Exploring Artificial Intelligence (AI) in English Language Arts (ELA) with StoryQ is a 10-hour curriculum module designed for high school ELA classes. The module introduces students to fundamental AI concepts and essential machine learning workflow using StoryQ, a web-based GUI environment for Grades 6-12 learners. In this module, students work with unstructured text data and learn to train, test, and improve text classification models such as intent recognition, clickbait filter, and sentiment analysis. As they interact with machine-learning language models deeply, students also gain a nuanced understanding of language and how to wield it, not just as a data structure, but as a tool in our human-human encounters as well. The current version contains eight lessons, all delivered through a full-featured online learning and teaching platform. Computers and Internet access are required to implement the module. The module was piloted in an ELA class in the Spring of 2022, and the student learning outcomes were positive. The module is currently undergoing revision and will be further tested and improved in Fall 2022.

StoryQ is a web-based machine learning and text mining tool that allows young learners (Grade 6-12) to engage in machine learning practices and work with unstructured text data without needing to code. StoryQ features dynamically linked data representations that promote meaningful inquiries and understandings across tables, graphs, and texts. These links create a unique user experience that makes machine learning models transparent, explainable, and fun to explore. This demo will showcase how key AI concepts such as representation, reasoning, feature space, feature weight, and machine learning are dynamically visualized in StoryQ and made accessible to young learners. A brief tutorial will be provided on how to use StoryQ to train, test, and troubleshoot text classification models using both standard feature extractors (e.g., N-grams) and special feature extraction tools and visualizations that have been specially designed to support young learners and non-computing teachers. This demo will also include sample learning activities designed for high school English Language Arts and History classes to showcase how machine learning concepts and practices can be introduced in non-computing classes. As the demands for AI scientists, engineers, and entrepreneurs have increased in recent years, as well as AI's increased presence in everyday lives, making access to how machine learning practices work is of paramount importance for young learners. This work is supported by an NSF ITEST project (DRL-1949110).

StoryQ is a web-based machine learning and text mining tool that allows young learners (Grade 6-12) to engage in machine learning practices and work with unstructured text data without coding. StoryQ features dynamically linked data representations that promote meaningful inquiries across tables, graphs, and texts--a unique user experience that makes machine learning models transparent, explainable, and fun to explore. This demo provides a brief tutorial on how to use StoryQ to train, test, and troubleshoot text classification models using both standard feature extractors (e.g., N-grams) and special feature extraction tools and visualizations designed to support young learners and non-computing teachers. This demo also includes sample learning activities designed for high school English Language Arts classes to showcase how machine learning concepts and practices can be introduced in non-computing classes. This work is supported by an NSF I-TEST project (DRL-1949110).

Recent work on entity coreference resolution (CR) follows current trends in Deep Learning applied to embeddings and relatively simple task-related features.SOTA models do not make use of hierarchical representations of discourse structure.In this work, we leverage automatically constructed discourse parse trees within a neural approach and demonstrate a significant improvement on two benchmark entity coreference-resolution datasets.We explore how the impact varies depending upon the type of mention.

We introduce a general method for the interpretation and comparison of neural models.The method is used to factor a complex neural model into its functional components, which are comprised of sets of co-firing neurons that cut across layers of the network architecture, and which we call neural pathways.The function of these pathways can be understood by identifying correlated task level and linguistic heuristics in such a way that this knowledge acts as a lens for approximating what the network has learned to apply to its intended task.As a case study for investigating the utility of these pathways, we present an examination of pathways identified in models trained for two standard tasks, namely Named Entity Recognition and Recognizing Textual Entailment.

Analysis of student writing, both for assessment and for enabling feedback have been of interest to the field of learning analytics. While much progress can be made through detection of local cues in writing, structured prediction approaches offer capabilities that are particularly well tailored to the needs of models aiming to offer substantive feedback on rhetorical structure. We thus cast the analysis of rhetorical structure in academic writing as a structured prediction task in which we employ models that leverage both local and global cues in writing. In particular, this paper presents a hierarchical neural architecture that performs this task. The evaluation demonstrates that the architecture achieves near-human performance while significantly surpassing state-of-the-art baselines. A multifaceted approach to model interpretation offers insights into the inner workings of the model.

Support of discussion based learning at scale benefits from automated analysis of discussion for enabling effective assignment of students to project teams, for triggering dynamic support of group learning processes, and for assessment of those learning processes. A major limitation of much past work in machine learning applied to automated analysis of discussion is the failure of the models to generalize to data outside of the parameters of the context in which the training data was collected. This limitation means that a separate training effort must be undertaken for each domain in which the models will be used. This paper focuses on a specific construct of discussion based learning referred to as Transactivity and provides a novel machine learning approach with performance that exceeds state-of-the-art performance within the same domain in which it was trained and a new domain, and does not suffer any reduction in performance when transferring to the new domain. These results stand as an advance over past work on automated detection of Transactivity and increase the value of trained models for supporting group learning at scale. Implications for practice in at-scale learning environments are discussed.