A lot about the learning process still remains unknown. The experiments described in this thesis investigated variables that affect instructional efficiency by employing specifically programmed computers to manage and control instructional variables within each experiment for 6- to 7-year old children. A Measurement Procedures Study was undertaken to ascertain when a response should be classified as "acquired." It was decided to classify a response as acquired if it could be performed correctly (without prompting) seven days after instruction. A review of the relationship between accuracy level during instruction and the rate of acquisition found that higher accuracy levels during instruction tend to be associated with higher rates of acquisition provided that non-copying prompting procedures are employed. The first experiment investigated the relationship between accuracy level during instruction and rate of acquisition by presenting a non-copying antecedent prompt (model of the correct spelling word) depending on a preselected target accuracy level. As an error-contingent prompt (model of the correct spelling word) was also provided it could not be ascertained whether transfer of stimulus control occurred as a result of the antecedent prompt, or the error-contingent prompt, or both. The second experiment was a repeat of the first experiment with the error-contingent prompt removed. It was found that it was possible to manage, although not completely control, the accuracy level during instruction by presenting a simultaneous non-copying prompt and that higher accuracy levels during instruction were associated with higher rates of acquisition. A review that examined the error-correction research found that a variety of correction procedures were effective. However, none of the 36 experiments which were reviewed controlled the number of response opportunities. Experiment 3 compared the effects on rate of acquisition of presenting an antecedent model or an error-contingent model. The results of Experiment 3 showed that when the number of learning opportunities was controlled there was little difference in effectiveness or efficiency between an antecedent model and an error-contingent model. Experiment 4 compared the effects of presenting an error-contingent model against an error-contingent model and a secondary response opportunity. It was found that an error-contingent model was at least as effective, although it was overall less efficient when response opportunities were controlled. A supplementary analysis was undertaken to review and compare the results obtained across the four experiments. Across experiments each newly acquired spelling response required about five practice responses, on average. It appears this was a critical variable for acquisition. Additionally, each acquired response was acquired over a two-day period. Although rates of acquisition differed between high-achieving children and low-achieving children, there was little difference in the number of practice responses required for acquisition between these two groups. It was observed that most of the 6- to 7-year old participants found error feedback aversive and this appeared to result in reduced attention to models of the correct spelling when these occurred following errors. The results from this series of investigations suggest that an opportunity for the transfer of stimulus control from the prompt (model of the correct spelling) to the practice stimulus (the spoken word) is more critical for acquisition than where the prompt occurs within the trial (that is, the antecedent or consequent position). It was suggested that future research could investigate (a) the variables which are necessary for the transfer of stimulus control, (b) the generality of the observation that children require five practice responses in order to acquire discrete academic responses, and (c) the effects on rates of acquisition and instructional efficiency of varying the distribution in time of practice responses for children who are learning various types of academic skills.

Thesis (Ph. D.)--University of Missouri-Columbia, 2000. / Typescript. Vita. Includes bibliographical references (leaves 112-118). Also available on the Internet.

No description available.

No description available.

With the increasing integration of artificial intelligence (AI) technologies into educational settings, it becomes important to ensure responsible and effective use of these systems. This thesis addresses two critical challenges within AI-driven educational applications: the effortless integration of different Large Language Models (LLMs) and the mitigation of inappropriate content. An AI assistant chatbot was developed, allowing teachers to design custom chatbots and set rules for them, enhancing students’ learning experiences. Evaluation of LangChain as a framework for LLM integration, alongside various prompt engineering techniques including zero-shot, few-shot, zero-shot chain-of-thought, and prompt chaining, revealed LangChain’s suitability for this task and highlighted prompt chaining as the most effective method for mitigating inappropriate content in this use case. Looking ahead, future research could focus on further exploring prompt engineering capabilities and strategies to ensure uniform learning outcomes for all students, as well as leveraging LangChain to enhance the adaptability and accessibility of educational applications.

My dissertation introduces the notion of "augmented authorship" and applications for prompt engineering with generative neural networks inspired by Gregory Ulmer's theories of electracy (2003) to the interdisciplinary fields that teach writing and rhetoric. With the goal of inspiring the general practice of electracy, I introduce prompt engineering as practice in flash reason (Ulmer 2008; 2012), a new collective prudence emerging from the apparatus of electracy. By situating electracy and flash reason as threshold concepts in writing studies, and by aligning principles of electracy with ACRL and NCTE digital literacy frameworks, I demonstrate how prompt engineering across modalities can help students meet digital literacy goals, before providing accessible models or "relays" in the form of AI-coauthored texts, course modules, and aesthetic models deployed in the game world Roblox.

