skip to main content


Search for: All records

Creators/Authors contains: "Alm, Cecilia Ovesdotter"

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. Much of the world’s population experiences some form of disability during their lifetime. Caution must be exercised while designing natural language processing (NLP) systems to prevent systems from inadvertently perpetuating ableist bias against people with disabilities, i.e., prejudice that favors those with typical abilities. We report on various analyses based on word predictions of a large-scale BERT language model. Statistically significant results demonstrate that people with disabilities can be disadvantaged. Findings also explore overlapping forms of discrimination related to interconnected gender and race identities. 
    more » « less
  2. To examine perceptions of faculty mentors of undergraduate research and their supervisors, this work discusses the results of surveys administered after 3 years of a summer CS-focused REU Site program. One survey was completed by administrators of faculty research mentors–deans and chairs–and the other was completed by faculty mentors. The surveys indicated a disconnect between how the groups assessed undergraduate research mentoring as an indicator of faculty productivity, and overt vs. covert recognition of undergraduate mentoring. Additional topics explored the effectiveness of internal communication of program outcomes and ways to improve it, as well as post-program continued mentoring engagement linking to perceptions of long-term student benefits. 
    more » « less
  3. Automated journalism technology is transforming news production and changing how audiences perceive the news. As automated text-generation models advance, it is important to understand how readers perceive human-written and machine-generated content. This study used OpenAI’s GPT-2 text-generation model (May 2019 release) and articles from news organizations across the political spectrum to study participants’ reactions to human- and machine-generated articles. As participants read the articles, we collected their facial expression and galvanic skin response (GSR) data together with self-reported perceptions of article source and content credibility. We also asked participants to identify their political affinity and assess the articles’ political tone to gain insight into the relationship between political leaning and article perception. Our results indicate that the May 2019 release of OpenAI’s GPT-2 model generated articles that were misidentified as written by a human close to half the time, while human-written articles were identified correctly as written by a human about 70 percent of the time. 
    more » « less
  4. We report on an emerging undergraduate research framework from the NSF Research Experiences for Undergraduate (REU) Site in Computational Sensing at Rochester Institute of Technology. Unobtrusive observation of people's physiological, behavioral, cognitive, and environmental data is increasingly enabling new computing experiences. This REU Site recognizes the accumulating need for training emerging researchers to gain experience in and grapple with systematic collection, processing, analysis, and interpretation of heterogeneous human-elicited information. Instead of merely leveraging traditional physiological measurements, our research program takes a holistic approach to the capture and integration of such sensing data. For instance, the data may also include linguistic and eye movement behaviors, or social and geospatial contextual information. These modalities provide rich information. An example of a multimodal data collection scenario from a project in the REU Site's first year is in Figure 1. This project applied sensing for observing and measuring cognitive reactions as participants engaged in tasks involving web-based video lecturing. 
    more » « less
  5. Online learning has gained increased popularity in recent years. However, with online learning, teacher observation and intervention is lost, creating a need for technologically observable characteristics that can compensate for this limitation. The present study used a wide array of sensing mechanisms including eye tracking, galvanic skin response (GSR) recording, facial expression analysis, and summary note-taking to monitor participants while they watched and recalled an online video lecture. We explored the link between these human-elicited responses and learning outcomes as measured by quiz questions. Results revealed GSR to be the best indicator of the challenge level of the lecture material. Yet, eye tracking and GSR remain difficult to capture when monitoring online learning as the requirement to remain still impacts natural behavior and leads to more stoic and unexpressive faces. Continued work on methods ensuring naturalistic capture are critical for broadening the use of sensor technology in online learning, as are ways to fuse these data with other input, such as structured and unstructured data from peer-to-peer or student-teacher interactions. 
    more » « less