Assessing inclusion and representativeness on digital platforms for health education: Evidence from YouTube.

BACKGROUND: Studies confirm that significant biases exist in online recommendation platforms, exacerbating pre-existing disparities and leading to less-than-optimal outcomes for underrepresented demographics. We study issues of bias in inclusion and representativeness in the context of healthcare information disseminated via videos on the YouTube social media platform, a widely used online channel for multi-media rich information. With one in three US adults using the Internet to learn about a health concern, it is critical to assess inclusivity and representativeness regarding how health information is disseminated by digital platforms such as YouTube. METHODS: Leveraging methods from fair machine learning (ML), natural language processing and voice and facial recognition methods, we examine inclusivity and representativeness of video content presenters using a large corpus of videos and their metadata on a chronic condition (diabetes) extracted from the YouTube platform. Regression models are used to determine whether presenter demographics impact video popularity, measured by the video's average daily view count. A video that generates a higher view count is considered to be more popular. RESULTS: The voice and facial recognition methods predicted the gender and race of the presenter with reasonable success. Gender is predicted through voice recognition (accuracy = 78 %, AUC = 76 %), while the gender and race predictions use facial recognition (accuracy = 93 %, AUC = 92 % and accuracy = 82 %, AUC = 80 %, respectively). The gender of the presenter is more significant for video views only when the face of the presenter is not visible while videos with male presenters with no face visibility have a positive relationship with view counts. Furthermore, videos with white and male presenters have a positive influence on view counts while videos with female and non - white group have high view counts. CONCLUSION: Presenters' demographics do have an influence on average daily view count of videos viewed on social media platforms as shown by advanced voice and facial recognition algorithms used for assessing inclusion and representativeness of the video content. Future research can explore short videos and those at the channel level because popularity of the channel name and the number of videos associated with that channel do have an influence on view counts.

Age differences in the recruitment of syntactic analysis and semantic plausibility during sentence comprehension.

Syntactic analysis and semantic plausibility provide important cues to build the meaningful representation of sentences. The purpose of this research is to explore the age-related differences in the use of syntactic analysis and semantic plausibility during sentence comprehension under different working memory load conditions. A sentence judgment task was implemented among a group of older and younger adults. Semantic plausibility (plausible, implausible) and syntactic consistency (consistent, inconsistent) were manipulated in the experimental stimuli, and working memory load (high, low) was varied by manipulating the presentation of the stimuli. The study revealed a stronger effect of semantic plausibility in older adults than in younger adults when working memory load was low. But no significant age difference in the effect of syntactic consistency was discovered. When working memory load was high, there was a stronger effect of semantic plausibility and a weaker effect of syntactic consistency in older adults than in younger adults, which suggests that older adults relied more on semantic plausibility and less on syntactic analysis than younger adults. The findings indicate that there is an age-related increase in the use of semantic plausibility, and a reduction in the use of syntactic analysis as working memory load increases.

How do people interpret implausible sentences?

People sometimes interpret implausible sentences nonliterally, for example treating The mother gave the candle the daughter as meaning the daughter receiving the candle. But how do they do so? We contrasted a nonliteral syntactic analysis account, according to which people compute a syntactic analysis appropriate for this nonliteral meaning, with a nonliteral semantic interpretation account, according to which they arrive at this meaning via purely semantic processing. The former but not the latter account postulates that people consider not only a literal-but-implausible double-object (DO) analysis in comprehending The mother gave the candle the daughter, but also a nonliteral-but-plausible prepositional-object (PO) analysis (i.e., including to before the daughter). In three structural priming experiments, participants heard a plausible or implausible DO or PO prime sentence. They then answered a comprehension question first or described a picture of a dative event first. In accord with the nonliteral syntactic analysis account, priming was reduced following implausible sentences than following plausible sentences and following nonliterally interpreted implausible sentences than literally interpreted implausible sentences. The results suggest that comprehenders constructed a nonliteral syntactic analysis, which we argue was predicted early in the sentence.

Biomedical and clinical English model packages for the Stanza Python NLP library.

OBJECTIVE: The study sought to develop and evaluate neural natural language processing (NLP) packages for the syntactic analysis and named entity recognition of biomedical and clinical English text. MATERIALS AND METHODS: We implement and train biomedical and clinical English NLP pipelines by extending the widely used Stanza library originally designed for general NLP tasks. Our models are trained with a mix of public datasets such as the CRAFT treebank as well as with a private corpus of radiology reports annotated with 5 radiology-domain entities. The resulting pipelines are fully based on neural networks, and are able to perform tokenization, part-of-speech tagging, lemmatization, dependency parsing, and named entity recognition for both biomedical and clinical text. We compare our systems against popular open-source NLP libraries such as CoreNLP and scispaCy, state-of-the-art models such as the BioBERT models, and winning systems from the BioNLP CRAFT shared task. RESULTS: For syntactic analysis, our systems achieve much better performance compared with the released scispaCy models and CoreNLP models retrained on the same treebanks, and are on par with the winning system from the CRAFT shared task. For NER, our systems substantially outperform scispaCy, and are better or on par with the state-of-the-art performance from BioBERT, while being much more computationally efficient. CONCLUSIONS: We introduce biomedical and clinical NLP packages built for the Stanza library. These packages offer performance that is similar to the state of the art, and are also optimized for ease of use. To facilitate research, we make all our models publicly available. We also provide an online demonstration (http://stanza.run/bio).

Automated Syntactic Analysis of Language Abilities in Persons with Mild and Subjective Cognitive Impairment.

In this work we analyze the syntactic complexity of transcribed Swedish-language picture descriptions using a variety of automated syntactic features, and investigate the features' predictive power in classifying narratives from people with subjective and mild cognitive impairment and healthy controls. Our results indicate that while there are no statistically significant differences, syntactic features can still be moderately successful at distinguishing the participant groups when used in a machine learning framework.

Comparative Analysis of Linguistic Features in Medical Knowledge Base / 医学信息学杂志

Taking Physician Data Query (PDQ(R)),which is provided for doctors and patients by National Cancer Institute (NCI),for example,the paper carries out comparative analysis on professional and public editions of PDQ from the perspective of vocabulary and syntax to make clear the significant statistical difference between the knowledge bases facing patients and doctors.

Statistical parsing of varieties of clinical Finnish.

OBJECTIVES: In this paper, we study the development and domain-adaptation of statistical syntactic parsers for three different clinical domains in Finnish. METHODS AND MATERIALS: The materials include text from daily nursing notes written by nurses in an intensive care unit, physicians' notes from cardiology patients' health records, and daily nursing notes from cardiology patients' health records. The parsing is performed with the statistical parser of Bohnet (http://code.google.com/p/mate-tools/, accessed: 22 November 2013). RESULTS: A parser trained only on general language performs poorly in all clinical subdomains, the labelled attachment score (LAS) ranging from 59.4% to 71.4%, whereas domain data combined with general language gives better results, the LAS varying between 67.2% and 81.7%. However, even a small amount of clinical domain data quickly outperforms this and also clinical data from other domains is more beneficial (LAS 71.3-80.0%) than general language only. The best results (LAS 77.4-84.6%) are achieved by using as training data the combination of all the clinical treebanks. CONCLUSIONS: In order to develop a good syntactic parser for clinical language variants, a general language resource is not mandatory, while data from clinical fields is. However, in addition to the exact same clinical domain, also data from other clinical domains is useful.