Matching Items (4)
Filtering by
- All Subjects: Knowledge
- Creators: Alvarez, Melanie
- Creators: Baral, Chitta
- Creators: Brewer, Gene A
Description
Collaborative learning is a potential technique for teachers to use to meet the diverse learning needs of the students in their classrooms. Previous studies have investigated the contexts in which the benefits of collaborative learning show greater presence. The most important factor found was the quality of the interactions. Studies have suggested that high achieving students are capable of improving the quality of interactions. This bears the question if prior knowledge plays an influence in the learning outcome of students in collaborative learning. Results show that high prior knowledge students do not face a detriment in having low prior knowledge students as a partner comparing to having another high prior knowledge student and that low prior knowledge students show significantly higher learning outcome when partnered with a high prior knowledge partner than with another low prior knowledge student. It is therefore likely that having a high prior knowledge student within a dyad improves the quality of interaction, resulting in greater learning outcome through collaborative learning.
ContributorsKeyvani, Kewmars (Author) / Chi, Michelene (Thesis director) / Wylie, Ruth (Committee member) / Barrett, The Honors College (Contributor) / Department of Psychology (Contributor) / School of Life Sciences (Contributor)
Created2014-05
Description
In natural language processing, language models have achieved remarkable success over the last few years. The Transformers are at the core of most of these models. Their success can be mainly attributed to an enormous amount of curated data they are trained on. Even though such language models are trained on massive curated data, they often need specific extracted knowledge to understand better and reason. This is because often relevant knowledge may be implicit or missing, which hampers machine reasoning. Apart from that, manual knowledge curation is time-consuming and erroneous. Hence, finding fast and effective methods to extract such knowledge from data is important for improving language models. This leads to finding ideal ways to utilize such knowledge by incorporating them into language models. Successful knowledge extraction and integration lead to an important question of knowledge evaluation of such models by developing tools or introducing challenging test suites to learn about their limitations and improve them further. So to improve the transformer-based models, understanding the role of knowledge becomes important. In the pursuit to improve language models with knowledge, in this dissertation I study three broad research directions spanning across the natural language, biomedical and cybersecurity domains: (1) Knowledge Extraction (KX) - How can transformer-based language models be leveraged to extract knowledge from data? (2) Knowledge Integration (KI) - How can such specific knowledge be used to improve such models? (3) Knowledge Evaluation (KE) - How can language models be evaluated for specific skills and understand their limitations?
I propose methods to extract explicit textual, implicit structural, missing textual, and missing structural knowledge from natural language and binary programs using transformer-based language models. I develop ways to improve the language model’s multi-step and commonsense reasoning abilities using external knowledge. Finally, I develop challenging datasets which assess their numerical reasoning skills in both in-domain and out-of-domain settings.
ContributorsPal, Kuntal Kumar (Author) / Baral, Chitta (Thesis advisor) / Wang, Ruoyu (Committee member) / Blanco, Eduardo (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2023
Description
Arizona is a unique state in that rain is not a normal occurrence throughout most of the year (NWS). Arizona averages from less than three months to half a month of measurable precipitation days per year (WRCC). With that, it is important to know the public’s understanding as well as their general trend of likeness towards the weather forecasts they receive. A questionnaire was distributed to 426 people in the state of Arizona to review what they understand from the forecasts and what they would like to see on social media and television.
ContributorsHermansen, Alexis Nicole (Author) / Alvarez, Melanie (Thesis director) / Cerveny, Randall (Committee member) / Hondula, David M. (Committee member) / Walter Cronkite School of Journalism & Mass Comm (Contributor) / School of Geographical Sciences and Urban Planning (Contributor, Contributor) / Barrett, The Honors College (Contributor)
Created2019-05
Description
Individuals encounter problems daily wherein varying numbers of constraints require delimitation of memory to target goal-satisfying information. Multiply-constrained problems, such as compound remote associates, are commonly used to study this type of problem solving. Since their development, multiply-constrained problems have been theoretically and empirically related to creative thinking, analytical problem solving, insight problem solving, intelligence, and a multitude of other cognitive abilities. Critically, in order to correctly solve a multiply-constrained problem the solver must have the solution available in memory and be able to target and access to that information. Experiment 1 determined that the cue – target relationship affects the likelihood that a problem is solved. Moreover, Experiment 2 identified that the association between cues and targets predicted inter- & intra-individual differences in multiply-constrained problem solving. Lastly, Experiment 3 found monetary incentives failed to improve problem solving performance likely due to knowledge serving as a limiting factor on performance. Additionally, problem solvers were shown to be able to reliably assess the likelihood they would solve a problem. Taken together all three studies demonstrated the importance of knowledge & knowledge structures on problem solving performance.
ContributorsEllis, Derek (Author) / Brewer, Gene A (Thesis advisor) / Homa, Donald (Committee member) / Blais, Chris (Committee member) / Goldinger, Stephen (Committee member) / Arizona State University (Publisher)
Created2021