In-context learning (ICL) has demonstrated excellent performance across various downstream NLP tasks, especially when synergized with powerful large language models (LLMs). Existing studies evaluate ICL methods primarily based on downstream task performance. This evaluation protocol overlooks the significant cost associated with the demonstration configuration process, i.e., tuning the demonstration as the ICL prompt. However, in this work, we point out that the evaluation protocol leads to unfair comparisons and potentially biased evaluation, because we surprisingly find the correlation between the configuration costs and task performance. Then we call for a two-dimensional evaluation paradigm that considers both of these aspects, facilitating a fairer comparison.Finally, based on our empirical finding that the optimized demonstration on one language model generalizes across language models of different sizes, we introduce a simple yet efficient strategy that can be applied to any ICL method as a plugin, yielding a better trade-off between the two dimensions according to the proposed evaluation paradigm.

Developing effective biomedical retrieval models is important for excelling at knowledge-intensive biomedical tasks but still challenging due to the lack of sufficient publicly annotated biomedical data and computational resources. We present BMRetriever, a series of dense retrievers for enhancing biomedical retrieval via unsupervised pre-training on large biomedical corpora, followed by instruction fine-tuning on a combination of labeled datasets and synthetic pairs. Experiments on 5 biomedical tasks across 11 datasets verify BMRetriever’s efficacy on various biomedical applications. BMRetriever also exhibits strong parameter efficiency, with the 410M variant outperforming baselines up to 11.7 times larger, and the 2B variant matching the performance of models with over 5B parameters. The training data and model checkpoints are released at https://huggingface.co/BMRetriever to ensure transparency, reproducibility, and application to new domains.

Despite their improved capabilities in generation and reasoning, adapting large language models (LLMs) to the biomedical domain remains challenging due to their immense size and privacy concerns. In this study, we propose MedAdapter, a unified post-hoc adapter for test-time adaptation of LLMs towards biomedical applications. Instead of fine-tuning the entire LLM, MedAdapter effectively adapts the original model by fine-tuning only a small BERT-sized adapter to rank candidate solutions generated by LLMs. Experiments on four biomedical tasks across eight datasets demonstrate that MedAdapter effectively adapts both white-box and black-box LLMs in biomedical reasoning, achieving average performance improvements of 18.24% and 10.96%, respectively, without requiring extensive computational resources or sharing data with third parties. MedAdapter also yields enhanced performance when combined with train-time adaptation, highlighting a flexible and complementary solution to existing adaptation methods. Faced with the challenges of balancing model performance, computational resources, and data privacy, MedAdapter provides an efficient, privacy-preserving, cost-effective, and transparent solution for adapting LLMs to the biomedical domain.

Clinicians often rely on data engineers to retrieve complex patient information from electronic health record (EHR) systems, a process that is both inefficient and time-consuming. We propose EHRAgent, a large language model (LLM) agent empowered with accumulative domain knowledge and robust coding capability. EHRAgent enables autonomous code generation and execution to facilitate clinicians in directly interacting with EHRs using natural language. Specifically, we formulate a multi-tabular reasoning task based on EHRs as a tool-use planning process, efficiently decomposing a complex task into a sequence of manageable actions with external toolsets. We first inject relevant medical information to enable EHRAgent to effectively reason about the given query, identifying and extracting the required records from the appropriate tables. By integrating interactive coding and execution feedback, EHRAgent then effectively learns from error messages and iteratively improves its originally generated code. Experiments on three real-world EHR datasets show that EHRAgent outperforms the strongest baseline by up to 29.6% in success rate, verifying its strong capacity to tackle complex clinical tasks with minimal demonstrations.

Prior research endeavors of the ensemble Large Language Models (LLMs) achieved great success by employing an individual language model (LM) rank before the text generation. However, the use of an individual LM ranker faces two primary challenges: (1) The time-intensive nature of the ranking process, stemming from the comparisons between models; (2) The issue of error propagation arising from the separate ranking and generation models within the framework. In order to overcome these challenges, we propose a novel ensemble framework, namely Unified Ranking and Generation (URG). URG represents an end-to-end framework that jointly ranks the outputs of LLMs and generates fine-grained fusion results, via utilizing a dedicated cross-attention-based module and noise mitigation training against irrelevant information stemming from bad ranking results. Through extensive experimentation and evaluation, we demonstrate the efficiency and effectiveness of our framework in both the ranking and generation tasks. With the close coordination of the ranking and generation modules, our end-to-end framework achieves the state-of-the-art (SOTA) performance on these tasks, and exhibits substantial enhancements to any of the ensembled models.

