularly under low-resource scenarios, indicating its potential to bridge communication gaps where parallel data are scarce or unavailable.
- TaP: A Taxonomy-Guided Framework for Automated and Scalable Preference Data Generation1
This work uses TaP-generated datasets to perform supervised and preference fine-tuning on multiple LLMs, and demonstrates that LLMs trained on TaP-generated datasets outperform those trained on existing open-source datasets.
- Enhanced spectrum prediction using deep learning models with multi-frequency supplementary inputs1
It is demonstrated that the predictive accuracy of the network can be improved dramatically by simply adding supplementary multi-frequency inputs to the existing dataset in the target spectrum predicting process.
- 1
- 1
- 1
- 1
- GhostRNN: Reducing State Redundancy in RNN with Cheap Operations1
An efficient RNN architecture, GhostRNN, which reduces hidden state redundancy with cheap operations, and observes that partial dimensions of hidden states are similar to the others in trained RNN models, suggesting that redundancy exists in specific RNNs.
- 1
- 1
- Joint Training and Decoding for Multilingual End-to-End Simultaneous Speech Translation1
This paper investigates end- to-end simultaneous speech translation in a one-to-many multilingual setting which is closer to applications in real scenarios and proposes an asynchronous training strategy on the proposed unified decoder architecture.
- 1
- 1
- 1
- Adversarially Improving NMT Robustness to ASR Errors with Confusion Sets1
An adversarial example generation method based on confusion sets that contain words easily confusable with a target word by ASR to conduct adversarial training for NMT models that can achieve improvements on the clean test set.
- CoCoID: Learning Contrastive Representations and Compact Clusters for Semi-Supervised Intent Discovery1
This paper proposes a novel semi-supervised intent discovery framework CoCoID with two essential components: contrastive user utter-ance representation learning and intra-cluster knowledge distillation that is able to learn contrastive representations and compact clusters for intent discovery.
- Learning Hard Retrieval Decoder Attention for Transformers1
An approach to learning a hard retrieval attention where an attention head only attends to one token in the sentence rather than all tokens, which is 1.43 times faster in decoding and preserves translation quality on a wide range of machine translation tasks when used in the decoder self- and cross-attention networks.
- 1
- Neural machine translation with constraints1
An outlook of the future trends in NMT is presented, which includes a series of NMT models proposed for considering a variety of useful information and knowledge constraints, which include variational NMT with constraints of latent variables, NMT advised by statistical machine translation, and N MT with syntactical constraints from the source language.
- Bridging Source and Target Word Embeddings for Neural Machine Translation.1
The proposed bridging models are able to significantly improve quality of both translation and word alignments and a direct link bridging model that directly connects source and target word embeddings so as to minimize their discrepancy is proposed.
- 1
- Improving Translation Selection with Supersenses.1
A supersense tagging method to annotate source words with coarse-grained ontological concepts that can significantly improve translation quality via correctly conveying the meaning of the source language to the target language.
- Adapted competitive learning on continuous semantic space for word sense induction1
This paper learns continuous semantic space representations for the ambiguous instances via recursive context composition, allowing them to capture lexical semantics and syntactic relations simultaneously simultaneously, and adapt rival penalization competitive learning to conduct instances based word sense clustering, allowing it to determine the number of word senses automatically.
- 1
- 1
- Chinese Statistical Parsing with Rich Linguistic Features1
In this model, non-recursive noun and verb phrases are annotated in the Penn Chinese Treebank because of their strong mark of boundaries and the context configuration frame provides a stronger representation of bilexical dependency structures.
- –
- –
- MASS: Deep Research for Social Sciences with Memory-Augmented Social Simulation–
Memory-Augmented Social Simulation (MASS) is proposed, an innovative paradigm that leverages highly realistic and research-oriented social simulations to enhance the creativity and empirical founding of LLMs-generated research.
- AdaDPI: Document-level Translation Adaptive Agent via Dynamic Parametric Internalization–
AdaDPI is proposed, an adaptive agentic framework that shifts the DocMT paradigm from static retrieval to dynamic parametric internalization, and designs a linguistic uncertainty monitor to actively detect critical discourse discontinuities by the model’s epistemic uncertainty.
- –
- –
- –
- SeMi-LoRA: Enhancing Low-Rank Adaptation via Separation and Mixing–
This work proposes SeMi-LoRA (Separation and Mixing LoRA), a novel framework that enables complete high-rank updates while preserving mergeability and consistently outperforms strong PEFT baselines with favorable parameter efficiency.
- ExpertIVS: Sociological Expert Driven Individual Value Simulation in Large Language Models–
ExpertIVS is proposed, a framework employing 14 Sociological Expert Agents to interpret World Values Survey responses through structured professional perspectives, rather than direct responses concatenation, enabling a shift from mere response concatenation to genuine sociological role-playing.
