Rui Zhang

Hongjin Su, Jungo Kasai, Chen Henry Wu et al. · allen-ai, cmu

Many recent approaches to natural language tasks are built on the remarkable abilities of large language models. Large language models can perform in-context learning, where they learn a new task from a few task demonstrations, without any parameter updates. This work examines the implications of in-context learning for the creation of datasets for new natural language tasks. Departing from recent in-context learning methods, we formulate an annotation-efficient, two-step framework: selective annotation that chooses a pool of examples to annotate from unlabeled data in advance, followed by prompt retrieval that retrieves task examples from the annotated pool at test time. Based on this framework, we propose an unsupervised, graph-based selective annotation method, voke-k, to select diverse, representative examples to annotate. Extensive experiments on 10 datasets (covering classification, commonsense reasoning, dialogue, and text/code generation) demonstrate that our selective annotation method improves the task performance by a large margin. On average, vote-k achieves a 12.9%/11.4% relative gain under an annotation budget of 18/100, as compared to randomly selecting examples to annotate. Compared to state-of-the-art supervised finetuning approaches, it yields similar performance with 10-100x less annotation cost across 10 tasks. We further analyze the effectiveness of our framework in various scenarios: language models with varying sizes, alternative selective annotation methods, and cases where there is a test data domain shift. We hope that our studies will serve as a basis for data annotations as large language models are increasingly applied to new tasks. Our code is available at https://github.com/HKUNLP/icl-selective-annotation.

Simeng Han, Hailey Schoelkopf, Yilun Zhao et al. · salesforce

Large language models (LLMs) have achieved remarkable performance on a variety of natural language understanding tasks. However, existing benchmarks are inadequate in measuring the complex logical reasoning capabilities of a model. We present FOLIO, a human-annotated, logically complex and diverse dataset for reasoning in natural language (NL), equipped with first-order logic (FOL) annotations. FOLIO consists of 1,430 examples (unique conclusions), each paired with one of 487 sets of premises used to deductively reason for the validity of each conclusion. The logical correctness of the premises and conclusions is ensured by their FOL annotations, which are automatically verified by an FOL inference engine. In addition to the main NL reasoning task, NL-FOL pairs in FOLIO constitute a new NL-FOL translation dataset. Our experiments on FOLIO systematically evaluate the FOL reasoning ability of supervised fine-tuning on medium-sized language models. For both NL reasoning and NL-FOL translation, we benchmark multiple state-of-the-art language models. Our results show that a subset of FOLIO presents a challenge for one of the most capable {Large Language Model (LLM)} publicly available, GPT-4.

Yilun Zhao, Yunxiang Li, Chenying Li et al.

Numerical reasoning over hybrid data containing both textual and tabular content (e.g., financial reports) has recently attracted much attention in the NLP community. However, existing question answering (QA) benchmarks over hybrid data only include a single flat table in each document and thus lack examples of multi-step numerical reasoning across multiple hierarchical tables. To facilitate data analytical progress, we construct a new large-scale benchmark, MultiHiertt, with QA pairs over Multi Hierarchical Tabular and Textual data. MultiHiertt is built from a wealth of financial reports and has the following unique characteristics: 1) each document contain multiple tables and longer unstructured texts; 2) most of tables contained are hierarchical; 3) the reasoning process required for each question is more complex and challenging than existing benchmarks; and 4) fine-grained annotations of reasoning processes and supporting facts are provided to reveal complex numerical reasoning. We further introduce a novel QA model termed MT2Net, which first applies facts retrieving to extract relevant supporting facts from both tables and text and then uses a reasoning module to perform symbolic reasoning over retrieved facts. We conduct comprehensive experiments on various baselines. The experimental results show that MultiHiertt presents a strong challenge for existing baselines whose results lag far behind the performance of human experts. The dataset and code are publicly available at https://github.com/psunlpgroup/MultiHiertt.

Xiaoyan Zhao, Yang Deng, Min Yang et al.

Relation extraction (RE) involves identifying the relations between entities from underlying content. RE serves as the foundation for many natural language processing (NLP) and information retrieval applications, such as knowledge graph completion and question answering. In recent years, deep neural networks have dominated the field of RE and made noticeable progress. Subsequently, the large pre-trained language models have taken the state-of-the-art RE to a new level. This survey provides a comprehensive review of existing deep learning techniques for RE. First, we introduce RE resources, including datasets and evaluation metrics. Second, we propose a new taxonomy to categorize existing works from three perspectives, i.e., text representation, context encoding, and triplet prediction. Third, we discuss several important challenges faced by RE and summarize potential techniques to tackle these challenges. Finally, we outline some promising future directions and prospects in this field. This survey is expected to facilitate researchers' collaborative efforts to address the challenges of real-world RE systems.

Sicheng Zhou, Nan Wang, Liwei Wang et al.

