Christopher M. Ormerod

Christopher Michael Ormerod, Alexander Kwako

Current research on generative language models (GLMs) for automated text scoring (ATS) has focused almost exclusively on querying proprietary models via Application Programming Interfaces (APIs). Yet such practices raise issues around transparency and security, and these methods offer little in the way of efficiency or customizability. With the recent proliferation of smaller, open-source models, there is the option to explore GLMs with computers equipped with modest, consumer-grade hardware, that is, for the "GPU poor." In this study, we analyze the performance and efficiency of open-source, small-scale GLMs for ATS. Results show that GLMs can be fine-tuned to achieve adequate, though not state-of-the-art, performance. In addition to ATS, we take small steps towards analyzing models' capacity for generating feedback by prompting GLMs to explain their scores. Model-generated feedback shows promise, but requires more rigorous evaluation focused on targeted use cases.

Christopher Ormerod, Amy Burkhardt, Mackenzie Young et al.

This study demonstrates the effectiveness of XLNet, a transformer-based language model, for annotating argumentative elements in persuasive essays. XLNet's architecture incorporates a recurrent mechanism that allows it to model long-term dependencies in lengthy texts. Fine-tuned XLNet models were applied to three datasets annotated with different schemes - a proprietary dataset using the Annotations for Revisions and Reflections on Writing (ARROW) scheme, the PERSUADE corpus, and the Argument Annotated Essays (AAE) dataset. The XLNet models achieved strong performance across all datasets, even surpassing human agreement levels in some cases. This shows XLNet capably handles diverse annotation schemes and lengthy essays. Comparisons between the model outputs on different datasets also revealed insights into the relationships between the annotation tags. Overall, XLNet's strong performance on modeling argumentative structures across diverse datasets highlights its suitability for providing automated feedback on essay organization.

Zifan Wang, Christopher Ormerod

Automated Short Answer Scoring (ASAS) is a critical component in educational assessment. While traditional ASAS systems relied on rule-based algorithms or complex deep learning methods, recent advancements in Generative Language Models (GLMs) offer new opportunities for improvement. This study explores the application of GLMs to ASAS, leveraging their off-the-shelf capabilities and performance in various domains. We propose a novel pipeline that combines vector databases, transformer-based encoders, and GLMs to enhance short answer scoring accuracy. Our approach stores training responses in a vector database, retrieves semantically similar responses during inference, and employs a GLM to analyze these responses and determine appropriate scores. We further optimize the system through fine-tuned retrieval processes and prompt engineering. Evaluation on the SemEval 2013 dataset demonstrates a significant improvement on the SCIENTSBANK 3-way and 2-way tasks compared to existing methods, highlighting the potential of GLMs in advancing ASAS technology.

Christopher Ormerod, Gitit Kehat

Transformer-based language models are architecturally constrained to process text of a fixed maximum length. Essays written by higher-grade students frequently exceed the maximum allowed length for many popular open-source models. A common approach to addressing this issue when using these models for Automated Essay Scoring is to truncate the input text. This raises serious validity concerns as it undermines the model's ability to fully capture and evaluate organizational elements of the scoring rubric, which requires long contexts to assess. In this study, we evaluate several models that incorporate architectural modifications of the standard transformer architecture to overcome these length limitations using the Kaggle ASAP 2.0 dataset. The models considered in this study include fine-tuned versions of XLNet, Longformer, ModernBERT, Mamba, and Llama models.

Christopher Ormerod

We propose a new way to assess certain short constructed responses to mathematics items. Our approach uses a pipeline that identifies the key values specified by the student in their response. This allows us to determine the correctness of the response, as well as identify any misconceptions. The information from the value identification pipeline can then be used to provide feedback to the teacher and student. The value identification pipeline consists of two fine-tuned language models. The first model determines if a value is implicit in the student response. The second model identifies where in the response the key value is specified. We consider both a generic model that can be used for any prompt and value, as well as models that are specific to each prompt and value. The value identification pipeline is a more accurate and informative way to assess short constructed responses than traditional rubric-based scoring. It can be used to provide more targeted feedback to students, which can help them improve their understanding of mathematics.

