The last decade has seen a rising awareness on computing's social and ethical impacts. If the ambition is for computing disciplines to reposition themselves to be able to educate about these impacts, however, we have to wonder whether we are duly training all the change-makers: not only the students, but also their teachers. While recent computing curricula such as the ACM/IEEE CC2020 and CS2023 have evolved to include Social, Ethical, and Professional (SEP) competencies, the shift has exposed a gap: students are expected to develop these competencies during their degrees, but these learning goals are supposed to be filled by faculty with limited preparation and support. In this paper we argue that, as a community, we are underexploring systematic discussions, research, and practices about the development of faculty SEP competencies. In the status quo, faculty teach SEP through a hidden curriculum: our curricula take socio-ethical stances, but they do so tacitly, implicitly, or with limited awareness. We call for direct action in building SEP competencies by proposing systemic, local, and individual actions such as creating incentives, establishing communities of practice, and promoting reflective practice. Our hope is that, by developing new practices, we can build sufficient critical mass in our community to reorient computing towards the social. Our wish for this paper is that it may serve as a call to action for the computer science education community to invest in its educators as the architects of hope for a socially responsible computing profession.

Computer Science Curricula 2023 identifies Society, Ethics, and Profession competencies essential to all computing topics and calls for their integration. However, limited research addresses faculty perspectives on developing these competencies. My PhD explores how educators acquire, refine, and implement SEP competencies, bridging the gap between ideal teaching and practice.

The development of non-technical skills, such as critical thinking and public speaking, is increasingly recognized in computing curricula as essential for computing professionals in the job market. Universities are adopting active learning methods to cultivate these skills, among which are debate-based teaching methods. Previous literature showed that debating is used in many knowledge domains, and can help develop skills of information synthesis, critical thinking, and effective verbal communication. In spite of these benefits, however, debating in computing education is still rare. This work explores a partial redesign of a debates-based teaching method, called Tech Battles, implemented in a 6-ECTS non-technical course at the University of Trento targeted to first-year Master's students in Computer Science with a minor in innovation and entrepreneurship. Battles have been part of the course since 2013, and over the years the teaching team upgraded the methodology, which now foresees a preparatory activity where students read a science fiction short story assigned by the teachers. In this paper, we discuss how this latest modification changes the course and its impacts: how the workflow of each debate evolved, our criteria for selecting short stories, preliminary observations gathered from the field, and the research methods that will be used to more thoroughly validate the updated method.

This innovative practice full paper aims to describe how an interdisciplinary group of engineering students developed sustainability competencies during a two-week Challenge-Based Learning summer school on analysing the impacts and sustainability of digital education and its infrastructure. Our contribution aims to create a “fair” narrative of this experience, taking into account multiple perspectives, and building a transparent and representative reconstruction of the students' competence development. Our work is qualitative in nature: we gathered pedagogical insights, participants' emotions, thoughts, and experiences, by relying on a variety of data sources (of which the primary was student diaries), ultimately creating a micro-ethnography of the course. This approach allowed us to identify and track the evolution of the GreenComp competencies over time, create team profiles, and gain insights on the evolution of the students' thoughts, perspectives, and social dynamics. While further research is needed to refine the approach, reduce researcher burden, and test it in other environments, this paper aims to start the process of creating an analysis method that lets teachers and researchers reconstruct the students' learnings, experiences, and personal reflections in a rich and transparent way, going beyond the creation of all-too-reductive performance metrics, and contributing to the ongoing discussion on the development of non-technical competencies in engineering education.

The advent of powerful neural classifiers has increased interest in problems that require both learning and reasoning.These problems are critical for understanding important properties of models, such as trustworthiness, generalization, interpretability, and compliance to safety and structural constraints. However, recent research observed that tasks requiring both learning and reasoning on background knowledge often suffer from reasoning shortcuts (RSs): predictors can solve the downstream reasoning task without associating the correct concepts to the high-dimensional data. To address this issue, we introduce rsbench, a comprehensive benchmark suite designed to systematically evaluate the impact of RSs on models by providing easy access to highly customizable tasks affected by RSs. Furthermore, rsbench implements common metrics for evaluating concept quality and introduces novel formal verification procedures for assessing the presence of RSs in learning tasks. Using rsbench, we highlight that obtaining high quality concepts in both purely neural and neuro-symbolic models is a far-from-solved problem. rsbench is available at: https://unitn-sml.github.io/rsbench.

