Despite doubts on data quality, instruction synthesis has been widely applied into instruction tuning (IT) of LLMs as an economic and rapid alternative. Recent endeavors focus on improving data quality for synthesized instruction pairs in English and have facilitated IT of English-centric LLMs. However, data quality issues in multilingual synthesized instruction pairs are even more severe, since the common synthesizing practice is to translate English synthesized data into other languages using machine translation (MT). Besides the known content errors in these English synthesized data, multilingual synthesized instruction data are further exposed to defects introduced by MT and face insufficient localization of the target languages, leading to cultural inequality in trained LLMs. In this paper, we propose MIDB, a Multilingual Instruction Data Booster to automatically address the quality issues in multilingual synthesized data. MIDB is trained on around 36.8k revision examples across 16 languages by human linguistic experts, thereby can boost the low-quality data by addressing content errors and MT defects, and improving localization in these synthesized data. Both automatic and human evaluation indicate that not only MIDB steadily improved instruction data quality in 16 languages, but also the instruction-following and cultural-understanding abilities of multilingual LLMs fine-tuned on MIDB-boosted data were significantly enhanced, suggesting an improved linguistic and cultural equality.

Pairwise evaluation of large language models (LLMs) has become the dominant paradigm for benchmarking open-ended tasks, yet non-transitive preferences—where evaluators prefer A over B, B over C, but C over A—fundamentally undermine ranking reliability. We show that this critical issue stems largely from low-quality data that contains inherently ambiguous preference pairs. To address this challenge, we propose ELSPR, a principled graph-theoretic framework that models pairwise preferences as tournament graphs and systematically identifies problematic training data. ELSPR quantifies non-transitivity through strongly connected components (SCCs) analysis and measures overall preference clarity using a novel normalized directed graph structural entropy metric. Our filtering methodology selectively removes preference data that induce non-transitivity while preserving transitive preferences. Extensive experiments on the AlpacaEval benchmark demonstrate that models fine-tuned on ELSPR-filtered data achieve substantial improvements: a 13.8% reduction in non-transitivity, a 0.088 decrease in structural entropy, and significantly enhanced discriminative power in real-world evaluation systems. Human validation confirms that discarded data exhibit dramatically lower inter-annotator agreement (34.4% vs. 52.6%) and model-human consistency (51.2% vs. 80.6%) compared to cleaned data. These findings establish ELSPR as an effective data self-purification approach for developing more robust, consistent, and human-aligned LLM evaluation systems.

Post-training hybridization of large language models (LLMs) often replaces quadratic selfattention with sliding-window attention (SWA) to reduce KV cache usage and improve latency.Existing hybridization schemes are typically defined either at the layer level (e.g., interleaving) or at the head level via static rankings from local to global.Layer-level schemes ignore that local and global dependencies are routed through heads within the same layer, while static head-level rankings suffer from entanglement: a head's local/global behavior can change after hybridization.We propose BOSCH, Black-box Binary Optimization for Short-context Head Selection, a training-free method that formulates the problem as a Large Neighborhood Search and decomposes it into three subproblems: (i) layer-importance detection via small-budget black-box probes, (ii) adaptive per-layer SWA-ratio assignment based on these sensitivities, and (iii) grouped headlevel optimization within ratio buckets.Extensive experiments on 4 LLMs ranging from 1.7B to 30B parameters, across 4 SWA ratios, show that BOSCH consistently outperforms layerlevel heuristics and 6 strong static head-level methods, with larger gains at higher SWA ratios.Under continual pretraining, BOSCH recover original long-context performance faster and to a higher level.Analysis of the selected heads reveals substantial turnover for BOSCH across different SWA ratios, underscoring the importance of performing head-level selection for each target ratio rather than relying on fixed locality rankings.

Linrui Ma, Chun Hei Lo, Xinyu Wang, Peng Lu, Xihao Yuan, Hanting Chen, Kai Han, Xinghao Chen, Chengjun Zhan, Hanlin xu, Yichun Yin, Lifeng Shang, Feng Wen, Boxing Chen, Yufei Cui. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Multilingual instruction fine-tuning (IFT) empowers large language models to generalize across diverse linguistic and cultural contexts; however, high-quality, systematically curated multilingual IFT datasets remain scarce. To address this gap, we propose M-DaQ (Multilingual Diversity and Quality), a diversity-aware sampling framework that jointly optimizes instruction-response quality and cross-lingual semantic diversity. M-DaQ leverages a fine-tuned Quality Scoring Model alongside a maximal marginal relevance-inspired selection strategy to construct balanced, high-fidelity training data. Furthermore, we present the first systematic investigation of the Superficial Alignment Hypothesis in multilingual settings. Extensive evaluations across 18 languages demonstrate that models trained on M-DaQ-curated data achieve average win rates exceeding 60% against strong baselines on Alpaca-Eval and MT-Bench. Complementary human evaluations corroborate these gains, highlighting significant improvements in cultural relevance, contextual appropriateness, and instruction-following capability. The code are publicly released to facilitate reproducibility and future research.

Deployment of Large Language Models (LLMs) has major computational costs, due to their rapidly expanding size. Compression of LLMs reduces the memory footprint, latency, and energy required for their inference. Post-training Quantization (PTQ) techniques have been developed to compress LLMs while avoiding expensive re-training. Most PTQ approaches formulate the quantization error based on a layer-wise Euclidean loss, ignoring the model output. Then, each layer is calibrated using its layer-wise Hessian to update the weights towards minimizing the quantization error. The Hessian is also used for detecting the most salient weights to quantization. Such PTQ approaches are prone to accuracy drop in low-precision quantization. We propose Output-adaptive Calibration (OAC) to incorporate the model output in the calibration process. We formulate the quantization error based on the distortion of the output cross-entropy loss. OAC approximates the output-adaptive Hessian for each layer under reasonable assumptions to reduce the computational complexity. The output-adaptive Hessians are used to update the weight matrices and detect the salient weights towards maintaining the model output. Our proposed method outperforms the state-of-the-art baselines such as SpQR and BiLLM, especially, at extreme low-precision (2-bit and binary) quantization.

