To combat COVID-19, both clinicians and scientists need to digest vast amounts of relevant biomedical knowledge in scientific literature to understand the disease mechanism and related biological functions. We have developed a novel and comprehensive knowledge discovery framework, COVID-KG to extract finegrained multimedia knowledge elements (entities and their visual chemical structures, relations and events) from scientific literature. We then exploit the constructed multimedia knowledge graphs (KGs) for question answering and report generation, using drug repurposing as a case study. Our framework also provides detailed contextual sentences, subfigures, and knowledge subgraphs as evidence. All of the data, KGs, reports 1 , resources, and shared services are publicly available 2 .

To combat COVID-19, both clinicians and scientists need to digest the vast amount of relevant biomedical knowledge in literature to understand the disease mechanism and the related biological functions. We have developed a novel and comprehensive knowledge discovery framework, COVID-KG to extract finegrained multimedia knowledge elements (entities, relations and events) from scientific literature. We then exploit the constructed multimedia knowledge graphs (KGs) for question answering and report generation, using drug repurposing as a case study. Our framework also provides detailed contextual sentences, subfigures and knowledge subgraphs as evidence. All of the data, KGs, reports 1 , resources and shared services are publicly available 2 .

Objective: The mismatch negativity (MMN) is considered as a promising biomarker that can inform future therapeutic studies. However, there is a large variability among patients with first episode psychosis (FEP). Also, most studies report a single electrode site and on comparing case-control group differences. Few have taken advantage of the full wealth of multi-channel EEG signals to examine observable patterns. None, to our knowledge, have used machine learning (ML) approaches to investigate neurophysiological derived subgroups with distinct cognitive and functional outcome characteristics. In this study, we applied ML to empirically stratify individuals into homogeneous subgroups based on multi-channel MMN data. We then characterized the functional, cognitive, and clinical profiles of these neurobiologically derived subgroups. We also explored the underlying low frequency range responses (delta, theta, alpha) during MMN. Methods: Clinical, neurocognitive, functioning data of 33 healthy controls and 20 FEP patients were collected. 90% of the patients had 6-month follow-up data. Neurocognition, social cognition, and functioning measures were assessed using the NCCB Cognitive Battery, the Awareness of Social Inference Test, UCSD Performance-Based Skills Assessment, and Multnomah Community Ability Scale. Symptom severity was collected using the PANSS. MMN amplitude and single-trial derived low frequency activity across 24 frontocentral channels were used as main variables in the ML kmeans clustering analyses.

No abstract

In this task, we identify a challenge that is reflective of linguistic and cognitive competencies that humans have when speaking and reasoning. Particularly, given the intuition that textual and visual information mutually inform each other for semantic reasoning, we formulate a Competence-based Question Answering challenge, designed to involve rich semantic annotation and aligned text-video objects. The task is to answer questions from a collection of English language cooking recipes and videos, where each question belongs to a "question family" reflecting a specific reasoning competence. The data and task result is publicly available. 1