Contrastive learning has been used to learn a high-quality representation of the image in computer vision. However, contrastive learning is not widely utilized in natural language processing due to the lack of a general method of data augmentation for text data. In this work, we explore the method of employing contrastive learning to improve the text representation from the BERT model for relation extraction. The key knob of our framework is a unique contrastive pre-training step tailored for the relation extraction tasks by seamlessly integrating linguistic knowledge into the data augmentation. Furthermore, we investigate how large-scale data constructed from the external knowledge bases can enhance the generality of contrastive pre-training of BERT. The experimental results on three relation extraction benchmark datasets demonstrate that our method can improve the BERT model representation and achieve state-of-the-art performance. In addition, we explore the interpretability of models by showing that BERT with contrastive pre-training relies more on rationales for prediction. Our code and data are publicly available at: https://github.com/AnonymousForNow.

Concept normalization, the task of linking textual mentions of concepts to concepts in an ontology, is critical for mining and analyzing biomedical texts. We propose a vector-space model for concept normalization, where mentions and concepts are encoded via transformer networks that are trained via a triplet objective with online hard triplet mining. The transformer networks refine existing pre-trained models, and the online triplet mining makes training efficient even with hundreds of thousands of concepts by sampling training triples within each mini-batch. We introduce a variety of strategies for searching with the trained vector-space model, including approaches that incorporate domain-specific synonyms at search time with no model retraining. Across five datasets, our models that are trained only once on their corresponding ontologies are within 3 points of state-of-the-art models that are retrained for each new domain. Our models can also be trained for each domain, achieving new state-of-the-art on multiple datasets.

Recent research on robust representations of biomedical names has focused on modeling large amounts of fine-grained conceptual distinctions using complex neural encoders. In this paper, we explore the opposite paradigm: training a simple encoder architecture using only small sets of names sampled from high-level biomedical concepts. Our encoder post-processes pretrained representations of biomedical names, and is effective for various types of input representations, both domain-specific or unsupervised. We validate our proposed few-shot learning approach on multiple biomedical relatedness benchmarks, and show that it allows for continual learning, where we accumulate information from various conceptual hierarchies to consistently improve encoder performance. Given these findings, we propose our approach as a low-cost alternative for exploring the impact of conceptual distinctions on robust biomedical name representations.

The accelerating growth of big data in the biomedical domain, with an endless amount of electronic health records and more than 30 million citations and abstracts in PubMed, introduces the need for automatic structuring of textual biomedical data. In this paper, we develop a method for detecting relations between food and disease entities from raw text. Due to the lack of annotated data on food with respect to health, we explore the feasibility of transfer learning by training BERT-based models on existing datasets annotated for the presence of cause and treat relations among different types of biomedical entities, and using them to recognize the same relations between food and disease entities in a dataset created for the purposes of this study. The best models achieve macro averaged F1 scores of 0.847 and 0.900 for the cause and treat relations, respectively.

We explore whether state-of-the-art BERT models encode sufficient domain knowledge to correctly perform domain-specific inference. Although BERT implementations such as BioBERT are better at domain-based reasoning than those trained on general-domain corpora, there is still a wide margin compared to human performance on these tasks. To bridge this gap, we explore whether supplementing textual domain knowledge in the medical NLI task: a) by further language model pretraining on the medical domain corpora, b) by means of lexical match algorithms such as the BM25 algorithm, c) by supplementing lexical retrieval with dependency relations, or d) by using a trained retriever module, can push this performance closer to that of humans. We do not find any significant difference between knowledge supplemented classification as opposed to the baseline BERT models, however. This is contrary to the results for evidence retrieval on other tasks such as open domain question answering (QA). By examining the retrieval output, we show that the methods fail due to unreliable knowledge retrieval for complex domain-specific reasoning. We conclude that the task of unsupervised text retrieval to bridge the gap in existing information to facilitate inference is more complex than what the state-of-the-art methods can solve, and warrants extensive research in the future.

Automatic ICD coding is the task of assigning codes from the International Classification of Diseases (ICD) to medical notes. These codes describe the state of the patient and have multiple applications, e.g., computer-assisted diagnosis or epidemiological studies. ICD coding is a challenging task due to the complexity and length of medical notes. Unlike the general trend in language processing, no transformer model has been reported to reach high performance on this task. Here, we investigate in detail ICD coding using PubMedBERT, a state-of-the-art transformer model for biomedical language understanding. We find that the difficulty of fine-tuning the model on long pieces of text is the main limitation for BERT-based models on ICD coding. We run extensive experiments and show that despite the gap with current state-of-the-art, pretrained transformers can reach competitive performance using relatively small portions of text. We point at better methods to aggregate information from long texts as the main need for improving BERT-based ICD coding.

