Challenges and Advances in Information Extraction from Scientific Literature: a Review

Abstract

Scientific articles have long been the primary means of disseminating scientific discoveries. Over the centuries, valuable data and potentially groundbreaking insights have been collected and buried deep in the mountain of publications. In materials engineering, such data are spread across technical handbooks specification sheets, journal articles, and laboratory notebooks in myriad formats. Extracting information from papers on a large scale has been a tedious and time-consuming job to which few researchers have wanted to devote their limited time and effort, yet is an activity that is essential for modern data-driven design practices. However, in recent years, significant progress has been made by the computer science community on techniques for automated information extraction from free text. Yet, transformative application of these techniques to scientific literature remains elusive-due not to a lack of interest or effort but to technical and logistical challenges. Using the challenges in the materials science literature as a driving motivation, we review the gaps between state-of-the-art information extraction methods and the practical application of such methods to scientific texts, and offer a comprehensive overview of work that can be undertaken to close these gaps.

Challenges in Information-Mining the Materials Literature: A Case Study and Perspective

“…While computers generally still struggle to grasp the meaning of a text fully, NLP algorithms can endow them with knowledge about the sentence’s grammar and sentiment. A more detailed description of NLP for materials science can be found in works by Olivetti and co-workers, ,, Cole and co-workers, and Hong and co-workers . A core aspect of NLP is tokenization, which consists of breaking a larger text down into smaller subunits (Figure ).…”

“…While there is rapid progress in the development of state-of-the-art text-mining algorithms for materials science, there remain several challenges . For instance, the MGI advocates for increased standardization of the reporting of data and metadatathe lack of which currently inhibits many text-mining activities .…”

“…While there is rapid progress in the development of state-ofthe-art text-mining algorithms for materials science, there remain several challenges. 13 For instance, the MGI advocates for increased standardization of the reporting of data and metadatathe lack of which currently inhibits many textmining activities. 14 This perspective illustrates common problems one may encounter in data collection from scientific publications by focusing on a case study of extracting crystallization solvents for organic materials.…”

DeepShovel: An Online Collaborative Platform for Data Extraction in Geoscience Literature with AI Assistance

“…With the development of artificial intelligence, more and more researchers pay attention to the data and phenomena in the published literature. Artificial intelligence methods are introduced in this process to help promote the collection of data from the published literature and build a scientific database in biology [29], chemistry [41,50], and many other disciplines [8,23,56]. In addition, there are also some related works focusing on literature retrieval and dissertation comprehension to extract information from scientific literature [32,52,53].…”

Uncertainty Informed Screening for Safer Solvents used in the Synthesis of Perovskite Based Solar Cells via Machine Learning

“…But this information hasn't been organized and it isn't easy to do so. The challenges on information extraction from the literature has been well documented in the article by Hong et al [32]. The ever-increasing number of scientific articles mean that different solvents and synthesis techniques are being discussed and it is difficult to keep track of them manually.…”