Stance detection is a challenging task that aims to identify public opinion from social media platforms with respect to specific targets. Previous work on stance detection largely focused on pure texts. In this paper, we study multi-modal stance detection for tweets consisting of texts and images, which are prevalent in today’s fast-growing social media platforms where people often post multi-modal messages. To this end, we create five new multi-modal stance detection datasets of different domains based on Twitter, in which each example consists of a text and an image. In addition, we propose a simple yet effective Targeted Multi-modal Prompt Tuning framework (TMPT), where target information is leveraged to learn multi-modal stance features from textual and visual modalities. Experimental results on our five benchmark datasets show that the proposed TMPT achieves state-of-the-art performance in multi-modal stance detection.

Clinical natural language processing faces challenges like complex medical terminology and clinical contexts. Recently, large language models (LLMs) have shown promise in this domain. Yet, their direct deployment can lead to privacy issues and are constrained by resources. To address this challenge, we delve into synthetic clinical text generation with LLMs for clinical NLP tasks. We propose an innovative, resource-efficient approach, ClinGen, which infuses knowledge into the process. Our model involves clinical knowledge extraction and context-informed LLM prompting. Both clinical topics and writing styles are drawn from external domain-specific knowledge graphs and LLMs to guide data generation. Our extensive empirical study across 8 clinical NLP tasks and 18 datasets reveals that ClinGen consistently enhances performance across various tasks by 7.7%-8.7% on average, effectively aligning the distribution of real datasets and enriching the diversity of generated training instances.

Although Large Language Models (LLMs) exhibit remarkable adaptability across domains, these models often fall short in structured knowledge extraction tasks such as named entity recognition (NER). This paper explores an innovative, cost-efficient strategy to harness LLMs with modest NER capabilities for producing superior NER datasets. Our approach diverges from the basic class-conditional prompts by instructing LLMs to self-reflect on the specific domain, thereby generating domain-relevant attributes (such as category and emotions for movie reviews), which are utilized for creating attribute-rich training data. Furthermore, we preemptively generate entity terms and then develop NER context data around these entities, effectively bypassing the LLMs’ challenges with complex structures. Our experiments across both general and niche domains reveal significant performance enhancements over conventional data generation methods while being more cost-effective than existing alternatives.

Hierarchical text classification (HTC) is a complex subtask under multi-label text classification, characterized by a hierarchical label taxonomy and data imbalance. The best-performing models aim to learn a static representation by combining document and hierarchical label information. However, the relevance of document sections can vary based on the hierarchy level, necessitating a dynamic document representation. To address this, we propose HiGen, a text-generation-based framework utilizing language models to encode dynamic text representations. We introduce a level-guided loss function to capture the relationship between text and label name semantics. Our approach incorporates a task-specific pretraining strategy, adapting the language model to in-domain knowledge and significantly enhancing performance for classes with limited examples. Furthermore, we present a new and valuable dataset called ENZYME, designed for HTC, which comprises articles from PubMed with the goal of predicting Enzyme Commission (EC) numbers. Through extensive experiments on the ENZYME dataset and the widely recognized WOS and NYT datasets, our methodology demonstrates superior performance, surpassing existing approaches while efficiently handling data and mitigating class imbalance. We release our code and dataset here: https://github.com/viditjain99/HiGen.

Retrieval-augmented generation enhances large language models (LLMs) by incorporating relevant information from external knowledge sources. This enables LLMs to adapt to specific domains and mitigate hallucinations in knowledge-intensive tasks. However, existing retrievers are often misaligned with LLMs due to separate training processes and the inherent black-box nature of LLMs. To address this challenge, we propose ARL2, a retriever learning technique that harnesses LLMs as labelers. ARL2 leverages LLMs to annotate and score adaptive relevance evidence, enabling the retriever to learn from robust LLM supervision. Furthermore, ARL2 incorporates a self-training strategy to minimize the cost of API calls. Extensive experiments demonstrate the effectiveness of ARL2, achieving accuracy improvements of 5.4% on NQ and 4.6% on MMLU compared to the state-of-the-art methods. Additionally, ARL2 exhibits robust transfer learning capabilities and strong zero-shot generalization abilities.