We present emrKBQA, a dataset for answering physician questions from a structured patient record. It consists of questions, logical forms and answers. The questions and logical forms are generated based on real-world physician questions and are slot-filled and answered from patients in the MIMIC-III KB through a semi-automated process. This community-shared release consists of over 940000 question, logical form and answer triplets with 389 types of questions and ~7.5 paraphrases per question type. We perform experiments to validate the quality of the dataset and set benchmarks for question to logical form learning that helps answer questions on this dataset.

The MEDIQA 2021 shared tasks at the BioNLP 2021 workshop addressed three tasks on summarization for medical text: (i) a question summarization task aimed at exploring new approaches to understanding complex real-world consumer health queries, (ii) a multi-answer summarization task that targeted aggregation of multiple relevant answers to a biomedical question into one concise and relevant answer, and (iii) a radiology report summarization task addressing the development of clinically relevant impressions from radiology report findings. Thirty-five teams participated in these shared tasks with sixteen working notes submitted (fifteen accepted) describing a wide variety of models developed and tested on the shared and external datasets. In this paper, we describe the tasks, the datasets, the models and techniques developed by various teams, the results of the evaluation, and a study of correlations among various summarization evaluation measures. We hope that these shared tasks will bring new research and insights in biomedical text summarization and evaluation.

This paper describes our contribution for the MEDIQA-2021 Task 1 question summarization competition. We model the task as conditional generation problem. Our concrete pipeline performs a finetuning of the large pretrained generative transformers PEGASUS (Zhang et al.,2020a) and BART (Lewis et al.,2020). We used the resulting models as strong baselines and experimented with (i) integrating structured knowledge via entity embeddings, (ii) ensembling multiple generative models with the generator-discriminator framework and (iii) disentangling summarization and interrogative prediction to achieve further improvements. Our best performing model, a fine-tuned vanilla PEGASUS, reached the second place in the competition with an ROUGE-2-F1 score of 15.99. We observed that all of our additional measures hurt performance (up to 5.2 pp) on the official test set. In course of a post-hoc experimental analysis which uses a larger validation set results indicate slight performance improvements through the proposed extensions. However, further analysis is need to provide stronger evidence.

In this article, we describe our systems for the MEDIQA 2021 Shared Tasks. First, we will describe our method for the second task, Multi-Answer Summarization (MAS). For extractive summarization, two series of methods are applied. The first one follows (CITATION). First a RoBERTa model is first applied to give a local ranking of the candidate sentences. Then a Markov Chain model is applied to evaluate the sentences globally. The second method applies cross-sentence contextualization to improve the local ranking and discard the global ranking step. Our methods achieve the 1st Place in the MAS task. For the question summarization (QS) and radiology report summarization (RRS) tasks, we explore how end-to-end pre-trained seq2seq model perform. A series of tricks for improving the fine-tuning performances are validated.

This paper presents our winning system at the Radiology Report Summarization track of the MEDIQA 2021 shared task. Radiology report summarization automatically summarizes radiology findings into free-text impressions. This year’s task emphasizes the generalization and transfer ability of participating systems. Our system is built upon a pre-trained Transformer encoder-decoder architecture, i.e., PEGASUS, deployed with an additional domain adaptation module to particularly handle the transfer and generalization issue. Heuristics like ensemble and text normalization are also used. Our system is conceptually simple yet highly effective, achieving a ROUGE-2 score of 0.436 on test set and ranked the 1st place among all participating systems.

Medical question summarization is an important but difficult task, where the input is often complex and erroneous while annotated data is expensive to acquire. We report our participation in the MEDIQA 2021 question summarization task in which we are required to address these challenges. We start from pre-trained conditional generative language models, use knowledge bases to help correct input errors, and rerank single system outputs to boost coverage. Experimental results show significant improvement in string-based metrics.

The success of pretrained word embeddings has motivated their use in the biomedical domain, with contextualized embeddings yielding remarkable results in several biomedical NLP tasks. However, there is a lack of research on quantifying their behavior under severe “stress” scenarios. In this work, we systematically evaluate three language models with adversarial examples – automatically constructed tests that allow us to examine how robust the models are. We propose two types of stress scenarios focused on the biomedical named entity recognition (NER) task, one inspired by spelling errors and another based on the use of synonyms for medical terms. Our experiments with three benchmarks show that the performance of the original models decreases considerably, in addition to revealing their weaknesses and strengths. Finally, we show that adversarial training causes the models to improve their robustness and even to exceed the original performance in some cases.