In recent years, ransomware has emerged as a formidable data security threat, causing significant data privacy breaches that inflict substantial financial, reputational, and operational damages on society. Many studies employ dynamic feature analysis for ransomware detection. However, these methods utilize neither the internal semantic information (semantic information inherent in the features), nor external semantics (the wealth of existing knowledge and expert experience with regard to ransomware detection). Moreover, conventional methods rely on training data from known ransomware families, while zero-day ransomware often has unknown data distribution patterns, posing detection challenges. In this paper, we propose a Semantics-based Ransomware Detection and family Classification (SRDC) framework that can utilize both internal and external semantics of software. To bolster semantic analysis in zero-day attacks, we also design a procedure called LLM-assisted task-adaptive pre-training (LATAP). In LATAP, ransomware semantics from human experts and LLMs are employed to pre-train the detection model (GPT-2). By fully utilizing semantics, the proposed SRDC framework outperforms the SOTA methods by 12.15% for ransomware family classification tasks, and by 4.03% for zero-day ransomware detection tasks. SRDC also exhibits excellent data efficiency, requiring only two ransom families for training, which is only 35% of the data required by existing methods, to achieve a 90%+ accuracy of zero-day ransomware detection in nine unseen ransom families.

Log analysis represents a critical sub-domain within AI applications that facilitates automatic approaches to fault and error management of large-scaled software systems, saving labors of traditional manual methods. While existing solutions using large language models (LLMs) show promise, they are limited by a significant domain gap between natural and log languages (the latter contains rich domain-specific tokens such as status codes, IP addresses, resource pathes), which restricts their effectiveness in real-world applications. However, directly adapting general-purpose LLMs to log analysis using raw logs may degrade their performance due to inconsistent token distribution. In this paper, we present a domain adaptation approach that addresses these limitations by integrating interpretable domain knowledge into open-source LLMs through continual pre-training (CPT), which bridges this domain gap by adapting LLMs on interpretable natural texts with log knowledge (instead of raw logs) to reduce distribution discrepancy. To achieve this, we developed NLPLog, a comprehensive dataset containing over 250,000 question-answer pairs on log-related knowledge. Our resulting model, SuperLog, achieves the best performance across four log analysis tasks, with an average accuracy improvement of 12.01% over the second-best model. Ablation study also suggests advantages of domain adaption using interpretable log knowledge over using raw logs.

Large Language Models (LLMs) have demonstrated remarkable performance across various natural language processing (NLP) tasks. However, their deployment is challenging due to the substantial computational resources required. Power-of-two (PoT) quantization is a general tool to counteract this difficulty. Albeit previous works on PoT quantization can be efficiently dequantized on CPUs using fixed-point addition, it showed less effectiveness on GPUs. The reason is entanglement of the sign bit and sequential bit manipulations needed for dequantization. We propose a novel POT quantization framework for LLM weights that (i) outperforms state-of-the-art accuracy in extremely low-precision number formats, and (ii) enables faster inference through more efficient dequantization. To maintain the accuracy of the quantized model, we introduce a two-step post-training algorithm: (i) initialize the quantization scales with a robust starting point, and (ii) refine these scales using a minimal calibration set. The performance of our PoT post-training algorithm surpasses the current state-of-the-art in integer quantization, particularly at low precisions such as 2- and 3-bit formats. Our PoT quantization accelerates the dequantization step required for the floating point inference and leads to 3.67× speed up on a NVIDIA V100, and 1.63× on a NVIDIA RTX 4090, compared to uniform integer dequantization.

Benyamin Jamialahmadi, Parsa Kavehzadeh, Mehdi Rezagholizadeh, Parsa Farinneya, Hossein Rajabzadeh, Aref Jafari, Boxing Chen, Marzieh S. Tahaei. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Softmax self-attention often assigns disproportionate weight to semantically uninformative tokens such as special tokens and punctuation, a phenomenon known as attention noise.While recent methods like Cog Attention and the Differential Transformer have addressed this by introducing negative attention scores, they risk discarding useful information.In this paper, we propose the Integral Transformer, a novel selfattention mechanism that denoises attention by integrating signals sampled from the logit distribution.Our approach mitigates noise while preserving the contributions of special tokens critical for model performance.Extensive experiments demonstrate that our model outperforms vanilla, Cog, and Differential attention variants on well-established knowledge and reasoning language benchmarks.Moreover, our analysis reveals that employing vanilla self-attention in the lower Transformer layers enhances performance and that the Integral Transformer effectively balances attention distributions and reduces rank collapse in upper layers.

Yilun Liu, Minggui He, Feiyu Yao, Yuhe Ji, Shimin Tao, Jingzhou Du, Justin Li, Jian Gao, Zhang Li, Hao Yang, Boxing Chen, Osamu Yoshie. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Over recent years, the Transformer has become a fundamental building block for sequence modeling architectures. Yet at its core is the use of self-attention, whose memory and computational cost grow quadratically with the sequence length $N$, rendering it prohibitively expensive for long sequences. A promising approach is top-$k$ attention, which selects only the $k$ most relevant tokens and achieves performance comparable to vanilla self-attention while significantly reducing space and computational demands. However, causal masks require the current query token to only attend to past tokens, preventing existing top-$k$ attention methods from efficiently searching for the most relevant tokens in parallel, thereby limiting training efficiency. In this work, we propose ZETA, leveraging Z-Order Curves for Efficient Top-k Attention, to enable parallel querying of past tokens for entire sequences. We first theoretically show that the choice of key and query dimensions involves a trade-off between the curse of dimensionality and the preservation of relative distances after projection. In light of this insight, we propose reducing the dimensionality of keys and queries in contrast to values and further leveraging Z-order curves to map low-dimensional keys and queries into one-dimensional space, which permits parallel sorting, thereby largely improving the efficiency for top-$k$ token selection. Experimental results demonstrate that ZETA~matches the performance of standard attention on synthetic tasks Associative Recall and outperforms attention and its variants on Long-Range Arena and WikiText-103 language modeling.