Comparative reasoning plays a crucial role in predicting text preferences; however, large language models (LLMs) often demonstrate inconsistencies in their reasoning, leading to incorrect preference predictions. While approaches like Chain-of-Thought improve accuracy in many settings, they struggle to consistently distinguish the similarities and differences of complex texts. We introduce SC², a model that prompts LLMs to predict text preferences by generating structured intermediate comparisons. SC² begins by proposing aspects for comparison, followed by generating textual comparisons under each aspect. We select consistent comparisons with a pairwise comparator that ensures each comparison of a given aspect clearly distinguishes differences between texts, significantly reducing hallucination and improving consistency. Our empirical studies across various NLP tasks, including summarization, retrieval, and automatic rating, demonstrate that SC²‘s enhanced performance in text preference prediction is significant.

Large language models (LLMs) have shown remarkable capabilities in various natural language understanding tasks with a few demonstration examples via in-context learning. Common strategies to boost such “in-context” learning ability are to ensemble multiple model decoded results and require the model to generate an explanation along with the prediction. However, these models often treat different class predictions equally and neglect the potential discrepancy between the explanations and predictions. To fully unleash the power of explanations, we propose EASE, an Explanation-Aware Soft Ensemble framework to empower in-context learning with LLMs. We design two techniques, explanation-guided ensemble, and soft probability aggregation, to mitigate the effect of unreliable explanations and improve the consistency between explanations and final predictions. Experiments on seven natural language understanding tasks and four varying-size LLMs demonstrate the effectiveness of our proposed framework.

We present RAM-EHR, a Retrieval AugMentation pipeline to improve clinical predictions on Electronic Health Records (EHRs). RAM-EHR first collects multiple knowledge sources, converts them into text format, and uses dense retrieval to obtain information related to medical concepts. This strategy addresses the difficulties associated with complex names for the concepts. RAM-EHR then augments the local EHR predictive model co-trained with consistency regularization to capture complementary information from patient visits and summarized knowledge. Experiments on two EHR datasets show the efficacy of RAM-EHR over previous knowledge-enhanced baselines (3.4% gain in AUROC and 7.2% gain in AUPR), emphasizing the effectiveness of the summarized knowledge from RAM-EHR for clinical prediction tasks.

Cross-lingual pre-training methods mask and predict tokens in multilingual text to generalize diverse multilingual information. However, due to the lack of sufficient aligned multilingual resources in the pre-training process, these methods may not fully explore the multilingual correlation of masked tokens, resulting in the limitation of multilingual information interaction. In this paper, we propose a lifelong multilingual multi-granularity semantic alignment approach, which continuously extracts massive aligned linguistic units from noisy data via a maximum co-occurrence probability algorithm. Then, the approach releases a version of the multilingual multi-granularity semantic alignment resource, supporting seven languages, namely English, Czech, German, Russian, Romanian, Hindi and Turkish. Finally, we propose how to use this resource to improve the translation performance on WMT14 18 benchmarks in twelve directions. Experimental results show an average of 0.3 1.1 BLEU improvements in all translation benchmarks. The analysis and discussion also demonstrate the superiority and potential of the proposed approach. The resource used in this work will be publicly available.

With the growing privacy concerns surrounding natural language understanding (NLU) applications, the need to train high-quality models while safeguarding data privacy has reached unprecedented importance. Federated learning (FL) offers a promising approach to collaborative model training by exchanging model gradients. However, many studies show that eavesdroppers in FL could develop sophisticated data reconstruction attack (DRA) to accurately reconstruct clients’ data from the shared gradients. Regrettably, current DRA methods in federated NLU have been mostly conducted on public datasets, lacking a comprehensive evaluation of real-world privacy datasets. To address this limitation, this paper presents a pioneering study that reexamines the performance of these DRA methods as well as corresponding defense methods. Specifically, we introduce a novel real-world privacy dataset called FedAttack which leads to a significant discovery: existing DRA methods usually fail to accurately recover the original text of real-world privacy data. In detail, the tokens within a recovery sentence are disordered and intertwined with tokens from other sentences in the same training batch. Moreover, our experiments demonstrate that the performance of DRA is also influenced by different languages and domains. By discovering these findings, our work lays a solid foundation for further research into the development of more practical DRA methods and corresponding defenses.