NLP has emerged as an essential tool to extract knowledge from the exponentially increasing volumes of biomedical texts. Many NLP tasks, such as named entity recognition and named entity normalization, are especially challenging in the biomedical domain partly because of the prolific use of acronyms. Long names for diseases, bacteria, and chemicals are often replaced by acronyms. We propose Biomedical Local Acronym Resolver (BLAR), a high-performing acronym resolver that leverages state-of-the-art (SOTA) pre-trained language models to accurately resolve local acronyms in biomedical texts. We test BLAR on the Ab3P corpus and achieve state-of-the-art results compared to the current best-performing local acronym resolution algorithms and models.

Social media contains unfiltered and unique information, which is potentially of great value, but, in the case of misinformation, can also do great harm. With regards to biomedical topics, false information can be particularly dangerous. Methods of automatic fact-checking and fake news detection address this problem, but have not been applied to the biomedical domain in social media yet. We aim to fill this research gap and annotate a corpus of 1200 tweets for implicit and explicit biomedical claims (the latter also with span annotations for the claim phrase). With this corpus, which we sample to be related to COVID-19, measles, cystic fibrosis, and depression, we develop baseline models which detect tweets that contain a claim automatically. Our analyses reveal that biomedical tweets are densely populated with claims (45 % in a corpus sampled to contain 1200 tweets focused on the domains mentioned above). Baseline classification experiments with embedding-based classifiers and BERT-based transfer learning demonstrate that the detection is challenging, however, shows acceptable performance for the identification of explicit expressions of claims. Implicit claim tweets are more challenging to detect.

Recent advancements in pretraining strategies in NLP have shown a significant improvement in the performance of models on various text mining tasks. We apply ‘replaced token detection’ pretraining technique proposed by ELECTRA and pretrain a biomedical language model from scratch using biomedical text and vocabulary. We introduce BioELECTRA, a biomedical domain-specific language encoder model that adapts ELECTRA for the Biomedical domain. WE evaluate our model on the BLURB and BLUE biomedical NLP benchmarks. BioELECTRA outperforms the previous models and achieves state of the art (SOTA) on all the 13 datasets in BLURB benchmark and on all the 4 Clinical datasets from BLUE Benchmark across 7 different NLP tasks. BioELECTRA pretrained on PubMed and PMC full text articles performs very well on Clinical datasets as well. BioELECTRA achieves new SOTA 86.34%(1.39% accuracy improvement) on MedNLI and 64% (2.98% accuracy improvement) on PubMedQA dataset.

To keep pace with the increased generation and digitization of documents, automated methods that can improve search, discovery and mining of the vast body of literature are essential. Keyphrases provide a concise representation by identifying salient concepts in a document. Various supervised approaches model keyphrase extraction using local context to predict the label for each token and perform much better than the unsupervised counterparts. Unfortunately, this method fails for short documents where the context is unclear. Moreover, keyphrases, which are usually the gist of a document, need to be the central theme. We propose a new extraction model that introduces a centrality constraint to enrich the word representation of a Bidirectional long short-term memory. Performance evaluation on 2 publicly available datasets demonstrate our model outperforms existing state-of-the art approaches.

Disease name recognition and normalization is a fundamental process in biomedical text mining. Recently, neural joint learning of both tasks has been proposed to utilize the mutual benefits. While this approach achieves high performance, disease concepts that do not appear in the training dataset cannot be accurately predicted. This study introduces a novel end-to-end approach that combines span representations with dictionary-matching features to address this problem. Our model handles unseen concepts by referring to a dictionary while maintaining the performance of neural network-based models. Experiments using two major datasaets demonstrate that our model achieved competitive results with strong baselines, especially for unseen concepts during training.

We describe a simple procedure for the automatic creation of word-level alignments between printed documents and their respective full-text versions. The procedure is unsupervised, uses standard, off-the-shelf components only, and reaches an F-score of 85.01 in the basic setup and up to 86.63 when using pre- and post-processing. Potential areas of application are manual database curation (incl. document triage) and biomedical expression OCR.