The function-calling agent is obtained by performing agent tuning to the large language model (LLM) on function-calling dataset. However, even state-of-the-art datasets (e.g., xlam-function-calling-60k datasets) still contain numerous misleading examples of low-quality data, wasting significant computational resources and result in an unnecessary carbon footprint. Furthermore, such inductive bad data negatively impacts the performance of the agent. In this paper, we propose a set of scoring criteria specifically tailored to evaluate function-calling data and use these criteria to develop a data filtering framework. By applying this framework to filter out low-quality data, we fine-tuned SuperFC, which demonstrates substantial improvements in both sustainability and performance. The SuperFC-7B training process reduced training time from 455 minutes to 85 minutes, resulting in a 80.02% reduction in carbon footprint. Simultaneously, fine-tuning on high-quality data subsets led to performance improvements of up to 3.68%. Additionally, we provide an in-depth analysis of the causes behind the low quality of synthetic function-calling data, offering valuable insights for future data synthesis in this domain. We have also released a high-quality function-calling dataset, available at: https://github.com/Zire-Young/SuperFC

Peng Lu, Ivan Kobyzev, Mehdi Rezagholizadeh, Boxing Chen, Philippe Langlais. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

The quadratic complexity of the attention mechanism in Transformer models has motivated the development of alternative architectures with sub-quadratic scaling, such as state-space models. Among these, Mamba has emerged as a leading architecture, achieving state-of-the-art results across a range of language modeling tasks. However, Mamba’s performance significantly deteriorates when applied to contexts longer than those seen during pre-training, revealing a sharp sensitivity to context length extension. Through detailed analysis, we attribute this limitation to the out-of-distribution behavior of its state-space dynamics, particularly within the parameterization of the state transition matrix $A$. Unlike recent works which attribute this sensitivity to the vanished accumulation of discretization time steps, $\exp(-\sum_{t=1}^N{\Delta}_t)$, we establish a connection between state convergence behavior as the input length approaches infinity and the spectrum of the transition matrix $A$, offering a well-founded explanation of its role in length extension. Next, to overcome this challenge, we propose an approach that applies spectrum scaling to pre-trained Mamba models to enable robust long-context generalization by selectively modulating the spectrum of $A$ matrices in each layer. We show that this can significantly improve performance in settings where simply modulating ${\Delta}_t$ fails, validating our insights and providing avenues for better length generalization of state-space models with structured transition matrices.

Mohammad Dehghan, Mohammad Alomrani, Sunyam Bagga, David Alfonso-Hermelo, Khalil Bibi, Abbas Ghaddar, Yingxue Zhang, Xiaoguang Li, Jianye Hao, Qun Liu, Jimmy Lin, Boxing Chen, Prasanna Parthasarathi, Mahdi Biparva, Mehdi Rezagholizadeh. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

Yuan Ge, Yilun Liu, Chi Hu, Weibin Meng, Shimin Tao, Xiaofeng Zhao, Mahong Xia, Zhang Li, Boxing Chen, Hao Yang, Bei Li, Tong Xiao, JingBo Zhu. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.

In this paper, we study how open-source large language models (LLMs) can be effectively deployed for improving query rewriting in conversational search, especially for ambiguous queries.We introduce CHIQ, a two-step method that leverages the capabilities of LLMs to resolve ambiguities in the conversation history before query rewriting.This approach contrasts with prior studies that predominantly use closed-source LLMs to directly generate search queries from conversation history.We demonstrate on five well-established benchmarks that CHIQ leads to state-of-the-art results across most settings, showing highly competitive performances with systems leveraging closed-source LLMs.Our study provides a first step towards leveraging open-source LLMs in conversational search, as a competitive alternative to the prevailing reliance on commercial LLMs for query rewriting.

Cross-lingual information retrieval (CLIR) continues to be an actively studied topic in information retrieval (IR), and there have been consistent efforts in curating test collections to support its research. However, there is a lack of high-quality human-annotated CLIR resources for African languages: the few existing collections are mostly curated synthetically or from sources with limited corpora for these languages. We present CIRAL, a test collection for cross-lingual retrieval with English queries and passages in four African languages: Hausa, Somali, Swahili, and Yoruba. CIRAL's corpora are obtained from Indigenous African websites and consist of a total of over 2.5 million passages. We gathered over 1,600 queries and 30k high-quality binary relevance judgments annotated by native speakers of the languages. Additional pools were also obtained at CIRAL's shared task, which was hosted at the Forum for Information Retrieval Evaluation 2023 to encourage community participation in CLIR for African languages. We describe the design and curation process of our test collection and provide reproducible baselines that demonstrate CIRAL's utility in evaluating the effectiveness of systems. CIRAL is available at https://github.com/ciralproject/ciral.

Kexin Yang, Dayiheng Liu, Wenqiang Lei, Baosong Yang, Mingfeng Xue, Boxing Chen, Jun Xie. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

A popular approach to streaming speech translation is to employ a single offline model with a wait-k policy to support different latency requirements, which is simpler than training multiple online models with different latency constraints.However, there is a mismatch problem in using a model trained with complete utterances for streaming inference with partial input.We demonstrate that speech representations extracted at the end of a streaming input are significantly different from those extracted from a complete utterance.To address this issue, we propose a new approach called Future-Aware Streaming Translation (FAST) that adapts an offline ST model for streaming input.FAST includes a Future-Aware Inference (FAI) strategy that incorporates future context through a trainable masked embedding, and a Future-Aware Distillation (FAD) framework that transfers future context from an approximation of full speech to streaming input.Our experiments on the MuST-C EnDe, EnEs, and EnFr benchmarks show that FAST achieves better trade-offs between translation quality and latency than strong baselines.Extensive analyses suggest that our methods effectively alleviate the aforementioned mismatch problem between offline training and online inference.1

Self-supervised speech pre-training enables deep neural network models to capture meaningful and disentangled factors from raw waveform signals. The learned universal speech representations can then be used across numerous down-stream tasks. These representations, however, are sensitive to distribution shifts caused by environmental factors, such as noise and/or room reverberation. Their large sizes, in turn, make them unfeasible for edge applications. In this work, we propose a knowledge distillation methodology termed RobustDistiller which compresses universal representations while making them more robust against environmental artifacts via a multi-task learning objective. The proposed layer-wise distillation recipe is evaluated on top of three well-established universal representations, as well as with three downstream tasks. Experimental results show the proposed methodology applied on top of the WavLM Base+ teacher model outperforming all other benchmarks across noise types and levels, as well as reverberation times. Oftentimes, the obtained results with the student model (24M parameters) achieved results inline with those of the teacher model (95M).