Objective: The generalizability of clinical large language models is usually ignored during the model development process. This study evaluated the generalizability of BERT-based clinical NLP models across different clinical settings through a breast cancer phenotype extraction task. Materials and Methods: Two clinical corpora of breast cancer patients were collected from the electronic health records from the University of Minnesota and the Mayo Clinic, and annotated following the same guideline. We developed three types of NLP models (i.e., conditional random field, bi-directional long short-term memory and CancerBERT) to extract cancer phenotypes from clinical texts. The models were evaluated for their generalizability on different test sets with different learning strategies (model transfer vs. locally trained). The entity coverage score was assessed with their association with the model performances. Results: We manually annotated 200 and 161 clinical documents at UMN and MC, respectively. The corpora of the two institutes were found to have higher similarity between the target entities than the overall corpora. The CancerBERT models obtained the best performances among the independent test sets from two clinical institutes and the permutation test set. The CancerBERT model developed in one institute and further fine-tuned in another institute achieved reasonable performance compared to the model developed on local data (micro-F1: 0.925 vs 0.932). Conclusions: The results indicate the CancerBERT model has the best learning ability and generalizability among the three types of clinical NLP models. The generalizability of the models was found to be correlated with the similarity of the target entities between the corpora.

Mingchen Li, Rui Zhang

Recent advancements in language models (LMs) have led to the emergence of powerful models such as Small LMs (e.g., T5) and Large LMs (e.g., GPT-4). These models have demonstrated exceptional capabilities across a wide range of tasks, such as name entity recognition (NER) in the general domain. (We define SLMs as pre-trained models with fewer parameters compared to models like GPT-3/3.5/4, such as T5, BERT, and others.) Nevertheless, their efficacy in the medical section remains uncertain and the performance of medical NER always needs high accuracy because of the particularity of the field. This paper aims to provide a thorough investigation to compare the performance of LMs in medical few-shot NER and answer How far is LMs from 100\% Few-shot NER in Medical Domain, and moreover to explore an effective entity recognizer to help improve the NER performance. Based on our extensive experiments conducted on 16 NER models spanning from 2018 to 2023, our findings clearly indicate that LLMs outperform SLMs in few-shot medical NER tasks, given the presence of suitable examples and appropriate logical frameworks. Despite the overall superiority of LLMs in few-shot medical NER tasks, it is important to note that they still encounter some challenges, such as misidentification, wrong template prediction, etc. Building on previous findings, we introduce a simple and effective method called \textsc{RT} (Retrieving and Thinking), which serves as retrievers, finding relevant examples, and as thinkers, employing a step-by-step reasoning process. Experimental results show that our proposed \textsc{RT} framework significantly outperforms the strong open baselines on the two open medical benchmark datasets

Abdullah Al Ishtiaq, Sarkar Snigdha Sarathi Das, Syed Md Mukit Rashid et al.

In this paper, we present Hermes, an end-to-end framework to automatically generate formal representations from natural language cellular specifications. We first develop a neural constituency parser, NEUTREX, to process transition-relevant texts and extract transition components (i.e., states, conditions, and actions). We also design a domain-specific language to translate these transition components to logical formulas by leveraging dependency parse trees. Finally, we compile these logical formulas to generate transitions and create the formal model as finite state machines. To demonstrate the effectiveness of Hermes, we evaluate it on 4G NAS, 5G NAS, and 5G RRC specifications and obtain an overall accuracy of 81-87%, which is a substantial improvement over the state-of-the-art. Our security analysis of the extracted models uncovers 3 new vulnerabilities and identifies 19 previous attacks in 4G and 5G specifications, and 7 deviations in commercial 4G basebands.

Mingchen Li, Huixue Zhou, Rui Zhang

Biomedical triple extraction systems aim to automatically extract biomedical entities and relations between entities. The exploration of applying large language models (LLM) to triple extraction is still relatively unexplored. In this work, we mainly focus on sentence-level biomedical triple extraction. Furthermore, the absence of a high-quality biomedical triple extraction dataset impedes the progress in developing robust triple extraction systems. To address these challenges, initially, we compare the performance of various large language models. Additionally, we present GIT, an expert-annotated biomedical triple extraction dataset that covers a wider range of relation types.

Nan Zhang, Yanchi Liu, Xujiang Zhao et al.

Large Language Models (LLMs) have exhibited remarkable proficiency across a wide array of NLP tasks. However, the escalation in model size also engenders substantial deployment costs. While few efforts have explored model pruning techniques to reduce the size of LLMs, they mainly center on general or task-specific weights. This leads to suboptimal performance due to lacking specificity on the target domain or generality on different tasks when applied to domain-specific challenges. This work introduces an innovative unstructured dual-pruning methodology, D-Pruner, for domain-specific compression on LLM. It extracts a compressed, domain-specific, and task-agnostic LLM by identifying LLM weights that are pivotal for general capabilities, like linguistic capability and multi-task solving, and domain-specific knowledge. More specifically, we first assess general weight importance by quantifying the error incurred upon their removal with the help of an open-domain calibration dataset. Then, we utilize this general weight importance to refine the training loss, so that it preserves generality when fitting into a specific domain. Moreover, by efficiently approximating weight importance with the refined training loss on a domain-specific calibration dataset, we obtain a pruned model emphasizing generality and specificity. Our comprehensive experiments across various tasks in healthcare and legal domains show the effectiveness of D-Pruner in domain-specific compression. Our code is available at https://github.com/psunlpgroup/D-Pruner.

Dayong Wu, Jiaqi Li, Baoxin Wang et al.

Large language models (LLMs) have shown remarkable achievements across various language tasks.To enhance the performance of LLMs in scientific literature services, we developed the scientific literature LLM (SciLit-LLM) through pre-training and supervised fine-tuning on scientific literature, building upon the iFLYTEK Spark LLM. Furthermore, we present a knowledge service system Spark Research Assistant (SparkRA) based on our SciLit-LLM. SparkRA is accessible online and provides three primary functions: literature investigation, paper reading, and academic writing. As of July 30, 2024, SparkRA has garnered over 50,000 registered users, with a total usage count exceeding 1.3 million.