Pretrained language models have shown success in many natural language processing tasks. Many works explore to incorporate the knowledge into the language models. In the biomedical domain, experts have taken decades of effort on building large-scale knowledge bases. For example, UMLS contains millions of entities with their synonyms and defines hundreds of relations among entities. Leveraging this knowledge can benefit a variety of downstream tasks such as named entity recognition and relation extraction. To this end, we propose KeBioLM, a biomedical pretrained language model that explicitly leverages knowledge from the UMLS knowledge bases. Specifically, we extract entities from PubMed abstracts and link them to UMLS. We then train a knowledge-aware language model that firstly applies a text-only encoding layer to learn entity representation and then applies a text-entity fusion encoding to aggregate entity representation. In addition, we add two training objectives as entity detection and entity linking. Experiments on the named entity recognition and relation extraction tasks from the BLURB benchmark demonstrate the effectiveness of our approach. Further analysis on a collected probing dataset shows that our model has better ability to model medical knowledge.

Transformer-based neural language models have led to breakthroughs for a variety of natural language processing (NLP) tasks. However, most models are pretrained on general domain data. We propose a methodology to produce a model focused on the clinical domain: continued pretraining of a model with a broad representation of biomedical terminology (PubMedBERT) on a clinical corpus along with a novel entity-centric masking strategy to infuse domain knowledge in the learning process. We show that such a model achieves superior results on clinical extraction tasks by comparing our entity-centric masking strategy with classic random masking on three clinical NLP tasks: cross-domain negation detection, document time relation (DocTimeRel) classification, and temporal relation extraction. We also evaluate our models on the PubMedQA dataset to measure the models’ performance on a non-entity-centric task in the biomedical domain. The language addressed in this work is English.

Several previous studies on explanation for recurrent neural networks focus on approaches that find the most important input segments for a network as its explanations. In that case, the manner in which these input segments combine with each other to form an explanatory pattern remains unknown. To overcome this, some previous work tries to find patterns (called rules) in the data that explain neural outputs. However, their explanations are often insensitive to model parameters, which limits the scalability of text explanations. To overcome these limitations, we propose a pipeline to explain RNNs by means of decision lists (also called rules) over skipgrams. For evaluation of explanations, we create a synthetic sepsis-identification dataset, as well as apply our technique on additional clinical and sentiment analysis datasets. We find that our technique persistently achieves high explanation fidelity and qualitatively interpretable rules.

Chinese word segmentation (CWS) and medical concept recognition are two fundamental tasks to process Chinese electronic medical records (EMRs) and play important roles in downstream tasks for understanding Chinese EMRs. One challenge to these tasks is the lack of medical domain datasets with high-quality annotations, especially medical-related tags that reveal the characteristics of Chinese EMRs. In this paper, we collected a Chinese EMR corpus, namely, ACEMR, with human annotations for Chinese word segmentation and EMR-related tags. On the ACEMR corpus, we run well-known models (i.e., BiLSTM, BERT, and ZEN) and existing state-of-the-art systems (e.g., WMSeg and TwASP) for CWS and medical concept recognition. Experimental results demonstrate the necessity of building a dedicated medical dataset and show that models that leverage extra resources achieve the best performance for both tasks, which provides certain guidance for future studies on model selection in the medical domain.

The impact of design choices on the performance of biomedical language models recently has been a subject for investigation. In this paper, we empirically study biomedical domain adaptation with large transformer models using different design choices. We evaluate the performance of our pretrained models against other existing biomedical language models in the literature. Our results show that we achieve state-of-the-art results on several biomedical domain tasks despite using similar or less computational cost compared to other models in the literature. Our findings highlight the significant effect of design choices on improving the performance of biomedical language models.

First-hand experience related to any changes of one’s health condition and understanding such experience can play an important role in advancing medical science and healthcare. Monitoring the safe use of medication drugs is an important task of pharmacovigilance, and first-hand experience of effects about consumers’ medication intake can be valuable to gain insight into how our human body reacts to medications. Social media have been considered as a possible alternative data source for gathering personal experience with medications posted by users. Identifying personal experience tweets is a challenging classification task, and efforts have made to tackle the challenges using supervised approaches requiring annotated data. There exists abundance of unlabeled Twitter data, and being able to use such data for training without suffering in classification performance is of great value, which can reduce the cost of laborious annotation process. We investigated two semi-supervised learning methods, with different mixes of labeled and unlabeled data in the training set, to understand the impact on classification performance. Our results from both pseudo-label and consistency regularization methods show that both methods generated a noticeable improvement in F1 score when the labeled set was small, and consistency regularization could still provide a small gain even a larger labeled set was used.