As pre-trained models automate many code intel-ligence tasks, a widely used paradigm is to fine-tune a model on the task dataset for each programming language. A recent study reported that multilingual fine-tuning benefits a range of tasks and models. However, we find that multilingual fine-tuning leads to performance degradation on recent models UniXcoder and CodeT5. To alleviate the potentially catastrophic forgetting issue in multilingual models, we fix all pre-trained model parameters, insert the parameter-efficient structure adapter, and fine-tune it. Updating only 0.6% of the overall parameters compared to full-model fine-tuning for each programming language, adapter tuning yields consistent improvements on code search and sum-marization tasks, achieving state-of-the-art results. In addition, we experimentally show its effectiveness in cross-lingual and low-resource scenarios. Multilingual fine-tuning with 200 samples per programming language approaches the results fine-tuned with the entire dataset on code summarization. Our experiments on three probing tasks show that adapter tuning significantly outperforms full-model fine-tuning and effectively overcomes catastrophic forgetting.

Artificial neural networks have shown promising results in a variety of natural language understanding (NLU) tasks. Despite their successes, conventional neural-based NLU models are criticized for high energy consumption, making them laborious to be widely applied in low-power electronics, such as smartphones and intelligent terminals. In this paper, we introduce a potential direction to alleviate this bottleneck by proposing a spiking encoder. The core of our model is bi-directional spiking neural network (SNN) which transforms numeric values into discrete spiking signals and replaces massive multiplications with much cheaper additive operations. We examine our model on sentiment classification and machine translation tasks. Experimental results reveal that our model achieves comparable classification and translation accuracy to advancedTransformerbaseline, whereas significantly reduces the required computational energy to 0.82%.

End-to-end speech-to-text translation (E2E-ST) is becoming increasingly popular due to the potential of its less error propagation, lower latency, and fewer parameters. Given the triplet training corpus〈speech, transcription, translation〉, the conventional high-quality E2E-ST system leverages the〈speech, transcription〉pair to pre-train the model and then utilizes the〈speech, translation〉pair to optimize it further. However, this process only involves two-tuple data at each stage, and this loose coupling fails to fully exploit the association between triplet data. In this paper, we attempt to model the joint probability of transcription and translation based on the speech input to directly leverage such triplet data. Based on that, we propose a novel regularization method for model training to improve the agreement of dual-path decomposition within triplet data, which should be equal in theory. To achieve this goal, we introduce two Kullback-Leibler divergence regularization terms into the model training objective to reduce the mismatch between output probabilities of dual-path. Then the well-trained model can be naturally transformed as the E2E-ST models by a pre-defined early stop tag. Experiments on the MuST-C benchmark demonstrate that our proposed approach significantly outperforms state-of-the-art E2E-ST baselines on all 8 language pairs while achieving better performance in the automatic speech recognition task.

Yu Wan, Dayiheng Liu, Baosong Yang, Haibo Zhang, Boxing Chen, Derek Wong, Lidia Chao. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2022.

k-Nearest-Neighbor Machine Translation (kNN-MT) has been recently proposed as a non-parametric solution for domain adaptation in neural machine translation (NMT).It aims to alleviate the performance degradation of advanced MT systems in translating out-ofdomain sentences by coordinating with an additional token-level feature-based retrieval module constructed from in-domain data.Previous studies (Khandelwal et al., 2021; Zheng et al., 2021a) have already demonstrated that non-parametric NMT is even superior to models fine-tuned on out-of-domain data.In spite of this success, kNN retrieval is at the expense of high latency, in particular for large datastores.To make it practical, in this paper, we explore a more efficient kNN-MT and propose to use clustering to improve the retrieval efficiency.Concretely, we first propose a cluster-based Compact Network for feature reduction in a contrastive learning manner to compress context features into 90+% lower dimensional vectors.We then suggest a cluster-based pruning solution to filter out 10%~40% redundant nodes in large datastores while retaining translation quality.Our proposed methods achieve better or comparable performance while reducing up to 57% inference latency against the advanced non-parametric MT model on several machine translation benchmarks.Experimental results indicate that the proposed methods maintain the most useful information of the original datastore and the Compact Network shows good generalization on unseen domains.Codes are available at https: //github.com/tjunlp-lab/PCKMT.Datastore distribution C-II("bank") C-III("bank") C-VI("bill") Compact Layer Compact Layer centroid ✖ Datastore distribution C-VI("bill") C-II("bank") C-I("sell") C-III("bank") Datastore distribution C-VI("bill") C-II("bank") C-I("sell") C-III("bank") Join Repartition Prune C-III C-II A pruning example w.r.t."bank" ✔ C-I("sell")

Few-shot abstractive summarization has become a challenging task in natural language generation. To support it, we developed a novel soft prompts architecture coupled with a prompt pre-training plus prompt fine-tuning paradigm, which is effective and tunes only extremely light parameters. To meet the structure of the generation models, the soft prompts comprise continuous input embeddings across an encoder and a decoder. Importantly, a new inner-prompt placed in the text is introduced to capture document-level information. The aim is to devote attention to understanding the document that better prompts the model to generate document-related content. In the training process, the prompt pre-training with self-supervised pseudo-data firstly teaches the model basic summarizing capability. Then, with few-shot examples, only the designed lightweight soft prompts are fine-tuned. Experimental results on the CNN/DailyMail and XSum datasets show that our method, with only 0.1% of the parameters, outperforms full-model tuning where all model parameters are tuned. It also surpasses Prompt Tuning by a large margin and delivers competitive results against Prefix-Tuning with 3% of the parameters.