Knowledge Distillation (KD) serves as an efficient method for transferring language knowledge from open-source large language models (LLMs) to more computationally efficient models. However, challenges arise when attempting to apply vanilla KD methods to transfer knowledge from closed-source Multilingual Neural Machine Translation (MNMT) models based on LLMs. In this scenario, the soft labels and training data are not accessible, making it difficult to achieve effective knowledge transfer. To address this issue, this paper proposes a Teacher Assistant enhanced Knowledge Distillation (TAeKD) method to augment the knowledge transfer capacity from closed-source MNMT models. Specifically, TAeKD designs a fusion model that integrates translation outputs from multiple closed-source models to generate soft labels and training samples. Furthermore, a quality assessment learning mechanism is introduced to enhance the generalization of the fusion model and elevate the quality of the fusion data used to train the student model. To facilitate research on knowledge transfer from MNMT models, we also introduce FuseData, a benchmark consisting of a blend of translations from multiple closed-source systems. The experimental results show that TAeKD outperforms the previous state-of-the-art KD methods on both WMT22 and FLORES-101 test sets.

We present PATRON, a prompt-based data selection method for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available. In PATRON, we design (1) a prompt-based uncertainty propagation approach to estimate the importance of data points and (2) a partition-then-rewrite (PTR) strategy to promote sample diversity when querying for annotations. Experiments on six text classification datasets show that PATRON outperforms the strongest cold-start data selection baselines by up to 6.9%. Besides, with 128 labels only, PATRON achieves 91.0% and 92.1% of the fully supervised performance based on vanilla fine-tuning and prompt-based learning respectively. Our implementation of PATRON will be published upon acceptance.

Prompt-based methods have shown their efficacy in transferring general knowledge within pre-trained language models (PLMs) for low-resource scenarios. Typically, prompt-based methods convert downstream tasks to cloze-style problems and map all labels to verbalizers.However, when applied to zero-shot entity and relation extraction, vanilla prompt-based methods may struggle with the limited coverage of verbalizers to labels and the slow inference speed. In this work, we propose a novel Discriminate Soft Prompts (DSP) approach to take advantage of the prompt-based methods to strengthen the transmission of general knowledge. Specifically, we develop a discriminative prompt method, which reformulates zero-shot tasks into token discrimination tasks without having to construct verbalizers.Furthermore, to improve the inference speed of the prompt-based methods, we design a soft prompt co-reference strategy, which leverages soft prompts to approximately refer to the vector representation of text tokens. The experimental results show that, our model outperforms baselines on two zero-shot entity recognition datasets with higher inference speed, and obtains a 7.5% average relation F1-score improvement over previous state-of-the-art models on Wiki-ZSL and FewRel.

With increasing concerns about data privacy, there is an increasing necessity of fine-tuning pre-trained language models (PLMs) for adapting to downstream tasks located in end-user devices or local clients without transmitting data to the central server. This urgent necessity therefore calls the research of investigating federated learning (FL) for PLMs. However, large PLMs bring the curse of prohibitive communication overhead and local model adaptation costs for the FL system. To this end, we investigate the parameter-efficient tuning (PETuning) of PLMs and develop a corresponding federated benchmark for four representative PETuning methods, dubbed FedPETuning. Specifically, FedPETuning provides the first holistic empirical study of representative PLMs tuning methods in FL, covering privacy attacks, performance comparisons, and resource-constrained analysis. Intensive experimental results have indicated that FedPETuning can efficiently defend against privacy attacks and maintains acceptable performance with reducing heavy resource consumption. The open-source code and data are available at https://github.com/SMILELab-FL/FedPETuning.

With the development of large language models (LLMs), zero-shot learning has attracted much attention for various NLP tasks. Different from prior works that generate training data with billion-scale natural language generation (NLG) models, we propose a retrieval-enhanced framework to create training data from a general-domain unlabeled corpus. To realize this, we first conduct contrastive pretraining to learn an unsupervised dense retriever for extracting the most relevant documents using class-descriptive verbalizers. We then further pro- pose two simple strategies, namely Verbalizer Augmentation with Demonstrations and Self- consistency Guided Filtering to improve the topic coverage of the dataset while removing noisy examples. Experiments on nine datasets demonstrate that ReGen achieves 4.3% gain over the strongest baselines and saves around 70% of the time when compared with baselines using large NLG models. Besides, REGEN can be naturally integrated with recently proposed large language models to boost performance.