Sarkar Snigdha Sarathi Das, Ryo Kamoi, Bo Pang et al.

The effectiveness of large language models (LLMs) is closely tied to the design of prompts, making prompt optimization essential for enhancing their performance across a wide range of tasks. Many existing approaches to automating prompt engineering rely exclusively on textual feedback, refining prompts based solely on inference errors identified by large, computationally expensive LLMs. Unfortunately, smaller models struggle to generate high-quality feedback, resulting in complete dependence on large LLM judgment. Moreover, these methods fail to leverage more direct and finer-grained information, such as gradients, due to operating purely in text space. To this end, we introduce GReaTer, a novel prompt optimization technique that directly incorporates gradient information over task-specific reasoning. By utilizing task loss gradients, GReaTer enables self-optimization of prompts for open-source, lightweight language models without the need for costly closed-source LLMs. This allows high-performance prompt optimization without dependence on massive LLMs, closing the gap between smaller models and the sophisticated reasoning often needed for prompt refinement. Extensive evaluations across diverse reasoning tasks including BBH, GSM8k, and FOLIO demonstrate that GReaTer consistently outperforms previous state-of-the-art prompt optimization methods, even those reliant on powerful LLMs. Additionally, GReaTer-optimized prompts frequently exhibit better transferability and, in some cases, boost task performance to levels comparable to or surpassing those achieved by larger language models, highlighting the effectiveness of prompt optimization guided by gradients over reasoning. Code of GReaTer is available at https://github.com/psunlpgroup/GreaTer.

Haoyuan Li, Yusen Zhang, Rui Zhang et al.

Fairness in multi-document summarization (MDS) measures whether a system can generate a summary fairly representing information from documents with different social attribute values. Fairness in MDS is crucial since a fair summary can offer readers a comprehensive view. Previous works focus on quantifying summary-level fairness using Proportional Representation, a fairness measure based on Statistical Parity. However, Proportional Representation does not consider redundancy in input documents and overlooks corpus-level unfairness. In this work, we propose a new summary-level fairness measure, Equal Coverage, which is based on coverage of documents with different social attribute values and considers the redundancy within documents. To detect the corpus-level unfairness, we propose a new corpus-level measure, Coverage Parity. Our human evaluations show that our measures align more with our definition of fairness. Using our measures, we evaluate the fairness of thirteen different LLMs. We find that Claude3-sonnet is the fairest among all evaluated LLMs. We also find that almost all LLMs overrepresent different social attribute values. The code is available at https://github.com/leehaoyuan/coverage_fairness.

Haoyuan Li, Rui Zhang, Snigdha Chaturvedi

Fairness in multi-document summarization (MDS) is crucial for providing comprehensive views across documents with diverse social attribute values, which can significantly impact decision-making. For example, a summarization system that tends to overrepresent negative reviews of products can mislead customers into disregarding good products. Previous works measure fairness in MDS at two levels: summary-level and corpus-level. While summary-level fairness focuses on individual summaries, corpus-level fairness focuses on a corpus of summaries. Recent methods primarily focus on summary-level fairness. We propose FairPO, a preference tuning method that focuses on both summary-level and corpus-level fairness in MDS. To improve summary-level fairness, we propose to generate preference pairs by perturbing document sets. To improve corpus-level fairness, we propose fairness-aware preference tuning by dynamically adjusting the weights of preference pairs. Our experiments show that FairPO outperforms strong baselines while maintaining the critical qualities of summaries. The code is available at https://github.com/leehaoyuan/coverage_fairnes.

Tianbao Xie, Chen Henry Wu, Peng Shi et al.

Structured knowledge grounding (SKG) leverages structured knowledge to complete user requests, such as semantic parsing over databases and question answering over knowledge bases. Since the inputs and outputs of SKG tasks are heterogeneous, they have been studied separately by different communities, which limits systematic and compatible research on SKG. In this paper, we overcome this limitation by proposing the UnifiedSKG framework, which unifies 21 SKG tasks into a text-to-text format, aiming to promote systematic SKG research, instead of being exclusive to a single task, domain, or dataset. We use UnifiedSKG to benchmark T5 with different sizes and show that T5, with simple modifications when necessary, achieves state-of-the-art performance on almost all of the 21 tasks. We further demonstrate that multi-task prefix-tuning improves the performance on most tasks, largely improving the overall performance. UnifiedSKG also facilitates the investigation of zero-shot and few-shot learning, and we show that T0, GPT-3, and Codex struggle in zero-shot and few-shot learning for SKG. We also use UnifiedSKG to conduct a series of controlled experiments on structured knowledge encoding variants across SKG tasks. UnifiedSKG is easily extensible to more tasks, and it is open-sourced at https://github.com/hkunlp/unifiedskg.

Ruoqian Lin, Rui Zhang, Chunyang Wang et al.