The end-to-end speech translation (E2E-ST) has received increasing attention due to the potential of its less error propagation, lower latency and fewer parameters.However, the effectiveness of neural-based approaches to this task is severely limited by the available training corpus, especially for domain adaptation where in-domain triplet data is scarce or nonexistent.In this paper, we propose a novel nonparametric method that leverages in-domain text translation corpus to achieve domain adaptation for E2E-ST systems.To this end, we first incorporate an additional encoder into the pre-trained E2E-ST model to realize text translation modeling, based on which the decoder's output representations for text and speech translation tasks are unified by reducing the correspondent representation mismatch in available triplet training data.During domain adaptation, a k-nearest-neighbor (kNN) classifier is introduced to produce the final translation distribution using the external datastore built by the domain-specific text translation corpus, while the universal output representation is adopted to perform a similarity search.Experiments on the Europarl-ST benchmark demonstrate that when in-domain text translation data is involved only, our proposed approach significantly improves baseline by 12.82 BLEU on average in all translation directions, even outperforming the strong in-domain fine-tuning strategy.

End-to-end Speech Translation (ST) aims at translating the source language speech into target language text without generating the intermediate transcriptions.However, the training of end-to-end methods relies on parallel ST data, which are difficult and expensive to obtain.Fortunately, the supervised data for automatic speech recognition (ASR) and machine translation (MT) are usually more accessible, making zero-shot speech translation a potential direction.Existing zero-shot methods fail to align the two modalities of speech and text into a shared semantic space, resulting in much worse performance compared to the supervised ST methods.In order to enable zero-shot ST, we propose a novel Discrete Cross-Modal Alignment (DCMA) method that employs a shared discrete vocabulary space to accommodate and match both modalities of speech and text.Specifically, we introduce a vector quantization module to discretize the continuous representations of speech and text into a finite set of virtual tokens, and use ASR data to map corresponding speech and text to the same virtual token in a shared codebook.This way, source language speech can be embedded in the same semantic space as the source language text, which can be then transformed into target language text with an MT module.Experiments on multiple language pairs demonstrate that our zero-shot ST method significantly improves the SOTA, and even performs on par with the strong supervised ST baselines 1 .

Document-level translation models are still far from perfect. Most existing document-level neural machine translation (NMT) models leverage a fixed number of the previous or all global sentences to handle the context-independent problem in standard NMT. However, the translating of each source sentence benefits from various sizes of context. And study shows that inappropriate redundant context will increase model burden but not improve the translation performance. This work introduces a data-adaptive method that enables the model to adopt the necessary and helpful context. Specifically, we introduce a light predictor into two document-level translation models to select the explicit context. Experiments demonstrate the proposed approach can significantly improve the performance over the previous methods with a gain up to 1.99 BLEU points.

Cross-Lingual Summarization (CLS) is a task that extracts important information from a source document and summarizes it into a summary in another language. It is a challenging task that requires a system to understand, summarize, and translate at the same time, making it highly related to Monolingual Summarization (MS) and Machine Translation (MT). In practice, the training resources for Machine Translation are far more than that for cross-lingual and monolingual summarization. Thus incorporating the Machine Translation corpus into CLS would be beneficial for its performance. However, the present work only leverages a simple multi-task framework to bring Machine Translation in, lacking deeper exploration.

In this paper, we propose to formulate the task-oriented dialogue system as the purely natural language generation task, so as to fully leverage the large-scale pre-trained models like GPT-2 and simplify complicated delexicalization prepossessing. However, directly applying this method heavily suffers from the dialogue entity inconsistency caused by the removal of delexicalized tokens, as well as the catastrophic forgetting problem of the pre-trained model during fine-tuning, leading to unsatisfactory performance. To alleviate these problems, we design a novel GPT-Adapter-CopyNet network, which incorporates the lightweight adapter and CopyNet modules into GPT-2 to achieve better performance on transfer learning and dialogue entity generation. Experimental results conducted on the DSTC8 Track 1 benchmark and MultiWOZ dataset demonstrate that our proposed approach significantly outperforms baseline models with a remarkable performance on automatic and human evaluations.

Abstract Short texts (STs) present in a variety of scenarios, including query, dialog, and entity names. Most of the exciting studies in neural machine translation (NMT) are focused on tackling open problems concerning long sentences rather than short ones. The intuition behind is that, with respect to human learning and processing, short sequences are generally regarded as easy examples. In this article, we first dispel this speculation via conducting preliminary experiments, showing that the conventional state-of-the-art NMT approach, namely, Transformer (Vaswani et al. 2017), still suffers from over-translation and mistranslation errors over STs. After empirically investigating the rationale behind this, we summarize two challenges in NMT for STs associated with translation error types above, respectively: (1) the imbalanced length distribution in training set intensifies model inference calibration over STs, leading to more over-translation cases on STs; and (2) the lack of contextual information forces NMT to have higher data uncertainty on short sentences, and thus NMT model is troubled by considerable mistranslation errors. Some existing approaches, like balancing data distribution for training (e.g., data upsampling) and complementing contextual information (e.g., introducing translation memory) can alleviate the translation issues in NMT for STs. We encourage researchers to investigate other challenges in NMT for STs, thus reducing ST translation errors and enhancing translation quality.

