The literature on coronaviruses counts more than 300,000 publications. Finding relevant papers concerning arbitrary queries is essential to discovery helpful knowledge. Current best information retrieval (IR) use deep learning approaches and need supervised train sets with labeled data, namely to know a priori the queries and their corresponding relevant papers. Creating such labeled datasets is time-expensive and requires prominent experts’ efforts, resources insufficiently available under a pandemic time pressure. We present a new self-supervised solution, called SUBLIMER, that does not require labels to learn to search on corpora of scientific papers for most relevant against arbitrary queries. SUBLIMER is a novel efficient IR engine trained on the unsupervised COVID-19 Open Research Dataset (CORD19), using deep metric learning. The core point of our self-supervised approach is that it uses no labels, but exploits the bibliography citations from papers to create a latent space where their spatial proximity is a metric of semantic similarity; for this reason, it can also be applied to other domains of papers corpora. SUBLIMER, despite is self-supervised, outperforms the Precision@5 (P@5) and Bpref of the state-of-the-art competitors on CORD19, which, differently from our approach, require both labeled datasets and a number of trainable parameters that is an order of magnitude higher than our.

Although current state-of-the-art Transformerbased solutions succeeded in a wide range for single-document NLP tasks, they still struggle to address multi-input tasks such as multidocument summarization. Many solutions truncate the inputs, thus ignoring potential summary-relevant contents, which is unacceptable in the medical domain where each information can be vital. Others leverage linear model approximations to apply multi-input concatenation, worsening the results because all information is considered, even if it is conflicting or noisy with respect to a shared background. Despite the importance and social impact of medicine, there are no ad-hoc solutions for multi-document summarization. For this reason, we propose a novel discriminative marginalized probabilistic method (DAMEN) trained to discriminate critical information from a cluster of topic-related medical documents and generate a multi-document summary via token probability marginalization. Results prove we outperform the previous state-of-theart on a biomedical dataset for multi-document summarization of systematic literature reviews. Moreover, we perform extensive ablation studies to motivate the design choices and prove the importance of each module of our method. 1

No abstract

Long document summarization poses obstacles to current generative transformer-based models because of the broad context to process and understand. Indeed, detecting long-range dependencies is still challenging for today’s state-of-the-art solutions, usually requiring model expansion at the cost of an unsustainable demand for computing and memory capacities. This paper introduces Emma, a novel efficient memory-enhanced transformer-based architecture. By segmenting a lengthy input into multiple text fragments, our model stores and compares the current chunk with previous ones, gaining the capability to read and comprehend the entire context over the whole document with a fixed amount of GPU memory. This method enables the model to deal with theoretically infinitely long documents, using less than 18 and 13 GB of memory for training and inference, respectively. We conducted extensive performance analyses and demonstrate that Emma achieved competitive results on two datasets of different domains while consuming significantly less GPU memory than competitors do, even in low-resource settings.

Generative transformer-based models have reached cutting-edge performance in long document summarization. Nevertheless, this task is witnessing a paradigm shift in developing ever-increasingly computationally-hungry solutions, focusing on effectiveness while ignoring the economic, environmental, and social costs of yielding such results. Accordingly, such extensive resources impact climate change and raise barriers to small and medium organizations distinguished by low-resource regimes of hardware and data. As a result, this unsustainable trend has lifted many concerns in the community, which directs the primary efforts on the proposal of tools to monitor models' energy costs. Despite their importance, no evaluation measure considering models' eco-sustainability exists yet. In this work, we propose Carburacy, the first carbon-aware accuracy measure that captures both model effectiveness and eco-sustainability. We perform a comprehensive benchmark for long document summarization, comparing multiple state-of-the-art quadratic and linear transformers on several datasets under eco-sustainable regimes. Finally, thanks to Carburacy, we found optimal combinations of hyperparameters that let models be competitive in effectiveness with significantly lower costs.