Atom segmentation and localization, noise reduction and deblurring of atomic-resolution scanning transmission electron microscopy (STEM) images with high precision and robustness is a challenging task. Although several conventional algorithms, such has thresholding, edge detection and clustering, can achieve reasonable performance in some predefined sceneries, they tend to fail when interferences from the background are strong and unpredictable. Particularly, for atomic-resolution STEM images, so far there is no well-established algorithm that is robust enough to segment or detect all atomic columns when there is large thickness variation in a recorded image. Herein, we report the development of a training library and a deep learning method that can perform robust and precise atom segmentation, localization, denoising, and super-resolution processing of experimental images. Despite using simulated images as training datasets, the deep-learning model can self-adapt to experimental STEM images and shows outstanding performance in atom detection and localization in challenging contrast conditions and the precision consistently outperforms the state-of-the-art two-dimensional Gaussian fit method. Taking a step further, we have deployed our deep-learning models to a desktop app with a graphical user interface and the app is free and open-source. We have also built a TEM ImageNet project website for easy browsing and downloading of the training data.

Linyong Nan, Dragomir Radev, Rui Zhang et al.

We present DART, an open domain structured DAta Record to Text generation dataset with over 82k instances (DARTs). Data-to-Text annotations can be a costly process, especially when dealing with tables which are the major source of structured data and contain nontrivial structures. To this end, we propose a procedure of extracting semantic triples from tables that encodes their structures by exploiting the semantic dependencies among table headers and the table title. Our dataset construction framework effectively merged heterogeneous sources from open domain semantic parsing and dialogue-act-based meaning representation tasks by utilizing techniques such as: tree ontology annotation, question-answer pair to declarative sentence conversion, and predicate unification, all with minimum post-editing. We present systematic evaluation on DART as well as new state-of-the-art results on WebNLG 2017 to show that DART (1) poses new challenges to existing data-to-text datasets and (2) facilitates out-of-domain generalization. Our data and code can be found at https://github.com/Yale-LILY/dart.

Mingchen Li, Halil Kilicoglu, Hua Xu et al.

Large Language Models (LLMs) have swiftly emerged as vital resources for different applications in the biomedical and healthcare domains; however, these models encounter issues such as generating inaccurate information or hallucinations. Retrieval-augmented generation provided a solution for these models to update knowledge and enhance their performance. In contrast to previous retrieval-augmented LMs, which utilize specialized cross-attention mechanisms to help LLM encode retrieved text, BiomedRAG adopts a simpler approach by directly inputting the retrieved chunk-based documents into the LLM. This straightforward design is easily applicable to existing retrieval and language models, effectively bypassing noise information in retrieved documents, particularly in noise-intensive tasks. Moreover, we demonstrate the potential for utilizing the LLM to supervise the retrieval model in the biomedical domain, enabling it to retrieve the document that assists the LM in improving its predictions. Our experiments reveal that with the tuned scorer,\textsc{ BiomedRAG} attains superior performance across 5 biomedical NLP tasks, encompassing information extraction (triple extraction, relation extraction), text classification, link prediction, and question-answering, leveraging over 9 datasets. For instance, in the triple extraction task, \textsc{BiomedRAG} outperforms other triple extraction systems with micro-F1 scores of 81.42 and 88.83 on GIT and ChemProt corpora, respectively.

Zaifu Zhan, Jun Wang, Shuang Zhou et al.

Objective: To optimize in-context learning in biomedical natural language processing by improving example selection. Methods: We introduce a novel multi-mode retrieval-augmented generation (MMRAG) framework, which integrates four retrieval strategies: (1) Random Mode, selecting examples arbitrarily; (2) Top Mode, retrieving the most relevant examples based on similarity; (3) Diversity Mode, ensuring variation in selected examples; and (4) Class Mode, selecting category-representative examples. This study evaluates MMRAG on three core biomedical NLP tasks: Named Entity Recognition (NER), Relation Extraction (RE), and Text Classification (TC). The datasets used include BC2GM for gene and protein mention recognition (NER), DDI for drug-drug interaction extraction (RE), GIT for general biomedical information extraction (RE), and HealthAdvice for health-related text classification (TC). The framework is tested with two large language models (Llama2-7B, Llama3-8B) and three retrievers (Contriever, MedCPT, BGE-Large) to assess performance across different retrieval strategies. Results: The results from the Random mode indicate that providing more examples in the prompt improves the model's generation performance. Meanwhile, Top mode and Diversity mode significantly outperform Random mode on the RE (DDI) task, achieving an F1 score of 0.9669, a 26.4% improvement. Among the three retrievers tested, Contriever outperformed the other two in a greater number of experiments. Additionally, Llama 2 and Llama 3 demonstrated varying capabilities across different tasks, with Llama 3 showing a clear advantage in handling NER tasks. Conclusion: MMRAG effectively enhances biomedical in-context learning by refining example selection, mitigating data scarcity issues, and demonstrating superior adaptability for NLP-driven healthcare applications.

Zaifu Zhan, Rui Zhang

To efficiently select optimal dataset combinations for enhancing multi-task learning (MTL) performance in large language models, we proposed a novel framework that leverages a neural network to predict the best dataset combinations. The framework iteratively refines the selection, greatly improving efficiency, while being model-, dataset-, and domain-independent. Through experiments on 12 biomedical datasets across four tasks - named entity recognition, relation extraction, event extraction, and text classification-we demonstrate that our approach effectively identifies better combinations, even for tasks that may seem unpromising from a human perspective. This verifies that our framework provides a promising solution for maximizing MTL potential.

Zaifu Zhan, Shuang Zhou, Huixue Zhou et al.