Guangsheng Bao, Yue Zhang, Zhiyang Teng, Boxing Chen, Weihua Luo. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Linqing Chen, Junhui Li, Zhengxian Gong, Boxing Chen, Weihua Luo, Min Zhang, Guodong Zhou. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

A currently popular research area in end-toend speech translation is the use of knowledge distillation from a machine translation (MT) task to improve the speech translation (ST) task.However, such scenario obviously allows only a one way transfer, limiting the overall effectiveness of the approach by the performance of the pre-trained teacher model.Therefore, we pose that in this respect knowledge distillationbased approaches are sub-optimal.We propose an alternative-a trainable mutual-learning scenario, where the MT and ST models are collaboratively trained and are considered as peers, rather than teacher/student.This allows us to improve the performance of end-to-end ST more effectively than with a teacher-student paradigm.As a side benefit, performance of the MT model also improves.Experimental results show that in our mutual-learning scenario, models can effectively utilise the auxiliary information from peer models and achieve compelling results on MuST-C datasets.

Document context-aware machine translation remains challenging due to the lack of large-scale document parallel corpora. To make full use of source-side monolingual documents for context-aware NMT, we propose a Pre-training approach with Global Context (PGC). In particular, we first propose a novel self-supervised pre-training task, which contains two training objectives: (1) reconstructing the original sentence from a corrupted version; (2) generating a gap sentence from its left and right neighbouring sentences. Then we design a universal model for PGC which consists of a global context encoder, a sentence encoder and a decoder, with similar architecture to typical context-aware NMT models. We evaluate the effectiveness and generality of our pre-trained PGC model by adapting it to various downstream context-aware NMT models. Detailed experimentation on four different translation tasks demonstrates that our PGC approach significantly improves the translation performance of context-aware NMT. For example, based on the state-of-the-art SAN model, we achieve an averaged improvement of 1.85 BLEU scores and 1.59 Meteor scores on the four translation tasks.

Paraphrase generation plays key roles in NLP tasks such as question answering, machine translation, and information retrieval. In this paper, we propose a novel framework for paraphrase generation. It simultaneously decodes the output sentence using a pretrained wordset-to-sequence model and a round-trip translation model. We evaluate this framework on Quora, WikiAnswers, MSCOCO and Twitter, and show its advantage over previous state-of-the-art unsupervised methods and distantly-supervised methods by significant margins on all datasets. For Quora and WikiAnswers, our framework even performs better than some strongly supervised methods with domain adaptation. Further, we show that the generated paraphrases can be used to augment the training data for machine translation to achieve substantial improvements.

Neural machine translation (NMT) models are data-driven and require large-scale training corpus.In practical applications, NMT models are usually trained on a general domain corpus and then fine-tuned by continuing training on the in-domain corpus.However, this bears the risk of catastrophic forgetting that the performance on the general domain is decreased drastically.In this work, we propose a new continual learning framework for NMT models.We consider a scenario where the training is comprised of multiple stages and propose a dynamic knowledge distillation technique to alleviate the problem of catastrophic forgetting systematically.We also find that the bias exists in the output linear projection when fine-tuning on the in-domain corpus, and propose a bias-correction module to eliminate the bias.We conduct experiments on three representative settings of NMT application.Experimental results show that the proposed method achieves superior performance compared to baseline models in all settings.1

Pengcheng Yang, Pei Zhang, Boxing Chen, Jun Xie, Weihua Luo. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021.

Large-scale pre-trained language models (e.g., BERT) have attracted great attention in recent years. It is straightforward to fine-tune them on natural language understanding tasks such as text classification, however, effectively and efficiently incorporating them into natural language generation tasks such as neural machine translation remains a challenging problem. In this paper, we integrate two pre-trained BERT models from the source and target language domains into a sequence-to-sequence model by introducing light-weight adapter modules. The adapters are inserted between BERT layers and tuned on downstream tasks, while the parameters of BERT models are fixed during fine-tuning. As pre-trained language models are usually very deep, inserting adapters into all layers will result in a considerable scale of new parameters. To deal with this problem, we introduce latent variables to decide whether using adapters or not in each layer, which are learned during fine-tuning. In this way, the model is able to automatically determine which adapters to use, therefore hugely promoting the parameter efficiency and decoding speed. We evaluate the proposed framework on various neural machine translation tasks. Equipped with parallel sequence decoding, our model consistently outperforms autoregressive baselines while reducing the inference latency by half. With automatic adapter selection, the proposed model further achieves 20% speedup while still outperforming autoregressive baselines. When applied to autoregressive decoding, the proposed model can also achieve comparable performance with the state-of-the-art baseline models.

Transfer learning between different language pairs has shown its effectiveness for Neural Machine Translation (NMT) in low-resource scenario. However, existing transfer methods involving a common target language are far from success in the extreme scenario of zero-shot translation, due to the language space mismatch problem between transferor (the parent model) and transferee (the child model) on the source side. To address this challenge, we propose an effective transfer learning approach based on cross-lingual pre-training. Our key idea is to make all source languages share the same feature space and thus enable a smooth transition for zero-shot translation. To this end, we introduce one monolingual pre-training method and two bilingual pre-training methods to obtain a universal encoder for different languages. Once the universal encoder is constructed, the parent model built on such encoder is trained with large-scale annotated data and then directly applied in zero-shot translation scenario. Experiments on two public datasets show that our approach significantly outperforms strong pivot-based baseline and various multilingual NMT approaches.

Multi-modal machine translation aims at translating the source sentence into a different language in the presence of the paired image. Previous work suggests that additional visual information only provides dispensable help to translation, which is needed in several very special cases such as translating ambiguous words. To make better use of visual information, this work presents visual agreement regularized training. The proposed approach jointly trains the source-to-target and target-to-source translation models and encourages them to share the same focus on the visual information when generating semantically equivalent visual words (e.g. “ball” in English and “ballon” in French). Besides, a simple yet effective multi-head co-attention model is also introduced to capture interactions between visual and textual features. The results show that our approaches can outperform competitive baselines by a large margin on the Multi30k dataset. Further analysis demonstrates that the proposed regularized training can effectively improve the agreement of attention on the image, leading to better use of visual information.