Although fine-tuning pre-trained language models (PLMs) renders strong performance in many NLP tasks, it relies on excessive labeled data. Recently, researchers have resorted to active fine-tuning for enhancing the label efficiency of PLM fine-tuning, but existing methods of this type usually ignore the potential of unlabeled data. We develop AcTune, a new framework that improves the label efficiency of active PLM fine-tuning by unleashing the power of unlabeled data via self-training. AcTune switches between data annotation and model self-training based on uncertainty: the unlabeled samples of high-uncertainty are selected for annotation, while the ones from low-uncertainty regions are used for model self-training. Additionally, we design (1) a region-aware sampling strategy to avoid redundant samples when querying annotations and (2) a momentum-based memory bank to dynamically aggregate the model’s pseudo labels to suppress label noise in self-training. Experiments on 6 text classification datasets show that AcTune outperforms the strongest active learning and self-training baselines and improves the label efficiency of PLM fine-tuning by 56.2% on average. Our implementation is available at https://github.com/yueyu1030/actune.

Weakly-supervised learning (WSL) has shown promising results in addressing label scarcity on many NLP tasks, but manually designing a comprehensive, high-quality labeling rule set is tedious and difficult. We study interactive weakly-supervised learning—the problem of iteratively and automatically discovering novel labeling rules from data to improve the WSL model. Our proposed model, named PRBoost, achieves this goal via iterative prompt-based rule discovery and model boosting. It uses boosting to identify large-error instances and discovers candidate rules from them by prompting pre-trained LMs with rule templates. The candidate rules are judged by human experts, and the accepted rules are used to generate complementary weak labels and strengthen the current model. Experiments on four tasks show PRBoost outperforms state-of-the-art WSL baselines up to 7.1%, and bridges the gaps with fully supervised models.

We study the problem of extracting N-ary relation tuples from scientific articles. This task is challenging because the target knowledge tuples can reside in multiple parts and modalities of the document. Our proposed method ReSel decomposes this task into a two-stage procedure that first retrieves the most relevant paragraph/table and then selects the target entity from the retrieved component. For the high-level retrieval stage, ReSel designs a simple and effective feature set, which captures multi-level lexical and semantic similarities between the query and components. For the low-level selection stage, ReSel designs a cross-modal entity correlation graph along with a multi-view architecture, which models both semantic and document-structural relations between entities. Our experiments on three scientific information extraction datasets show that ReSel outperforms state-of-the-art baselines significantly.

We present a new zero-shot dense retrieval (ZeroDR) method, COCO-DR, to improve the generalization ability of dense retrieval by combating the distribution shifts between source training tasks and target scenarios. To mitigate the impact of document differences, COCO-DR continues pretraining the language model on the target corpora to adapt the model to target distributions via COtinuous COtrastive learning. To prepare for unseen target queries, COCO-DR leverages implicit Distributionally Robust Optimization (iDRO) to reweight samples from different source query clusters for improving model robustness over rare queries during fine-tuning. COCO-DR achieves superior average performance on BEIR, the zero-shot retrieval benchmark. At BERT_Base scale, COCO-DR Base outperforms other ZeroDR models with 60x larger size. At BERT_Large scale, COCO-DR Large outperforms the giant GPT-3 embedding model which has 500x more parameters. Our analysis shows the correlation between COCO-DR’s effectiveness in combating distribution shifts and improving zero-shot accuracy. Our code and model can be found at https://github.com/OpenMatch/COCO-DR.

In this paper, we investigate the instability in the standard dense retrieval training, which iterates between model training and hard negative selection using the being-trained model. We show the catastrophic forgetting phenomena behind the training instability, where models learn and forget different negative groups during training iterations. We then propose ANCE-Tele, which accumulates momentum negatives from past iterations and approximates future iterations using lookahead negatives, as “teleportations” along the time axis to smooth the learning process. On web search and OpenQA, ANCE-Tele outperforms previous state-of-the-art systems of similar size, eliminates the dependency on sparse retrieval negatives, and is competitive among systems using significantly more (50x) parameters. Our analysis demonstrates that teleportation negatives reduce catastrophic forgetting and improve convergence speed for dense retrieval training. The source code of this paper is available at https://github.com/OpenMatch/ANCE-Tele.