The advancement of Large Language Models (LLMs) has significantly impacted biomedical Natural Language Processing (NLP), enhancing tasks such as named entity recognition, relation extraction, event extraction, and text classification. In this context, the DeepSeek series of models have shown promising potential in general NLP tasks, yet their capabilities in the biomedical domain remain underexplored. This study evaluates multiple DeepSeek models (Distilled-DeepSeek-R1 series and Deepseek-LLMs) across four key biomedical NLP tasks using 12 datasets, benchmarking them against state-of-the-art alternatives (Llama3-8B, Qwen2.5-7B, Mistral-7B, Phi-4-14B, Gemma-2-9B). Our results reveal that while DeepSeek models perform competitively in named entity recognition and text classification, challenges persist in event and relation extraction due to precision-recall trade-offs. We provide task-specific model recommendations and highlight future research directions. This evaluation underscores the strengths and limitations of DeepSeek models in biomedical NLP, guiding their future deployment and optimization.

GuangHao Meng, Sunan He, Jinpeng Wang et al.

Vision-language retrieval (VLR) has attracted significant attention in both academia and industry, which involves using text (or images) as queries to retrieve corresponding images (or text). However, existing methods often neglect the rich visual semantics knowledge of entities, thus leading to incorrect retrieval results. To address this problem, we propose the Entity Visual Description enhanced CLIP (EvdCLIP), designed to leverage the visual knowledge of entities to enrich queries. Specifically, since humans recognize entities through visual cues, we employ a large language model (LLM) to generate Entity Visual Descriptions (EVDs) as alignment cues to complement textual data. These EVDs are then integrated into raw queries to create visually-rich, EVD-enhanced queries. Furthermore, recognizing that EVD-enhanced queries may introduce noise or low-quality expansions, we develop a novel, trainable EVD-aware Rewriter (EaRW) for vision-language retrieval tasks. EaRW utilizes EVD knowledge and the generative capabilities of the language model to effectively rewrite queries. With our specialized training strategy, EaRW can generate high-quality and low-noise EVD-enhanced queries. Extensive quantitative and qualitative experiments on image-text retrieval benchmarks validate the superiority of EvdCLIP on vision-language retrieval tasks.

Peichao Wang, Jiabao Wang, Yao Chen et al.

The data-driven method for infrared small target detection (IRSTD) has achieved promising results. However, due to the small scale of infrared small target datasets and the limited number of pixels occupied by the targets themselves, it is a challenging task for deep learning methods to directly learn from these samples. Utilizing human expert knowledge to assist deep learning methods in better learning is worthy of exploration. To effectively guide the model to focus on targets' spatial features, this paper proposes the Local Contrast Attention Enhanced infrared small target detection Network (LCAE-Net), combining prior knowledge with data-driven deep learning methods. LCAE-Net is a U-shaped neural network model which consists of two developed modules: a Local Contrast Enhancement (LCE) module and a Channel Attention Enhancement (CAE) module. The LCE module takes advantages of prior knowledge, leveraging handcrafted convolution operator to acquire Local Contrast Attention (LCA), which could realize background suppression while enhance the potential target region, thus guiding the neural network to pay more attention to potential infrared small targets' location information. To effectively utilize the response information throughout downsampling progresses, the CAE module is proposed to achieve the information fusion among feature maps' different channels. Experimental results indicate that our LCAE-Net outperforms existing state-of-the-art methods on the three public datasets NUDT-SIRST, NUAA-SIRST, and IRSTD-1K, and its detection speed could reach up to 70 fps. Meanwhile, our model has a parameter count and Floating-Point Operations (FLOPs) of 1.945M and 4.862G respectively, which is suitable for deployment on edge devices.

Zaifu Zhan, Shuang Zhou, Min Zeng et al.

Large language models have demonstrated remarkable capabilities in biomedical natural language processing, yet their rapid growth in size and computational requirements present a major barrier to adoption in healthcare settings where data privacy precludes cloud deployment and resources are limited. In this study, we systematically evaluated the impact of quantization on 12 state-of-the-art large language models, including both general-purpose and biomedical-specific models, across eight benchmark datasets covering four key tasks: named entity recognition, relation extraction, multi-label classification, and question answering. We show that quantization substantially reduces GPU memory requirements-by up to 75%-while preserving model performance across diverse tasks, enabling the deployment of 70B-parameter models on 40GB consumer-grade GPUs. In addition, domain-specific knowledge and responsiveness to advanced prompting methods are largely maintained. These findings provide significant practical and guiding value, highlighting quantization as a practical and effective strategy for enabling the secure, local deployment of large yet high-capacity language models in biomedical contexts, bridging the gap between technical advances in AI and real-world clinical translation.

Haotian Li, Rui Zhang, Lingzhi Wang et al.

Recent advances in knowledge graph completion (KGC) have emphasized text-based approaches to navigate the inherent complexities of large-scale knowledge graphs (KGs). While these methods have achieved notable progress, they frequently struggle to fully incorporate the global structural properties of the graph. Stochastic blockmodels (SBMs), especially the latent feature relational model (LFRM), offer robust probabilistic frameworks for identifying latent community structures and improving link prediction. This paper presents a novel probabilistic KGC framework utilizing sparse latent feature models, optimized via a deep variational autoencoder (VAE). Our proposed method dynamically integrates global clustering information with local textual features to effectively complete missing triples, while also providing enhanced interpretability of the underlying latent structures. Extensive experiments on four benchmark datasets with varying scales demonstrate the significant performance gains achieved by our method.