International audience

Query translation (QT) serves as a critical factor in successful cross-lingual information retrieval (CLIR).Due to the lack of parallel query samples, neural-based QT models are usually optimized with synthetic data which are derived from large-scale monolingual queries.Nevertheless, such kind of pseudo corpus is mostly produced by a general-domain translation model, making it be insufficient to guide the learning of QT model.In this paper, we extend the data augmentation with a domain transfer procedure, thus to revise synthetic candidates to search-aware examples.Specifically, the domain transfer model is built upon advanced Transformer, in which layer coordination and mixed attention are exploited to speed up the refining process and leverage parameters from a pre-trained cross-lingual language model.In order to examine the effectiveness of the proposed method, we collected French-to-English and Spanish-to-English QT test sets, each of which consists of 10,000 parallel query pairs with careful manual-checking.Qualitative and quantitative analyses reveal that our model significantly outperforms strong baselines and the related domain transfer methods on both translation quality and retrieval accuracy.1

Recent studies have proven that the training of neural machine translation (NMT) can be facilitated by mimicking the learning process of humans.Nevertheless, achievements of such kind of curriculum learning rely on the quality of artificial schedule drawn up with the handcrafted features, e.g.sentence length or word rarity.We ameliorate this procedure with a more flexible manner by proposing self-paced learning, where NMT model is allowed to 1) automatically quantify the learning confidence over training examples; and 2) flexibly govern its learning via regulating the loss in each iteration step.Experimental results over multiple translation tasks demonstrate that the proposed model yields better performance than strong baselines and those models trained with human-designed curricula on both translation quality and convergence speed. 1

In this paper, we focus on the domain-specific translation with low resources, where indomain parallel corpora are scarce or nonexistent.One common and effective strategy for this case is exploiting in-domain monolingual data with the back-translation method.However, the synthetic parallel data is very noisy because they are generated by imperfect out-of-domain systems, resulting in the poor performance of domain adaptation.To address this issue, we propose a novel iterative domain-repaired back-translation framework, which introduces the Domain-Repair (DR) model to refine translations in synthetic bilingual data.To this end, we construct corresponding data for the DR model training by round-trip translating the monolingual sentences, and then design the unified training framework to optimize paired DR and NMT models jointly.Experiments on adapting NMT models between specific domains and from the general domain to specific domains demonstrate the effectiveness of our proposed approach, achieving 15.79 and 4.47 BLEU improvements on average over unadapted models and back-translation. 1

Many document-level neural machine translation (NMT) systems have explored the utility of context-aware architecture, usually requiring an increasing number of parameters and computational complexity.However, few attention is paid to the baseline model.In this paper, we research extensively the pros and cons of the standard transformer in document-level translation, and find that the auto-regressive property can simultaneously bring both the advantage of the consistency and the disadvantage of error accumulation.Therefore, we propose a surprisingly simple long-short term masking self-attention on top of the standard transformer to both effectively capture the long-range dependence and reduce the propagation of errors.We examine our approach on the two publicly available document-level datasets.We can achieve a strong result in BLEU and capture discourse phenomena.

The performances of automatic speech recognition (ASR) systems are usually evaluated by the metric word error rate (WER) when the manually transcribed data are provided, which are, however, expensively available in the real scenario.In addition, the empirical distribution of WER for most ASR systems usually tends to put a significant mass near zero, making it difficult to simulate with a single continuous distribution.In order to address the two issues of ASR quality estimation (QE), we propose a novel neural zero-inflated model to predict the WER of the ASR result without transcripts.We design a neural zeroinflated beta regression on top of a bidirectional transformer language model conditional on speech features (speech-BERT).We adopt the pre-training strategy of token level masked language modeling for speech-BERT as well, and further fine-tune with our zero-inflated layer for the mixture of discrete and continuous outputs.The experimental results show that our approach achieves better performance on WER prediction compared with strong baselines.

While large scale pre-trained language models such as BERT have achieved great success on various natural language understanding tasks, how to efficiently and effectively incorporate them into sequence-to-sequence models and the corresponding text generation tasks remains a non-trivial problem. In this paper, we propose to address this problem by taking two different BERT models as the encoder and decoder respectively, and fine-tuning them by introducing simple and lightweight adapter modules, which are inserted between BERT layers and tuned on the task-specific dataset. In this way, we obtain a flexible and efficient model which is able to jointly leverage the information contained in the source-side and target-side BERT models, while bypassing the catastrophic forgetting problem. Each component in the framework can be considered as a plug-in unit, making the framework flexible and task agnostic. Our framework is based on a parallel sequence decoding algorithm named Mask-Predict considering the bi-directional and conditional independent nature of BERT, and can be adapted to traditional autoregressive decoding easily. We conduct extensive experiments on neural machine translation tasks where the proposed method consistently outperforms autoregressive baselines while reducing the inference latency by half, and achieves $36.49$/$33.57$ BLEU scores on IWSLT14 German-English/WMT14 German-English translation. When adapted to autoregressive decoding, the proposed method achieves $30.60$/$43.56$ BLEU scores on WMT14 English-German/English-French translation, on par with the state-of-the-art baseline models.

The performances of machine translation (MT) systems are usually evaluated by the metric BLEU when the golden references are provided. However, in the case of model inference or production deployment, golden references are usually expensively available, such as human annotation with bilingual expertise. In order to address the issue of translation quality estimation (QE) without reference, we propose a general framework for automatic evaluation of the translation output for the QE task in the Conference on Statistical Machine Translation (WMT). We first build a conditional target language model with a novel bidirectional transformer, named neural bilingual expert model, which is pre-trained on large parallel corpora for feature extraction. For QE inference, the bilingual expert model can simultaneously produce the joint latent representation between the source and the translation, and real-valued measurements of possible erroneous tokens based on the prior knowledge learned from parallel data. Subsequently, the features will further be fed into a simple Bi-LSTM predictive model for quality estimation. The experimental results show that our approach achieves the state-of-the-art performance in most public available datasets of WMT 2017/2018 QE task.