Large Language Models (LLM) and their abilities of text-to-SQL are today a very relevant topic, as utilizing an LLM as a database interface would facilitate easy access to the data in the database without any prior knowledge of SQL. What is being studied in this thesis, is how to best structure a prompt to increase the accuracy of an LLM on a text-to-SQL task. The methods of experimentation used in the study were experimentation with 5 different prompts, and a total of 22 questions asked about the database with the questions being of difficulties varying from easy to extra hard. The results showed that a simpler, less descriptive prompt performed better on the easy and medium questions, while a more descriptive prompt performed better on the hard and extra hard questions. The f indings did not fully align with the hypothesis that more descriptive prompts would have the most correct outputs. In conclusion, it seemed that prompts that contained less ”clutter” and were more straightforward were more effective on easy questions, while on harder questions a prompt with a better description and examples had a better impact.

Online video platforms receive hundreds of hours of uploads every minute, making manual moderation of inappropriate content impossible. The most vulnerable consumers of malicious video content are children from ages 1-5 whose attention is easily captured by bursts of color and sound. Prominent video hosting platforms like YouTube have taken measures to mitigate malicious content, but these videos often go undetected by current automated content moderation tools that are focused on removing explicit or copyrighted content. Scammers attempting to monetize their content may craft malicious children's videos that are superficially similar to educational videos, but include scary and disgusting characters, violent motions, loud music, and disturbing noises. A robust classification of malicious videos requires audio representations in addition to video features. However, recent content moderation approaches rarely employ multimodal architectures that explicitly consider non-speech audio cues. Additionally, there is a dearth of comprehensive datasets for content moderation tasks which include these audio-visual feature annotations. This dissertation addresses these challenges and makes several contributions to the problem of content moderation for children’s videos. The first contribution is identifying a set of malicious features that are harmful to preschool children but remain unaddressed and publishing a labeled dataset (Malicious or Benign) of cartoon video clips that include these features. We provide a user-friendly web-based video annotation tool which can easily be customized and used for video classification tasks with any number of ground truth classes. The second contribution is adapting state-of-the-art Vision-Language models to apply content moderation techniques on the MOB benchmark. We perform prompt engineering and an in-depth analysis of how context-specific language prompts affect the content moderation performance of different CLIP (Contrastive Language-Image Pre-training) variants. This dissertation introduces new benchmark natural language prompt templates for cartoon videos that can be used with Vision-Language models. Finally, we introduce a multimodal framework that includes the audio modality for more robust content moderation of children's cartoon videos and extend our dataset to include audio labels. We present ablations to demonstrate the enhanced performance of adding audio. The audio modality and prompt learning are incorporated while keeping the backbone modules of each modality frozen. Experiments were conducted on a multimodal version of the MOB (Malicious or Benign) dataset in both supervised and few-shot settings.

The prompt gamma-ray timing (PGT) method offers a relatively simple approach for range verification in proton therapy. Starting from the findings of previous experiments, several steps toward a clinical application of PGT have been performed in this work. First of all, several scintillation materials have been investigated in the context of PGT. The time resolution was determined at high photon energies in the MeV-region. In conclusion, the fast and bright scintillator CeBr3 is the material of choice in combination with a timing photomultiplier tube as light detector. A second study was conducted at Universitäts Protonen Therapie Dresden (UPTD) to characterize the proton bunch structure of a clinical beam concerning its time width and relative arrival time. The data is mandatory as input for simulation studies and to correct for phase drifts. The obtained data could furthermore be used for the first 2D imaging of a heterogeneous phantom based on prompt gamma-rays. In a last step, a PGT prototype system was designed using the findings from the first two studies. The prototype system is based on a newly developed digital spectrometer and a CeBr3 detector. The device is characterized at the ELBE bremsstrahlung beam. It was verified that the prototype operates within the specifications concerning time and resolution as well as throughput rate. Finally, for the first time the PGT system was used under clinical conditions in the treatment room of UPTD. Here, PGT data was obtained from the delivery of a three-dimensional treatment plan onto PMMA phantoms. The spot-by-spot analysis helped to investigate the performance of the prototype device under clinical conditions. As a result, range variations of 5 mm could be detected for the first time with an uncollimated system at clinically relevant doses. To summarize, the obtained results help to bring PGT closer to a clinical application.