Mingchen Li, Yang Ye, Jeremy Yeung et al.

Contrastive learning has become a popular solution for few-shot Name Entity Recognization (NER). The conventional configuration strives to reduce the distance between tokens with the same labels and increase the distance between tokens with different labels. The effect of this setup may, however, in the medical domain, there are a lot of entities annotated as OUTSIDE (O), and they are undesirably pushed apart to other entities that are not labeled as OUTSIDE (O) by the current contrastive learning method end up with a noisy prototype for the semantic representation of the label, though there are many OUTSIDE (O) labeled entities are relevant to the labeled entities. To address this challenge, we propose a novel method named Weighted Prototypical Contrastive Learning for Medical Few Shot Named Entity Recognization (W-PROCER). Our approach primarily revolves around constructing the prototype-based contractive loss and weighting network. These components play a crucial role in assisting the model in differentiating the negative samples from OUTSIDE (O) tokens and enhancing the discrimination ability of contrastive learning. Experimental results show that our proposed W-PROCER framework significantly outperforms the strong baselines on the three medical benchmark datasets.

Yi Nian, Xinyue Hu, Rui Zhang et al.

To date, there are no effective treatments for most neurodegenerative diseases. Knowledge graphs can provide comprehensive and semantic representation for heterogeneous data, and have been successfully leveraged in many biomedical applications including drug repurposing. Our objective is to construct a knowledge graph from literature to study relations between Alzheimer's disease (AD) and chemicals, drugs and dietary supplements in order to identify opportunities to prevent or delay neurodegenerative progression. We collected biomedical annotations and extracted their relations using SemRep via SemMedDB. We used both a BERT-based classifier and rule-based methods during data preprocessing to exclude noise while preserving most AD-related semantic triples. The 1,672,110 filtered triples were used to train with knowledge graph completion algorithms (i.e., TransE, DistMult, and ComplEx) to predict candidates that might be helpful for AD treatment or prevention. Among three knowledge graph completion models, TransE outperformed the other two (MR = 13.45, Hits@1 = 0.306). We leveraged the time-slicing technique to further evaluate the prediction results. We found supporting evidence for most highly ranked candidates predicted by our model which indicates that our approach can inform reliable new knowledge. This paper shows that our graph mining model can predict reliable new relationships between AD and other entities (i.e., dietary supplements, chemicals, and drugs). The knowledge graph constructed can facilitate data-driven knowledge discoveries and the generation of novel hypotheses.

Sijia Liu, Andrew Wen, Liwei Wang et al.

While we pay attention to the latest advances in clinical natural language processing (NLP), we can notice some resistance in the clinical and translational research community to adopt NLP models due to limited transparency, interpretability, and usability. In this study, we proposed an open natural language processing development framework. We evaluated it through the implementation of NLP algorithms for the National COVID Cohort Collaborative (N3C). Based on the interests in information extraction from COVID-19 related clinical notes, our work includes 1) an open data annotation process using COVID-19 signs and symptoms as the use case, 2) a community-driven ruleset composing platform, and 3) a synthetic text data generation workflow to generate texts for information extraction tasks without involving human subjects. The corpora were derived from texts from three different institutions (Mayo Clinic, University of Kentucky, University of Minnesota). The gold standard annotations were tested with a single institution's (Mayo) ruleset. This resulted in performances of 0.876, 0.706, and 0.694 in F-scores for Mayo, Minnesota, and Kentucky test datasets, respectively. The study as a consortium effort of the N3C NLP subgroup demonstrates the feasibility of creating a federated NLP algorithm development and benchmarking platform to enhance multi-institution clinical NLP study and adoption. Although we use COVID-19 as a use case in this effort, our framework is general enough to be applied to other domains of interest in clinical NLP.

Sarkar Snigdha Sarathi Das, Arzoo Katiyar, Rebecca J. Passonneau et al.

Named Entity Recognition (NER) in Few-Shot setting is imperative for entity tagging in low resource domains. Existing approaches only learn class-specific semantic features and intermediate representations from source domains. This affects generalizability to unseen target domains, resulting in suboptimal performances. To this end, we present CONTaiNER, a novel contrastive learning technique that optimizes the inter-token distribution distance for Few-Shot NER. Instead of optimizing class-specific attributes, CONTaiNER optimizes a generalized objective of differentiating between token categories based on their Gaussian-distributed embeddings. This effectively alleviates overfitting issues originating from training domains. Our experiments in several traditional test domains (OntoNotes, CoNLL'03, WNUT '17, GUM) and a new large scale Few-Shot NER dataset (Few-NERD) demonstrate that on average, CONTaiNER outperforms previous methods by 3%-13% absolute F1 points while showing consistent performance trends, even in challenging scenarios where previous approaches could not achieve appreciable performance.

Sicheng Zhou, Liwei Wang, Nan Wang et al.