Alexander Scarlatos, Nigel Fernandez, Christopher Ormerod et al.

Item (question) difficulties play a crucial role in educational assessments, enabling accurate and efficient assessment of student abilities and personalization to maximize learning outcomes. Traditionally, estimating item difficulties can be costly, requiring real students to respond to items, followed by fitting an item response theory (IRT) model to get difficulty estimates. This approach cannot be applied to the cold-start setting for previously unseen items either. In this work, we present SMART (Simulated Students Aligned with IRT), a novel method for aligning simulated students with instructed ability, which can then be used in simulations to predict the difficulty of open-ended items. We achieve this alignment using direct preference optimization (DPO), where we form preference pairs based on how likely responses are under a ground-truth IRT model. We perform a simulation by generating thousands of responses, evaluating them with a large language model (LLM)-based scoring model, and fit the resulting data to an IRT model to obtain item difficulty estimates. Through extensive experiments on two real-world student response datasets, we show that SMART outperforms other item difficulty prediction methods by leveraging its improved ability alignment.

Kai North, Christopher Ormerod

Automated scoring (AS) systems used in large-scale assessment have traditionally used small statistical models that require a large quantity of hand-scored data to make accurate predictions, which can be time-consuming and costly. Generative Large Language Models are trained on many tasks and have shown impressive abilities to generalize to new tasks with little to no data. While these models require substantially more computational power to make predictions, they still require some fine-tuning to meet operational standards. Evidence suggests that these models can exceed human-human levels of agreement even when fine-tuned on small amounts of data. With this in mind, we propose a model distillation pipeline in which a large generative model, a Teacher, teaches a much smaller model, a Student. The Teacher, trained on a small subset of the training data, is used to provide scores on the remaining training data, which is then used to train the Student. We call the resulting dataset "Cyborg Data", as it combines human and machine-scored responses. Our findings show that Student models trained on "Cyborg Data" show performance comparable to training on the entire dataset, while only requiring 10% of the original hand-scored data.

Christopher Ormerod

Traditional methods for determining assessment item parameters, such as difficulty and discrimination, rely heavily on expensive field testing to collect student performance data for Item Response Theory (IRT) calibration. This study introduces a novel approach that implicitly models these psychometric properties by fine-tuning Large Language Models (LLMs) to simulate student responses across a spectrum of latent abilities. Leveraging the Qwen-3 dense model series and Low-Rank Adaptation (LoRA), we train models to generate responses to multiple choice questions conditioned on discrete ability descriptors. We reconstruct the probability of a correct response as a function of student ability, effectively generating synthetic Item Characteristic Curves (ICCs) to estimate IRT parameters. Evaluation on a dataset of Grade 6 English Language Arts (ELA) items and the BEA 2024 Shared Task dataset demonstrates that this method competes with or outperforms baseline approaches. This simulation-based technique seems particularly effective at modeling item discrimination.

Christopher Ormerod

This study illustrates how incorporating feedback-oriented annotations into the scoring pipeline can enhance the accuracy of automated essay scoring (AES). This approach is demonstrated with the Persuasive Essays for Rating, Selecting, and Understanding Argumentative and Discourse Elements (PERSUADE) corpus. We integrate two types of feedback-driven annotations: those that identify spelling and grammatical errors, and those that highlight argumentative components. To illustrate how this method could be applied in real-world scenarios, we employ two LLMs to generate annotations -- a generative language model used for spell correction and an encoder-based token-classifier trained to identify and mark argumentative elements. By incorporating annotations into the scoring process, we demonstrate improvements in performance using encoder-based large language models fine-tuned as classifiers.