Abstractive Sentence Summarization (AS-SUM) targets at grasping the core idea of the source sentence and presenting it as the summary.It is extensively studied using statistical models or neural models based on the large-scale monolingual source-summary parallel corpus.But there is no cross-lingual parallel corpus, whose source sentence language is different to the summary language, to directly train a cross-lingual ASSUM system.We propose to solve this zero-shot problem by using resource-rich monolingual AS-SUM system to teach zero-shot cross-lingual ASSUM system on both summary word generation and attention.This teaching process is along with a back-translation process which simulates source-summary pairs.Experiments on cross-lingual ASSUM task show that our proposed method is significantly better than pipeline baselines and previous works, and greatly enhances the cross-lingual performances closer to the monolingual performances.We release the code and data at https://github.com/KelleyYin/ Cross-lingual-Summarization.

Recent advances in sequence modeling have highlighted the strengths of the transformer architecture, especially in achieving state-of-theart machine translation results.However, depending on the up-stream systems, e.g., speech recognition, or word segmentation, the input to translation system can vary greatly.The goal of this work is to extend the attention mechanism of the transformer to naturally consume the lattice in addition to the traditional sequential input.We first propose a general lattice transformer for speech translation where the input is the output of the automatic speech recognition (ASR) which contains multiple paths and posterior scores.To leverage the extra information from the lattice structure, we develop a novel controllable lattice attention mechanism to obtain latent representations.On the LDC Spanish-English speech translation corpus, our experiments show that lattice transformer generalizes significantly better and outperforms both a transformer baseline and a lattice LSTM.Additionally, we validate our approach on the WMT 2017 Chinese-English translation task with lattice inputs from different BPE segmentations.In this task, we also observe the improvements over strong baselines.

Bilingual Lexicon Induction (BLI) aims at inducing word translations in two distinct languages. The generated bilingual dictionaries via BLI are essential for cross-lingual NLP applications. Most existing methods assume that a mapping matrix can be learned to project the embedding of a word in the source language to that of a word in the target language which shares the same meaning. However, a single matrix may not be able to provide sufficiently large parameter space and to tailor to the semantics of words across different domains and topics due to the complicated nature of linguistic regularities. In this paper, we propose a Soft Piecewise Mapping Model (SPMM). It generates word alignments in two languages by learning multiple mapping matrices with orthogonal constraint. Each matrix encodes the embedding translation knowledge over a distribution of latent topics in the embedding spaces. Such learning problem can be formulated as an extended version of the Wahba's problem, with a closed-form solution derived. To address the limited size of training data for low-resourced languages and emerging domains, an iterative boosting method based on SPMM is used to augment training dictionaries. Experiments conducted on both general and domain-specific corpora show that SPMM is effective and outperforms previous methods.

In this paper, we propose a method which uses semi-supervised convolutional neural networks (CNNs) to select in-domain training data for statistical machine translation. This approach is particularly effective when only tiny amounts of in-domain data are available. The in-domain data and randomly sampled general-domain data are used to train a data selection model with semi-supervised CNN, then this model computes domain relevance scores for all the sentences in the generaldomain data set. The sentence pairs with top scores are selected to train the system. We carry out experiments on 4 language directions with three test domains. Compared with strong baseline systems trained with large amount of data, this method can improve the performance up to 3.1 BLEU. Its performances are significant better than three state-of-the-art language model based data selection methods. We also show that the in-domain data used to train the selection model could be as fewas 100sentences, whichmakesfinegrained topic-dependent translation adaptation possible.

In this paper, we explore bilingual sentiment knowledge for statistical machine translation (SMT). We propose to explicitly model the consistency of sentiment between the source and target side with a lexicon-based approach. The experiments show that the proposed model significantly improves Chinese-to-English NIST translation over a competitive baseline. © 2014 Association for Computational Linguistics.

Spoken Language Translation (SLT) is the research area that focuses on the translation of speech or text between two spoken languages. Phrase-based and syntax-based methods represent the state-of-the-art for statistical machine translation (SMT). The phrase-based method specializes in modeling local reorderings and translations of multiword expressions. The syntax-based method is enhanced by using syntactic knowledge, which can better model long word reorderings, discontinuous phrases, and syntactic structure. In this article, we leverage on the strength of these two methods and propose a strategy based on multiple hypotheses generation in a two-stage framework for spoken language translation. The hypotheses are generated in two stages, namely, decoding and regeneration. In the decoding stage, we apply state-of-the-art, phrase-based, and syntax-based methods to generate basic translation hypotheses. Then in the regeneration stage, much more hypotheses that cannot be captured by the decoding algorithms are produced from the basic hypotheses. We study three regeneration methods: redecoding, n-gram expansion, and confusion network in the second stage. Finally, an additional reranking pass is introduced to select the translation outputs by a linear combination of rescoring models. Experimental results on the Chinese-to-English IWSLT-2006 challenge task of translating the transcription of spontaneous speech show that the proposed mechanism achieves significant improvements over the baseline of about 2.80 BLEU-score.

Workflow management system can offer enterprises a method to implement better management. This paper focuses on existing workflow management system bugs in the collaboration, discusses the combinability between workflow management system and CSCW, puts forward a work-strategy in the workflow collaboration and a system reference model, and designs a CSCW-based workflow management system. This system is based on a knowledge library, uses the rules and knowledge in the knowledge library to realize the workflow collaboration. This system also applies in the area of stock so as to effectively solve the problem that in the current stock system the workflows cannot work in the collaboration.

The integration of grid database is a hot researching field of grid database management nowadays. The middleware of data access and integration then plays an important role in grid database integration. This paper discusses the integration and the access of grid database based on the research and the analysis of OSGA-DAI framework and its internal procedure. We put forward a system framework and model design to extend and improve OGSA-DAI functions. The system presents good performances for the access and the integration of grid database. Finally, combined with the relative key technology and programming skills, we give out a solution of the access and the integration of grid database