Accurate extraction of breast cancer patients' phenotypes is important for clinical decision support and clinical research. Current models do not take full advantage of cancer domain-specific corpus, whether pre-training Bidirectional Encoder Representations from Transformer model on cancer-specific corpus could improve the performances of extracting breast cancer phenotypes from texts data remains to be explored. The objective of this study is to develop and evaluate the CancerBERT model for extracting breast cancer phenotypes from clinical texts in electronic health records. This data used in the study included 21,291 breast cancer patients diagnosed from 2010 to 2020, patients' clinical notes and pathology reports were collected from the University of Minnesota Clinical Data Repository (UMN). Results: About 3 million clinical notes and pathology reports in electronic health records for 21,291 breast cancer patients were collected to train the CancerBERT model. 200 pathology reports and 50 clinical notes of breast cancer patients that contain 9,685 sentences and 221,356 tokens were manually annotated by two annotators. 20% of the annotated data was used as a test set. Our CancerBERT model achieved the best performance with macro F1 scores equal to 0.876 (95% CI, 0.896-0.902) for exact match and 0.904 (95% CI, 0.896-0.902) for the lenient match. The NER models we developed would facilitate the automated information extraction from clinical texts to further help clinical decision support. Conclusions and Relevance: In this study, we focused on the breast cancer-related concepts extraction from EHR data and obtained a comprehensive annotated dataset that contains 7 types of breast cancer-related concepts. The CancerBERT model with customized vocabulary could significantly improve the performance for extracting breast cancer phenotypes from clinical texts.

Dalton Schutte, Jake Vasilakes, Anu Bompelli et al.

OBJECTIVE: Leverage existing biomedical NLP tools and DS domain terminology to produce a novel and comprehensive knowledge graph containing dietary supplement (DS) information for discovering interactions between DS and drugs, or Drug-Supplement Interactions (DSI). MATERIALS AND METHODS: We created SemRepDS (an extension of SemRep), capable of extracting semantic relations from abstracts by leveraging a DS-specific terminology (iDISK) containing 28,884 DS terms not found in the UMLS. PubMed abstracts were processed using SemRepDS to generate semantic relations, which were then filtered using a PubMedBERT-based model to remove incorrect relations before generating our knowledge graph (SuppKG). Two pathways are used to identify potential DS-Drug interactions which are then evaluated by medical professionals for mechanistic plausibility. RESULTS: Comparison analysis found that SemRepDS returned 206.9% more DS relations and 158.5% more DS entities than SemRep. The fine-tuned BERT model obtained an F1 score of 0.8605 and removed 43.86% of the relations, improving the precision of the relations by 26.4% compared to pre-filtering. SuppKG consists of 2,928 DS-specific nodes. Manual review of findings identified 44 (88%) proposed DS-Gene-Drug and 32 (64%) proposed DS-Gene1-Function-Gene2-Drug pathways to be mechanistically plausible. DISCUSSION: The additional relations extracted using SemRepDS generated SuppKG that was used to find plausible DSI not found in the current literature. By the nature of the SuppKG, these interactions are unlikely to have been found using SemRep without the expanded DS terminology. CONCLUSION: We successfully extend SemRep to include DS information and produce SuppKG which can be used to find potential DS-Drug interactions.

Yefeng Wang, Yunpeng Zhao, Jiang Bian et al.

Objective: The objective of this study is to develop a deep learning pipeline to detect signals on dietary supplement-related adverse events (DS AEs) from Twitter. Material and Methods: We obtained 247,807 tweets ranging from 2012 to 2018 that mentioned both DS and AE. We annotated biomedical entities and relations on 2,000 randomly selected tweets. For the concept extraction task, we compared the performance of traditional word embeddings with SVM, CRF and LSTM-CRF classifiers to BERT models. For the relation extraction task, we compared GloVe vectors with CNN classifiers to BERT models. We chose the best performing models in each task to assemble an end-to-end deep learning pipeline to detect DS AE signals and compared the results to the known DS AEs from a DS knowledge base (i.e., iDISK). Results: In both tasks, the BERT-based models outperformed traditional word embeddings. The best performing concept extraction model is the BioBERT model that can identify supplement, symptom, and body organ entities with F1-scores of 0.8646, 0.8497, and 0.7104, respectively. The best performing relation extraction model is the BERT model that can identify purpose and AE relations with F1-scores of 0.8335 and 0.7538, respectively. The end-to-end pipeline was able to extract DS indication and DS AEs with an F1-score of 0.7459 and 0,7414, respectively. Comparing to the iDISK, we could find both known and novel DS-AEs. Conclusion: We have demonstrated the feasibility of detecting DS AE signals from Twitter with a BioBERT-based deep learning pipeline.

Esha Singh, Anu Bompelli, Ruyuan Wan et al.

Dietary supplements (DS) have been widely used by consumers, but the information around the efficacy and safety of DS is disparate or incomplete, thus creating barriers for consumers to find information effectively. Conversational agent (CA) systems have been applied to the healthcare domain, but there is no such a system to answer consumers regarding DS use, although widespread use of DS. In this study, we develop the first CA system for DS use

Rui Zhang, Dimitar Hristovski, Dalton Schutte et al.

