We study the effect of seven data augmentation (DA) methods in factoid question answering, focusing on the biomedical domain, where obtaining training instances is particularly difficult. We experiment with data from the BIOASQ challenge, which we augment with training instances obtained from an artificial biomedical machine reading comprehension dataset, or via back-translation, information retrieval, word substitution based on WORD2VEC embeddings, or masked language modeling, question generation, or extending the given passage with additional context. We show that DA can lead to very significant performance gains, even when using large pre-trained Transformers, contributing to a broader discussion of if/when DA benefits large pre-trained models. One of the simplest DA methods, WORD2VEC-based word substitution, performed best and is recommended. We release our artificial training instances and code.

Question answering (QA) systems for large document collections typically use pipelines that (i) retrieve possibly relevant documents, (ii) re-rank them, (iii) rank paragraphs or other snippets of the top-ranked documents, and (iv) select spans of the top-ranked snippets as exact answers. Pipelines are conceptually simple, but errors propagate from one component to the next, without later components being able to revise earlier decisions. We present an architecture for joint document and snippet ranking, the two middle stages, which leverages the intuition that relevant documents have good snippets and good snippets come from relevant documents. The architecture is general and can be used with any neural text relevance ranker. We experiment with two main instantiations of the architecture, based on POSIT-DRMM (PDRMM) and a BERT-based ranker. Experiments on biomedical data from BIOASQ show that our joint models vastly outperform the pipelines in snippet retrieval, the main goal for QA, with fewer trainable parameters, also remaining competitive in document retrieval. Furthermore, our joint PDRMM-based model is competitive with BERT-based models, despite using orders of magnitude fewer parameters. These claims are also supported by human evaluation on two test batches of BIOASQ. To test our key findings on another dataset, we modified the Natural Questions dataset so that it can also be used for document and snippet retrieval. Our joint PDRMM-based model again outperforms the corresponding pipeline in snippet retrieval on the modified Natural Questions dataset, even though it performs worse than the pipeline in document retrieval. We make our code and the modified Natural Questions dataset publicly available.

Science, technology and innovation (STI) policies have evolved in the past decade. We are now progressing towards policies that are more aligned with sustainable development through integrating social, economic and environmental dimensions. In this new policy environment, the need to keep track of innovation from its conception in Science and Research has emerged. Argumentation mining, an interdisciplinary NLP field, gives rise to the required technologies. In this study, we present the first STI-driven multidisciplinary corpus of scientific abstracts annotated for argumentative units (AUs) on the sustainable development goals (SDGs) set by the United Nations (UN). AUs are the sentences conveying the Claim(s) reported in the author’s original research and the Evidence provided for support. We also present a set of strong, BERT-based neural baselines achieving an f1-score of 70.0 for Claim and 62.4 for Evidence identification evaluated with 10-fold cross-validation. To demonstrate the effectiveness of our models, we experiment with different test sets showing comparable performance across various SDG policy domains. Our dataset and models are publicly available for research purposes.

Cat. 2 Show-case: We present the Data4Impact (D4I) platform, a novel end-to-end system for evidence-based, timely and accurate monitoring and evaluation of research and innovation (R&I) activities. Using the latest technological advances in Human Language Technology (HLT) and our data-driven methodology, we build a novel set of indicators in order to track funded projects and their impact on science, the economy and the society as a whole, during and after the project life-cycle. We develop our methodology by targeting Health-related EC projects from 2007 to 2019 to produce solutions that meet the needs of stakeholders (mainly policy-makers and research funders). Various D4I text analytics workflows process datasets and their metadata, extract valuable insights and estimate intermediate results and metrics, culminating in a set of robust indicators that the users can interact with through our dashboard, the D4I Monitor (available at monitor.data4impact.eu). Therefore, our approach, which can be generalized to different contexts, is multidimensional (technology, tools, indicators, dashboard) and the resulting system can provide an innovative solution for public administrators in their policy-making needs related to RDI funding allocation.

We introduceBIOMRC, a large-scale cloze-style biomedical MRC dataset. Care was taken to reduce noise, compared to the previous BIOREAD dataset of Pappas et al. (2018). Experiments show that simple heuristics do not perform well on the new dataset and that two neural MRC models that had been tested on BIOREAD perform much better on BIOMRC, indicating that the new dataset is indeed less noisy or at least that its task is more feasible. Non-expert human performance is also higher on the new dataset compared to BIOREAD, and biomedical experts perform even better. We also introduce a new BERT-based MRC model, the best version of which substantially outperforms all other methods tested, reaching or surpassing the accuracy of biomedical experts in some experiments. We make the new dataset available in three different sizes, also releasing our code, and providing a leaderboard.

Network Embedding (NE) methods, which map network nodes to low-dimensional feature vectors, have wide applications in network analysis and bioinformatics. Many existing NE methods rely only on network structure, overlooking other information associated with the nodes, e.g., text describing the nodes. Recent attempts to combine the two sources of information only consider local network structure. We extend NODE2VEC, a well-known NE method that considers broader network structure, to also consider textual node descriptors using recurrent neural encoders. Our method is evaluated on link prediction in two networks derived from UMLS. Experimental results demonstrate the effectiveness of the proposed approach compared to previous work.

We present AUEB’s submissions to the BioASQ 6 document and snippet retrieval tasks (parts of Task 6b, Phase A). Our models use novel extensions to deep learning architectures that operate solely over the text of the query and candidate document/snippets. Our systems scored at the top or near the top for all batches of the challenge, highlighting the effectiveness of deep learning for these tasks.