Christopher M. Ormerod, Milan Patel, Harry Wang

This article details the advances made to a system that uses artificial intelligence to identify alarming student responses. This system is built into our assessment platform to assess whether a student's response indicates they are a threat to themselves or others. Such responses may include details concerning threats of violence, severe depression, suicide risks, and descriptions of abuse. Driven by advances in natural language processing, the latest model is a fine-tuned language model trained on a large corpus consisting of student responses and supplementary texts. We demonstrate that the use of a language model delivers a substantial improvement in accuracy over the previous iterations of this system.

Christopher Ormerod

We investigate the effectiveness of ensembles of pretrained transformer-based language models on short answer questions using the Kaggle Automated Short Answer Scoring dataset. We fine-tune a collection of popular small, base, and large pretrained transformer-based language models, and train one feature-base model on the dataset with the aim of testing ensembles of these models. We used an early stopping mechanism and hyperparameter optimization in training. We observe that generally that the larger models perform slightly better, however, they still fall short of state-of-the-art results one their own. Once we consider ensembles of models, there are ensembles of a number of large networks that do produce state-of-the-art results, however, these ensembles are too large to realistically be put in a production environment.

Christopher Ormerod, Amir Jafari, Susan Lottridge et al.

We study the effects of data size and quality on the performance on Automated Essay Scoring (AES) engines that are designed in accordance with three different paradigms; A frequency and hand-crafted feature-based model, a recurrent neural network model, and a pretrained transformer-based language model that is fine-tuned for classification. We expect that each type of model benefits from the size and the quality of the training data in very different ways. Standard practices for developing training data for AES engines were established with feature-based methods in mind, however, since neural networks are increasingly being considered in a production setting, this work seeks to inform us as to how to establish better training data for neural networks that will be used in production.

Christopher M Ormerod, Akanksha Malhotra, Amir Jafari

Automated Essay Scoring (AES) is a cross-disciplinary effort involving Education, Linguistics, and Natural Language Processing (NLP). The efficacy of an NLP model in AES tests it ability to evaluate long-term dependencies and extrapolate meaning even when text is poorly written. Large pretrained transformer-based language models have dominated the current state-of-the-art in many NLP tasks, however, the computational requirements of these models make them expensive to deploy in practice. The goal of this paper is to challenge the paradigm in NLP that bigger is better when it comes to AES. To do this, we evaluate the performance of several fine-tuned pretrained NLP models with a modest number of parameters on an AES dataset. By ensembling our models, we achieve excellent results with fewer parameters than most pretrained transformer-based models.

Pedro Uria Rodriguez, Amir Jafari, Christopher M. Ormerod

In this paper, we present a new comparative study on automatic essay scoring (AES). The current state-of-the-art natural language processing (NLP) neural network architectures are used in this work to achieve above human-level accuracy on the publicly available Kaggle AES dataset. We compare two powerful language models, BERT and XLNet, and describe all the layers and network architectures in these models. We elucidate the network architectures of BERT and XLNet using clear notation and diagrams and explain the advantages of transformer architectures over traditional recurrent neural network architectures. Linear algebra notation is used to clarify the functions of transformers and attention mechanisms. We compare the results with more traditional methods, such as bag of words (BOW) and long short term memory (LSTM) networks.

Christopher M. Ormerod, Amy E. Harris

Automated scoring engines are increasingly being used to score the free-form text responses that students give to questions. Such engines are not designed to appropriately deal with responses that a human reader would find alarming such as those that indicate an intention to self-harm or harm others, responses that allude to drug abuse or sexual abuse or any response that would elicit concern for the student writing the response. Our neural network models have been designed to help identify these anomalous responses from a large collection of typical responses that students give. The responses identified by the neural network can be assessed for urgency, severity, and validity more quickly by a team of reviewers than otherwise possible. Given the anomalous nature of these types of responses, our goal is to maximize the chance of flagging these responses for review given the constraint that only a fixed percentage of responses can viably be assessed by a team of reviewers.