The amount of biomedical literature has vastly increased over the past few decades. As a result, the sheer quantity of accessible information is overwhelming, and complicates manual information retrieval. Automated methods seek to speed up information retrieval from biomedical literature. However, such automated methods are still too time-intensive to survey all existing biomedical literature. We present a methodology for automatically generating literature queries that select relevant papers based on biological data. By using differentially expressed genes to inform our literature searches, we focus information extraction on mechanistic signaling details that are crucial for the disease or context of interest.

With the growing availability of full-text articles, integrating abstracts and full texts of documents into a unified representation is essential for comprehensive search of scientific literature. However, previous studies have shown that naïvely merging abstracts with full texts of articles does not consistently yield better performance. Balancing the contribution of query terms appearing in the abstract and in sections of different importance in full text articles remains a challenge both with traditional bag-of-words IR approaches and for neural retrieval methods. In this work we establish the connection between the BM25 score of a query term appearing in a section of a full text document and the probability of that document being clicked or identified as relevant. Probability is computed using Pool Adjacent Violators (PAV), an isotonic regression algorithm, providing a maximum likelihood estimate based on the observed data. Using this probabilistic transformation of BM25 scores we show an improved performance on the PubMed Click dataset developed and presented in this study, as well as the 2007 TREC Genomics collection.

In this paper, we describe our approach to question summarization and multi-answer summarization in the context of the 2021 MEDIQA shared task (Ben Abacha et al., 2021). We propose two kinds of transfer learning for the abstractive summarization of medical questions. First, we train on HealthCareMagic, a large question summarization dataset collected from an online healthcare service platform. Second, we leverage the ability of the BART encoder-decoder architecture to model both generation and classification tasks to train on the task of Recognizing Question Entailment (RQE) in the medical domain. We show that both transfer learning methods combined achieve the highest ROUGE scores. Finally, we cast the question-driven extractive summarization of multiple relevant answer documents as an Answer Sentence Selection (AS2) problem. We show how we can preprocess the MEDIQA-AnS dataset such that it can be trained in an AS2 setting. Our AS2 model is able to generate extractive summaries achieving high ROUGE scores.

In this article, we will describe our system for MEDIQA2021 shared tasks. First, we will describe the method of the second task, multiple answer summary (MAS). For extracting abstracts, we follow the rules of (CITATION). First, the candidate sentences are roughly estimated by using the Roberta model. Then the Markov chain model is used to evaluate the sentences in a fine-grained manner. Our team won the first place in overall performance, with the fourth place in MAS task, the seventh place in RRS task and the eleventh place in QS task. For the QS and RRS tasks, we investigate the performanceS of the end-to-end pre-trained seq2seq model. Experiments show that the methods of adversarial training and reverse translation are beneficial to improve the fine tuning performance.

This study describes the model design of the NCUEE-NLP system for the MEDIQA challenge at the BioNLP 2021 workshop. We use the PEGASUS transformers and fine-tune the downstream summarization task using our collected and processed datasets. A total of 22 teams participated in the consumer health question summarization task of MEDIQA 2021. Each participating team was allowed to submit a maximum of ten runs. Our best submission, achieving a ROUGE2-F1 score of 0.1597, ranked third among all 128 submissions.

Recent strides in the healthcare domain, have resulted in vast quantities of streaming data available for use for building intelligent knowledge-based applications. However, the challenges introduced to the huge volume, velocity of generation, variety and variability of this medical data have to be adequately addressed. In this paper, we describe the model and results for our submission at MEDIQA 2021 Question Summarization shared task. In order to improve the performance of summarization of consumer health questions, our method explores the use of transfer learning to utilize the knowledge of NLP transformers like BART, T5 and PEGASUS. The proposed models utilize the knowledge of pre-trained NLP transformers to achieve improved results when compared to conventional deep learning models such as LSTM, RNN etc. Our team SB_NITK ranked 12th among the total 22 submissions in the official final rankings. Our BART based model achieved a ROUGE-2 F1 score of 0.139.

This paper describes experiments undertaken and their results as part of the BioNLP MEDIQA 2021 challenge. We participated in Task 3: Radiology Report Summarization. Multiple runs were submitted for evaluation, from solutions leveraging transfer learning from pre-trained transformer models, which were then fine tuned on a subset of MIMIC-CXR, for abstractive report summarization. The task was evaluated using ROUGE and our best performing system obtained a ROUGE-2 score of 0.392.