Self-training achieves enormous success in various semi-supervised and weakly-supervised learning tasks. The method can be interpreted as a teacher-student framework, where the teacher generates pseudo-labels, and the student makes predictions. The two models are updated alternatingly. However, such a straightforward alternating update rule leads to training instability. This is because a small change in the teacher may result in a significant change in the student. To address this issue, we propose DRIFT, short for differentiable self-training, that treats teacher-student as a Stackelberg game. In this game, a leader is always in a more advantageous position than a follower. In self-training, the student contributes to the prediction performance, and the teacher controls the training process by generating pseudo-labels. Therefore, we treat the student as the leader and the teacher as the follower. The leader procures its advantage by acknowledging the follower’s strategy, which involves differentiable pseudo-labels and differentiable sample weights. Consequently, the leader-follower interaction can be effectively captured via Stackelberg gradient, obtained by differentiating the follower’s strategy. Experimental results on semi- and weakly-supervised classification and named entity recognition tasks show that our model outperforms existing approaches by large margins.

Fine-tuned pre-trained language models (LMs) have achieved enormous success in many natural language processing (NLP) tasks, but they still require excessive labeled data in the fine-tuning stage. We study the problem of fine-tuning pre-trained LMs using only weak supervision, without any labeled data. This problem is challenging because the high capacity of LMs makes them prone to overfitting the noisy labels generated by weak supervision. To address this problem, we develop a contrastive self-training framework, COSINE, to enable fine-tuning LMs with weak supervision. Underpinned by contrastive regularization and confidence-based reweighting, our framework gradually improves model fitting while effectively suppressing error propagation. Experiments on sequence, token, and sentence pair classification tasks show that our model outperforms the strongest baseline by large margins and achieves competitive performance with fully-supervised fine-tuning methods. Our implementation is available on https://github.com/yueyu1030/COSINE.

Multi-hop reasoning approaches over knowledge graphs infer a missing relationship between entities with a multi-hop rule, which corresponds to a chain of relationships. We extend existing works to consider a generalized form of multi-hop rules, where each rule is a set of relation chains. To learn such generalized rules efficiently, we propose a two-step approach that first selects a small set of relation chains as a rule and then evaluates the confidence of the target relationship by jointly scoring the selected chains. A game-theoretical framework is proposed to this end to simultaneously optimize the rule selection and prediction steps. Empirical results show that our multi-chain multi-hop (MCMH) rules result in superior results compared to the standard single-chain approaches, justifying both our formulation of generalized rules and the effectiveness of the proposed learning framework.

Active learning is an important technique for low-resource sequence labeling tasks. However, current active sequence labeling methods use the queried samples alone in each iteration, which is an inefficient way of leveraging human annotations. We propose a simple but effective data augmentation method to improve label efficiency of active sequence labeling. Our method, SeqMix, simply augments the queried samples by generating extra labeled sequences in each iteration. The key difficulty is to generate plausible sequences along with token-level labels. In SeqMix, we address this challenge by performing mixup for both sequences and token-level labels of the queried samples. Furthermore, we design a discriminator during sequence mixup, which judges whether the generated sequences are plausible or not. Our experiments on Named Entity Recognition and Event Detection tasks show that SeqMix can improve the standard active sequence labeling method by 2.27%–3.75% in terms of F₁ scores. The code and data for SeqMix can be found at https://github.com/rz-zhang/SeqMix.

In this paper we present GumDrop, Georgetown University’s entry at the DISRPT 2019 Shared Task on automatic discourse unit segmentation and connective detection. Our approach relies on model stacking, creating a heterogeneous ensemble of classifiers, which feed into a metalearner for each final task. The system encompasses three trainable component stacks: one for sentence splitting, one for discourse unit segmentation and one for connective detection. The flexibility of each ensemble allows the system to generalize well to datasets of different sizes and with varying levels of homogeneity.

Generally the existing monolingual corpora are not suitable for large vocabulary continuous speech recognition (LVCSR) of code-switching speech. The motivation of this paper is to study the rules and constraints code-switching follows and design a corpus for code-switching LVCSR task. This paper presents the development of a Mandarin-English code-switching corpus. This corpus consists of four parts: 1) conversational meeting speech and its data; 2) project meeting speech data; 3) student interviews speech; 4) text data of on-line news. The speech was transcribed by an annotator and verified by Mandarin-English bilingual speakers manually. We propose an approach for automatically downloading from the web text data that contains code-switching. The corpus includes both intra-sentential code-switching (switch in the middle of a sentence) and inter-sentential code-switching (switch at the end of the sentence). The distribution of part-of-speech (POS) tags and code-switching reasons are reported.