Scientific findings recorded in literature help continuously guide scientific advancements, but manual approaches to accessing that knowledge are insufficient due to the sheer quantity of information and data available. Although pre-trained language models are being explored for their utility as knowledge bases and structured data repositories, there is a lack of research for this application in the biomedical domain. Therefore, the aim in this project was to determine how Generative Pre-trained Transformer models pre-trained on articles in the biomedical domain can be used to make relevant information more accessible. Several models (BioGPT, BioGPT-Large, and BioMedLM) were evaluated on the task of extracting chemical-protein relations between entities directly from the models through prompting. Prompts were formulated as a natural language text or an ordered triple, and provided in different settings (few-shot, one-shot, or zero-shot). Model-predictions were evaluated quantitatively as a multiclass classification task using a macro-averaged F1-score. The result showed that out of the explored methods, the best performance for extracting chemical-protein relations from article-abstracts was obtained using a triple-based text prompt on the largest model, BioMedLM, in the few-shot setting, albeit with low improvements from the baseline (+0.019 F1). There was no clear pattern for which prompt setting was favourable in terms of task performance, however, the triple based prompt was generally more robust than the natural language formulation. The task performance of the two smaller models underperformed the random baseline (by at best -0.026 and -0.001 F1). The impact of the prompt method was minimal in the smallest model, and the one-shot setting was the least sensitive to the prompt formulation in all models. However, there were more pronounced differences between the prompt methods in the few-shot setting of the larger models (+0.021-0.038 F1). The results suggested that the method of prompting and the size of the model impact the knowledge eliciting performance of a language model. Admittedly, the models mostly underperformed the baseline and future work needs to look into how to adapt generative language models to solve this task. Future research could investigate what impact automatic prompt-design methods and larger in-domain models have on the model performance. / De vetenskapliga upptäckter som presenteras inom litteraturen vägleder kontinuerligt vetenskapliga framsteg. Manuella tillvägagångssätt för att ta del av den kunskapen är otillräckliga på grund av den enorma mängd information och data som finns tillgänglig. Även om för-tränade språkmodeller utforskas för sin brukbarhet som kunskapsbaser och strukturerade dataförråd så finns det en brist på forskning inom den biomedicinska domänen. Målet med detta projekt var att utreda hur Generative Pre-trained Transformer (GPT) modeller för-tränade på biomedicinska artiklar kan användas för att öka tillgängligheten av relevant information inom denna domän. Olika modeller (BioGPT, BioGPT-Large, och BioMedLM) utvärderas på uppgiften att extrahera relationsinformation mellan entiteter direkt ur modellen genom en textprompt. En prompt formuleras genom naturlig text och som en ordnad trippel, och används i olika demonstrationsmiljöer (few-shot, one-shot, zero-shot). Modellförutsägelser utvärderas kvantitativt som ett multi-klass klassifikationsproblem genom ett genomsnittligt F1 värde. Resultatet indikerade att kemikalie-protein relationer från vetenskapliga artikelsammanfattningar kan extraheras med en högre sannolikhet än slumpen med en trippelbaserad prompt genom den största modellen, BioMedLM, i few-shot-miljön, dock med små förbättringar från baslinjen (+0.019 F1). Resultatet visade inga tydliga mönster gällande vilken demonstrationsmiljö som var mest gynnsam, men den trippelbaserade formuleringen var generellt mer robust än formuleringen som följde naturligt språk. Uppgiftsprestandan på de två mindre modellerna underpresterade den slumpmässiga baslinjen (med som bäst -0.026 och -0.001 F1). Effekten av valet av promptmetod var minimal med den minsta modellen, och one-shot-miljön var minst känslig för olika formuleringar hos alla modeller. Dock fanns det mer markanta skillnader mellan promptmetoder i few-shot-miljön hos de större modellerna (+0.021-0.038 F1). Resultatet antydde att valet av promptmetod och storleken på modell påverkar modellens förmåga att extrahera information. De utvärderade modellerna underpresterade dock baslinjen och fortsatt efterforskning behöver se över hur generativa språkmodeller kan anpassas för att lösa denna uppgift. Framtida forskning kan även undersöka vilken effekt automatiska promptdesignmetoder och större domänmodeller har på modellprestanda.