- –
- TENP: Trapezoidal Expert Neuron Pruning For Mixture-of-Experts–
This work proposes TENP, a structured Trapezoidal ExpertNeuron Pruning framework, which identifies and retain important experts, while applying expert neuron pruning (ENP) to less important experts, reserving model parameters in a trapezoidal pattern from shallow to deep layers.
- –
- SARA: Unlocking Multilingual Knowledge in Mixture-of-Experts via Semantically Anchored Routing Alignment–
SARA (Semantically Anchored Routing Alignment), a framework designed to transfer specialized capabilities from high-resource languages as anchors to low-resource languages, is proposed, providing a scalable pathway to enhance multilingual capabilities in sparse architectures.
- –
- –
- –
- –
- –
- Application and exploration of 5G+machine vision in the industrial field–
The development history and advantages of Machine Vision is introduced, the composition and key technologies of Machine Vision are analyzed, the application of Machine Vision integrated with 5G technology in industrial scenarios is studied, and the future development direction of Machine Vision is explored.
- –
- –
- Finedeep: Mitigating Sparse Activation in Dense LLMs via Multi-Layer Fine-Grained Experts–
This work proposes Finedeep, a deep-layered fine-grained expert architecture for dense models, which significantly outperforms traditional dense architectures in terms of perplexity and benchmark performance while maintaining a comparable number of parameters and floating-point operations.
- –
- –
- Entity Recognition in Traditional Chinese Medicine Package Inserts Based on Large Language Models–
Chain-of-Thought reasoning rules are introduced to build a structured decision-making process to guide the model in performing entity recognition for TCM package inserts, effectively enhancing the model's ability to distinguish semantically ambiguous entities.
- –
- CMGBench: Benchmarking Chinese Metaphor Generation for Large Language Models–
CMGBench offers a high-quality dataset comprising 810 examples, 3,354 annotations, and two types of metaphor expressions: direct metaphor expressions and implicit metaphor expressions, the latter of which has received limited attention in previous research.
- N2C2: Nearest Neighbor Enhanced Confidence Calibration for Cross-Lingual In-Context Learning–
A novel approach is proposed, N2C2, which employs a -nearest neighbors augmented classifier for prediction confidence calibration and surpasses fine tuning, prompt tuning and recent state-of-the-art methods in terms of accuracy and calibration errors.
- Very-Long-Distance Dependency Capturing Evaluation via Language Modeling Based on Gender Consistency–
- –
- –
- –
- –
- Meta-RTL: Reinforcement-Based Meta-Transfer Learning for Low-Resource Commonsense Reasoning–
A reinforcement-based multi-source meta-transfer learning framework for low-resource commonsense reasoning that substantially outperforms strong baselines and previous task selection strategies and achieves larger improvements on extremely low-resource settings.
- –
- –
- –
- –
- –
- –
- –
- AI-driven emergence of frequency information non-uniform distribution via THz metasurface spectrum prediction–
It is substantiated that the proposed methodology of simply adding supplementary multi-frequency inputs to the existing dataset during the target spectral prediction process can significantly enhance the predictive accuracy of the network.
- –
- A design of self-learning intelligent infrared remote controller–
A self-learning intelligent infrared remote control system for Android system to realize the management and control of household appliances by mobile terminal.
- –
- –
- –
- Translation Error Detection with Linguistic Features–
This chapter uses a maximum entropy classifier to predict translation errors by integrating word posterior probability feature and linguistic features and proposes to incorporate two groups of linguistic features into error detection: lexical and syntactic features.
- Linguistically Motivated Bracketing–
A linguistically motivated bracketing approach is introduced in this chapter that directly determines whether a source segment can be bracketed and translated as a unit or not.
- Translation Rule Selection with Document-Level Semantic Information–
This chapter presents a framework for translation rule selection based on document-level semantic knowledge, particularly the gist of a document, which incorporates global topic information into translation disambiguation and proposes a topic dissimilarity model to select desirable (less dissimilar) rules.
- –
- Semantically Informed Reordering–
A semantically informed reordering model is introduced that uses the position of a predicate as the reference axis to estimate positions of its associated arguments on the target side and predicts moving directions of arguments relative to their predicates with semantic features.
- BTG-Based SMT–
This chapter systematically introduces BTG-based SMT and describes two extensions to these traditional language models: a backward language model that augments the conventional forward language model, and a mutual information trigger model which captures long-distance dependencies that go beyond the scope of standard \(n\)-gram language models.
- Syntactically Annotated Reordering–
This chapter presents a syntactically annotated reordering approach for SMT, and introduces a syntax-based reordering analysis method and conducts a thorough study on the impacts of syntactic knowledge on phrase reordering with this 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 2026-10-12. 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.