Challenge-Based Learning (CBL) is being increasingly adopted in engineering education, including in mandatory courses. Recently, authors have described a tension between structured and open approaches to CBL, with significant pedagogical tradeoffs in both cases. This Work-In-Progress Innovative Practice article proposes “Choose Your Own Adventure in Challenge-Based Learning” (CYOA-CBL), a model which reconfigures the open/structured CBL duality as a spectrum, and empowers students to fine-tune their learning by making their course more structured or open. Through CYOA-CBL, we hope to give our contribution to making CBL more replicable and better aligned with diverse student motivations.

Despite being studied for over twenty years, Recommender Systems (RSs) still suffer from important issues that limit their applicability in real-world scenarios. Data sparsity, cold start, and explainability are some of the most impacting problems. Intuitively, these historical limitations can be mitigated by injecting prior knowledge into recommendation models. Neuro-Symbolic (NeSy) approaches are suitable candidates for achieving this goal. Specifically, they aim to integrate learning (e.g., neural networks) with symbolic reasoning (e.g., logical reasoning). Generally, the integration lets a neural model interact with a logical knowledge base, enabling reasoning capabilities. In particular, NeSy approaches have been shown to deal well with poor training data, and their symbolic component could enhance model transparency. This gives insights that NeSy systems could potentially mitigate the aforementioned RSs limitations. However, the application of such systems to RSs is still in its early stages, and most of the proposed architectures do not really exploit the advantages of a NeSy approach. To this end, we conducted preliminary experiments with a Logic Tensor Network (LTN), a novel NeSy framework. We used the LTN to train a vanilla Matrix Factorization model using a First-Order Logic knowledge base as an objective. In particular, we encoded facts to enable the regularization of the latent factors using content information, obtaining promising results. In this paper, we review existing NeSy recommenders, argue about their limitations, show our preliminary results with the LTN, and propose interesting future works in this novel research area. In particular, we show how the LTN can be intuitively used to regularize models, perform cross-domain recommendation, ensemble learning, and explainable recommendation, reduce popularity bias, and easily define the loss function of a model.

This paper proposes an effective bottom-up extension of the popular FIND-S algorithm to learn (monotone) DNF-type rulesets.The algorithm greedily finds a partition of the positive examples.The produced monotone DNF is a set of conjunctive rules, each corresponding to the most specific rule consistent with a part of positive and all negative examples.We also propose two principled extensions of this method, approximating the Bayes Optimal Classifier by aggregating monotone DNF decision rules.Finally, we provide a methodology to improve the explainability of the learned rules while retaining their generalization capabilities.An extensive comparison with state-of-the-art symbolic and statistical methods on several benchmark data sets shows that our proposal provides an excellent balance between explainability and accuracy.

Deep generative modeling (DGM) is an increasingly popular approach that can create novel and unseen data, starting from a given data set. As the technology shows promising applications, many ethical issues also arise. For example, their misuse can enable disinformation campaigns and powerful phishing attempts. Research also shows different biases affect deep learning models, leading to social issues such as misrepresentation. In this work, we formulate a novel setting to deal with similar problems, showing that a repurposed anomaly detection system effectively generates novel data, avoiding generating specified unwanted data. We propose Variational Auto-encoding Binary Classifiers (V-ABC): a novel model that repurposes and extends the Auto-encoding Binary Classifier (ABC) anomaly detector using the Variational Auto-encoder (VAE). We survey the limitations of existing approaches and explore many tools to show the model's inner workings in an interpretable way. This proposal has excellent potential for generative applications: models that rely on user-generated data could automatically filter out unwanted content, such as offensive language, obscene images, and misleading information.