Objective: To discover candidate drugs to repurpose for COVID-19 using literature-derived knowledge and knowledge graph completion methods. Methods: We propose a novel, integrative, and neural network-based literature-based discovery (LBD) approach to identify drug candidates from both PubMed and COVID-19-focused research literature. Our approach relies on semantic triples extracted using SemRep (via SemMedDB). We identified an informative subset of semantic triples using filtering rules and an accuracy classifier developed on a BERT variant, and used this subset to construct a knowledge graph. Five SOTA, neural knowledge graph completion algorithms were used to predict drug repurposing candidates. The models were trained and assessed using a time slicing approach and the predicted drugs were compared with a list of drugs reported in the literature and evaluated in clinical trials. These models were complemented by a discovery pattern-based approach. Results: Accuracy classifier based on PubMedBERT achieved the best performance (F1= 0.854) in classifying semantic predications. Among five knowledge graph completion models, TransE outperformed others (MR = 0.923, Hits@1=0.417). Some known drugs linked to COVID-19 in the literature were identified, as well as some candidate drugs that have not yet been studied. Discovery patterns enabled generation of plausible hypotheses regarding the relationships between the candidate drugs and COVID-19. Among them, five highly ranked and novel drugs (paclitaxel, SB 203580, alpha 2-antiplasmin, pyrrolidine dithiocarbamate, and butylated hydroxytoluene) with their mechanistic explanations were further discussed. Conclusion: We show that an LBD approach can be feasible for discovering drug candidates for COVID-19, and for generating mechanistic explanations. Our approach can be generalized to other diseases as well as to other clinical questions.

Yadan Fan, Sicheng Zhou, Yifan Li et al.

The objective of our work is to demonstrate the feasibility of utilizing deep learning models to extract safety signals related to the use of dietary supplements (DS) in clinical text. Two tasks were performed in this study. For the named entity recognition (NER) task, Bi-LSTM-CRF (Bidirectional Long-Short-Term-Memory Conditional Random Fields) and BERT (Bidirectional Encoder Representations from Transformers) models were trained and compared with CRF model as a baseline to recognize the named entities of DS and Events from clinical notes. In the relation extraction (RE) task, two deep learning models, including attention-based Bi-LSTM and CNN (Convolutional Neural Network), and a random forest model were trained to extract the relations between DS and Events, which were categorized into three classes: positive (i.e., indication), negative (i.e., adverse events), and not related. The best performed NER and RE models were further applied on clinical notes mentioning 88 DS for discovering DS adverse events and indications, which were compared with a DS knowledge base. For the NER task, deep learning models achieved a better performance than CRF, with F1 scores above 0.860. The attention-based Bi-LSTM model performed the best in the relation extraction task, with the F1 score of 0.893. When comparing DS event pairs generated by the deep learning models with the knowledge base for DS and Event, we found both known and unknown pairs. Deep learning models can detect adverse events and indication of DS in clinical notes, which hold great potential for monitoring the safety of DS use.

Michihiro Yasunaga, Jungo Kasai, Rui Zhang et al.

Scientific article summarization is challenging: large, annotated corpora are not available, and the summary should ideally include the article's impacts on research community. This paper provides novel solutions to these two challenges. We 1) develop and release the first large-scale manually-annotated corpus for scientific papers (on computational linguistics) by enabling faster annotation, and 2) propose summarization methods that integrate the authors' original highlights (abstract) and the article's actual impacts on the community (citations), to create comprehensive, hybrid summaries. We conduct experiments to demonstrate the efficacy of our corpus in training data-driven models for scientific paper summarization and the advantage of our hybrid summaries over abstracts and traditional citation-based summaries. Our large annotated corpus and hybrid methods provide a new framework for scientific paper summarization research.

Catherine Finegan-Dollak, Jonathan K. Kummerfeld, Li Zhang et al.

To be informative, an evaluation must measure how well systems generalize to realistic unseen data. We identify limitations of and propose improvements to current evaluations of text-to-SQL systems. First, we compare human-generated and automatically generated questions, characterizing properties of queries necessary for real-world applications. To facilitate evaluation on multiple datasets, we release standardized and improved versions of seven existing datasets and one new text-to-SQL dataset. Second, we show that the current division of data into training and test sets measures robustness to variations in the way questions are asked, but only partially tests how well systems generalize to new queries; therefore, we propose a complementary dataset split for evaluation of future work. Finally, we demonstrate how the common practice of anonymizing variables during evaluation removes an important challenge of the task. Our observations highlight key difficulties, and our methodology enables effective measurement of future development.

Tao Yu, Zifan Li, Zilin Zhang et al.

Interacting with relational databases through natural language helps users of any background easily query and analyze a vast amount of data. This requires a system that understands users' questions and converts them to SQL queries automatically. In this paper we present a novel approach, TypeSQL, which views this problem as a slot filling task. Additionally, TypeSQL utilizes type information to better understand rare entities and numbers in natural language questions. We test this idea on the WikiSQL dataset and outperform the prior state-of-the-art by 5.5% in much less time. We also show that accessing the content of databases can significantly improve the performance when users' queries are not well-formed. TypeSQL gets 82.6% accuracy, a 17.5% absolute improvement compared to the previous content-sensitive model.

Rui Zhang, Honglak Lee, Lazaros Polymenakos et al.

In this paper, we study the problem of addressee and response selection in multi-party conversations. Understanding multi-party conversations is challenging because of complex speaker interactions: multiple speakers exchange messages with each other, playing different roles (sender, addressee, observer), and these roles vary across turns. To tackle this challenge, we propose the Speaker Interaction Recurrent Neural Network (SI-RNN). Whereas the previous state-of-the-art system updated speaker embeddings only for the sender, SI-RNN uses a novel dialog encoder to update speaker embeddings in a role-sensitive way. Additionally, unlike the previous work that selected the addressee and response separately, SI-RNN selects them jointly by viewing the task as a sequence prediction problem. Experimental results show that SI-RNN significantly improves the accuracy of addressee and response selection, particularly in complex conversations with many speakers and responses to distant messages many turns in the past.