We address the challenge of automated classification of diffuse large B-cell lymphoma (DLBCL) into its two primary subtypes: activated B-cell-like (ABC) and germinal center B-cell-like (GCB). Accurate classification between these subtypes is essential for determining the appropriate therapeutic strategy, given their distinct molecular profiles and treatment responses. Our proposed deep learning model demonstrates robust performance, achieving an average area under the curve (AUC) of$(87.4 \pm 5.7) \%$during cross-validation. It shows a high positive predictive value (PPV), highlighting its potential for clinical application, such as triaging for molecular testing. To gain biological insights, we performed an analysis of morphological features of ABC and GCB subtypes. We segmented cell nuclei using a pre-trained deep neural network and compared the statistics of geometric and color features for ABC and GCB. We found that the distributions of these features were not very different for the two subtypes, which suggests that the visual differences between them are more subtle. These results underscore the potential of our method to assist in more precise subtype classification and can contribute to improved treatment management and outcomes for patients of DLBCL.

The current standard for detecting human epidermal growth factor receptor 2 (HER2) status in breast cancer patients relies on HER2 expression identified through immunohistochemistry (IHC) or amplification identified through fluorescence in situ hybridization (FISH). However, hematoxylin and eosin (H&E) tumor stains are more widely available, and accurately predicting HER2 status using H&E could reduce costs and expedite treatment selection. Deep Learning algorithms for H&E have shown effectiveness in predicting various cancer features and clinical outcomes, including moderate success in HER2 status prediction. In this work, we employed a customized weak supervision classification technique combined with MoCov2 contrastive learning for self-supervised feature extraction training to predict HER2 status. We trained our pipeline on 182 publicly available H&E whole slide images (WSIs) from The Cancer Genome Atlas (TCGA), for which annotations by the pathology team at Yale School of Medicine are publicly available. Our pipeline achieved an Area Under the Curve (AUC) of 0.85$\pm {0. 0 2}$across four different test folds. Additionally, we tested our model on 44 H&E slides from the TCGA-BRCA dataset, which had an HER2 score of$2+$and included corresponding HER2 status and FISH test results. These cases are considered equivocal for IHC, requiring an expensive FISH test on their IHC slides for disambiguation. Our pipeline demonstrated an AUC of 0.81 on these challenging H & E slides. Reducing the need for FISH test can have significant implications in cancer treatment equity for underserved populations.

Artificial intelligence (AI) and machine learning (ML) have a deepening impact in our world. For empowered citizenship and career readiness, elementary and middle school students need to understand these technologies. This poster reports on five original interactive AI and ML software exhibits tested by 125 elementary and middle school students aged 7 to 14 years. Four themes emerged: Students recognized that AI and ML systems can process data from cameras (perception); they saw that these systems responded to their training input (trust); they appreciated the practical import of AI/ML systems (affective and cognitive attitudes); and students were introduced to models and modes (specialization).

Machine learning (ML) systems are increasingly in use in society. For young learners to be informed citizens and have full career potential it is important for them to understand these concepts. To support this learning, we created "ChemAIstry,'' an interactive software tool for children which demonstrates training and classification in machine learning. Students select which everyday items are safe to bring into a chemistry lab (e.g., a lab coat is safe; pizza is not). These selections serve as training input for a decision tree classifier. After training, students see how the trained model performs in classifying new objects. ChemAIstry was tested with 40 students aged 7 to 14 years at a public K?8 school. The software captured student selections during training. We analyzed these interactions to yield a "Correspondence Score,'' a measure of student understanding of the classification task. We screen-recorded student use of the software and audio-recorded our conversations with them during this use. Our analysis of these data indicates that students were able to understand the concept of model training, including that items were subsequently classified based on their training input. More than half of the student trials indicated that students correctly understood the task. This suggests ChemAIstry was effective in introducing students to these ideas in machine learning. We recommend continued development of related tools for curriculum integration of AI in K-8 education.

In this Research-to-Practice Full Paper, we report on CS Pathways, a middle school computer science (CS) curriculum developed as part of a researcher-practitioner partnership among two public universities and three urban school districts in the Northeast USA. The curriculum serves middle school students to develop apps for social impact. The partnership focuses on bridging the gap between STEM (Science, Technology, Engineering, and Mathematics) and community and gender gap within STEM. The project is based on Culturally Responsive Pedagogy (CRP), which includes the importance of recognizing students’ culture in all facets of learning. The project employs a researcher-practitioner partnership (RPP) model, which recognizes that transformational change can occur in educational ecosystems that connect research, policy, practice, and community work.The project curriculum was collaboratively developed by CS researchers, teacher-practitioners, and school administrators. Middle school students develop their own apps that support socially relevant activities in their communities. Using the RPP process, continuous feedback from researchers, teacher-practitioners, and students shaped the curriculum design. Key feedback was collected via one-on-one meetings with the teacher-practitioners, which bridged the visions and knowledge among different groups of the project partners.The curriculum includes the areas of computing and society, digital tools and collaboration, computing systems, and computational thinking. The curriculum helps students develop a critical consciousness of the role they can play in affecting their communities through computing, and empower them to move beyond simply learning to code [1]. The curriculum strives to demonstrate how to integrate computing across middle school subjects in a culturally-responsive way and spread a powerful message of Computer Science for All. This paper advocates for the need for culturally-responsive computing, describes how it is integrated into teachers’ instruction, and presents the CS Pathways curriculum design.