This paper describes the solution of the QIAI lab sent to the Radiology Report Summarization (RRS) challenge at MEDIQA 2021. This paper aims to investigate whether using multimodality during training improves the summarizing performances of the model at test-time. Our preliminary results shows that taking advantage of the visual features from the x-rays associated to the radiology reports leads to higher evaluation metrics compared to a text-only baseline system. These improvements are reported according to the automatic evaluation metrics METEOR, BLEU and ROUGE scores. Our experiments can be fully replicated at the following address: https://github.com/jbdel/vilmedic.

The quest for seeking health information has swamped the web with consumers’ healthrelated questions, which makes the need for efficient and reliable question answering systems more pressing. The consumers’ questions, however, are very descriptive and contain several peripheral information (like patient’s medical history, demographic information, etc.), that are often not required for answering the question. Furthermore, it contributes to the challenges of understanding natural language questions for automatic answer retrieval. Also, it is crucial to provide the consumers with the exact and relevant answers, rather than the entire pool of answer documents to their question. One of the cardinal tasks in achieving robust consumer health question answering systems is the question summarization and multi-document answer summarization. This paper describes the participation of the U.S. National Library of Medicine (NLM) in Consumer Question and Multi-Answer Summarization tasks of the MEDIQA 2021 challenge at NAACL-BioNLP workshop. In this work, we exploited the capabilities of pre-trained transformer models and introduced a transfer learning approach for the abstractive Question Summarization and extractive Multi-Answer Summarization tasks by first pre-training our model on a task-specific summarization dataset followed by fine-tuning it for both the tasks via incorporating medical entities. We achieved the second, sixth and the fourth position for the Question Summarization task in terms ROUGE-1, ROUGE-2 and ROUGE-L scores respectively.

Although recent advances in abstractive summarization systems have achieved high scores on standard natural language metrics like ROUGE, their lack of factual consistency remains an open challenge for their use in sensitive real-world settings such as clinical practice. In this work, we propose a novel approach to improve factual correctness of a summarization system by re-ranking the candidate summaries based on a factual vector of the summary. We applied this process during our participation in MEDIQA 2021 Task 3: Radiology Report Summarization, where the task is to generate an impression summary of a radiology report, given findings and background as inputs. In our system, we first used a transformer-based encoder-decoder model to generate top N candidate impression summaries for a report, then trained another transformer-based model to predict a 14-observations-vector of the impression based on the findings and background of the report, and finally, utilized this vector to re-rank the candidate summaries. We also employed a source-specific ensembling technique to accommodate for distinct writing styles from different radiology report sources. Our approach yielded 2nd place in the challenge.

This paper describes a system developed to summarize multiple answers challenge in the MEDIQA 2021 shared task collocated with the BioNLP 2021 Workshop. We propose an extractive summarization architecture based on several scores and state-of-the-art techniques. We also present our novel prosper-thy-neighbour strategies to improve performance. Our model has been proven to be effective with the best ROUGE-1/ROUGE-L scores, being the shared task runner up by ROUGE-2 F1 score (over 13 participated teams).

This paper details a Consumer Health Question (CHQ) summarization model submitted to MEDIQA 2021 for shared task 1: Question Summarization. Many CHQs are composed of multiple sentences with typos or unnecessary information, which can interfere with automated question answering systems. Question summarization mitigates this issue by removing this unnecessary information, aiding automated systems in generating a more accurate summary. Our summarization approach focuses on applying multiple pre-processing techniques, including question focus identification on the input and the development of an ensemble method to combine question focus with an abstractive summarization method. We use the state-of-art abstractive summarization model, PEGASUS (Pre-training with Extracted Gap-sentences for Abstractive Summarization), to generate abstractive summaries. Our experiments show that using our ensemble method, which combines abstractive summarization with question focus identification, improves performance over using summarization alone. Our model shows a ROUGE-2 F-measure of 11.14% against the official test dataset.

This paper describes a system developed to summarize multiple answers challenge in the MEDIQA 2021 shared task collocated with the BioNLP 2021 Workshop. We present an abstractive summarization model based on BART, a denoising auto-encoder for pre-training sequence-to-sequence models. As focusing on the summarization of answers to consumer health questions, we propose a query-driven filtering phase to choose useful information from the input document automatically. Our approach achieves potential results, rank no.2 (evaluated on extractive references) and no.3 (evaluated on abstractive references) in the final evaluation.