EVIDENCEMINER: Textual Evidence Discovery for Life Sciences

Abstract

Traditional search engines for life sciences (e.g., PubMed) are designed for document retrieval and do not allow direct retrieval of specific statements. Some of these statements may serve as textual evidence that is key to tasks such as hypothesis generation and new finding validation. We present EVIDENCEM-INER, a web-based system that lets users query a natural language statement and automatically retrieves textual evidence from a background corpora for life sciences. EVIDENCEMINER is constructed in a completely automated way without any human effort for training data annotation. It is supported by novel data-driven methods for distantly supervised named entity recognition and open information extraction. The entities and patterns are pre-computed and indexed offline to support fast online evidence retrieval. The annotation results are also highlighted in the original document for better visualization. EVIDENCEMINER also includes analytic functionalities such as the most frequent entity and relation summarization. EVIDENCEMINER can help scientists uncover essential research issues, leading to more effective research and more indepth quantitative analysis. The system of EVIDENCEMINER is available at https:// evidenceminer.firebaseapp.com/ 1 .

COVID-19-Related Scientific Literature Exploration: Short Survey and Comparative Study

“… Data Bias : Some applications (e.g., [ 54 ]) can also benefit from reducing data bias, especially gender bias. Smart Querying : Some applications [ 56 ] use query functionalities that tend to be limited to simple word matching. This can be problematic in cases where the intent of the user is not evident in the query.…”

“…Smart Querying : Some applications [ 56 ] use query functionalities that tend to be limited to simple word matching. This can be problematic in cases where the intent of the user is not evident in the query.…”

“… Granularity/Levels of Representations : We also noticed that the works used different levels of granularity, which depends on the intended tasks and the available computational resources. For example, to achieve the task of document retrieval, some works opted for simple document level representations [ 53 ], while other works either used more granular representations [ 12 , 32 , 37 , 40 , 50 , 54 , 55 , 56 ] or a mix of more granular representations with document level representations [ 16 , 24 , 38 , 39 , 44 , 51 , 52 , 57 ]. Using KGs : Knowledge graphs were used in multiple works for different purposes.…”

“…[ 38 ] used a KG ( CKG [ 36 ]) embedding in tandem with textual representations for document recommendation, while [ 37 ] ( CovEx KG [ 37 ]), [ 32 ] ( Vapur KG [ 32 ]) and [ 44 ] ( Citation KG [ 44 ]) traversed their respective KGs looking for similar entities to retrieve relevant papers. The authors of [ 56 ] ( Blender-KG [ 34 ]) used a KG to extend queries and make the search more efficient. Recommendation Modules : Many search engines [ 32 , 37 , 38 ] use recommendation modules to offer more user-oriented results.…”

Descriptive Knowledge Graph in Biomedical Domain

“…Document-level retrieval usually takes keyphrases (Shen et al, 2018) or questions (Voorhees et al, 2021;Levy et al, 2021) as queries and finds relevant documents. Using such systems, researchers need to read the retrieved documents to find relevant information, which is still time-consuming (Wang et al, 2020b). Sentence-level Wang et al, 2020b;Lahav et al, 2022) retrieval usually takes entities, entity types, or sentences as queries and finds sentences that contain the entities and entity types or are semantically similar to the input sentence.…”