Search for: All records

We fabricate and measure electrically-gated tunnel junctions in which the insulating barrier is a sliding van der Waals ferroelectric made from parallel-stacked bilayer hexagonal boron nitride and the electrodes are single-layer graphene. Despite the nominally-symmetric tunnel-junction structure, these devices can exhibit substantial electroresistance upon reversing the ferroelectric polarization. The magnitude and sign of tunneling electroresistance are tunable by bias and gate voltage. We show that this behavior can be understood within a simple tunneling model that takes into account the quantum capacitance of the graphene electrodes, so that the tunneling densities of states in the electrodes are separately modified as a function of bias and gate voltage. 

Virus-like particles (VLPs) are self-assembling nanoparticles derived from viruses with potential as scaffolds for myriad applications. They are also excellent testbeds for engineering protein superstructures. Engineers often employ techniques such as amino acid substitutions and insertions/deletions. Yet evolution also utilizes circular permutation, a powerful natural strategy that has not been fully explored in engineering self-assembling protein nanoparticles. Here, we demonstrate this technique using the MS2 VLP as a model self-assembling, proteinaceous nanoparticle. We constructed a comprehensive circular permutation library of the fused MS2 coat protein dimer construct. The strategy revealed terminal locations, validated via cryo-electron microscopy, that enabled C-terminal peptide tagging and led to a protein encapsulation strategy via covalent bonding – a feature the native coat protein does not permit. Our systematic study demonstrates the power of circular permutation for engineering features as well as quantitatively and systematically exploring VLP structural determinants. 

This original research article focuses on the investigation of the use of generative artificial intelligence (GAI) use among students in communication-intensive STEM courses and how this engagement shapes their scientific communication practices, competencies, confidence, and science identity. Using a mixed-methods approach, patterns were identified in how students perceived their current science identity and use of incorporating artificial intelligence (AI) into writing, oral, and technical tasks. Thematic analysis reveals that students use AI for a range of STEM communication endeavors such as structuring lab reports, brainstorming presentation ideas, and verifying code. While many minoritized students explicitly describe AI as a confidence-boosting, timesaving, and competence-enhancing tool, others—particularly those from privileged backgrounds—downplay its influence, despite evidence of its significant role in their science identity. These results suggest the reframing of science identity as being shaped by technological usage and social contingency. This research illuminates both the potential and pitfalls of AI-use in shaping the next generation of scientists. 

ABSTRACT Motor skill expertise can facilitate more automatic movement, engaging less cortical activity while producing appropriate motor output. Accordingly, cortical-evoked N1 responses to balance perturbation, assessed using electroencephalography (EEG), are smaller in young and older adults with better balance. These responses may thus reflect individual balance challenge versus functional, or objective, task difficulty. However, the effect of balance expertise on cortical responses to balance perturbation has not been studied. We hypothesized that balance ability gained though long-term training facilitates more automatic balance control. Using professional modern dancers as balance experts, we compared cortical-evoked responses and biomechanics of the balance-correcting response between modern dancers and nondancers. We predicted that modern dancers would have smaller cortical-evoked responses and better balance recovery at equivalent levels of balance challenge. Support-surface perturbations were normalized to individual challenge levels by delivering perturbations scaled to 60% and 140% of each individual’s step threshold. In contrast to our prediction, dancers exhibited larger N1 responses compared to nondancers while demonstrating similar biomechanical responses. Our results suggest dancers have greater cortical sensitivity to balance perturbations than nondancers. Further, dancer N1 responses modulated across perturbation magnitudes according to differences in objective task difficulty. In contrast, nondancer N1 responses modulated as a function of individual challenge level. Our findings suggest dance training increases sensitivity of the initial, cortical N1 response to balance perturbation, supporting postural alignment to an objective reference. The N1 response may reflect differences in balance-error processing that are altered with specific long-term training and may have implications for rehabilitation. NEW & NOTEWORTHYModern dancers show larger cortical responses to balance perturbations than nondancers, suggesting a greater sensitivity to perturbations. These results contrast with evidence of larger cortical-evoked responses in young adults with poorer balance, consistent with the cortical N1 response being a balance error assessment signal. Whereas nondancers scaled cortical responses by individual differences in N1 amplitude, dancers’ cortical responses were scaled to objective differences in perturbation magnitude, suggesting increased postural awareness due to training. 

Engineering education has faced significant and deep-rooted challenges, including outdated curricula and pedagogical practices, limited access for underrepresented groups, and persistent diversity gaps, that collectively undermine its ability to equip future generations of engineers for a rapidly evolving world. The changes that are needed to reform engineering education are monumental and highlight not only the need for systemic transformation of educational structures but also a fundamental shift in the mindsets of those leading the change. Faculty, professional staff, and administrators must develop knowledge and skills that go beyond their disciplinary training to drive sustainable reform. This article presents a professional development curriculum that has, for over a decade, equipped academic change agents with the tools to implement lasting change. Drawing on experiences from teams supported by the National Science Foundation’s Revolutionizing Engineering Departments (NSF RED) program, the article highlights proven strategies that academic change agents can master and situates them within the broader literature on change in higher education. Specifically, we focus on how academic change agents can develop capacity for systems thinking, build their ability to communicate effectively with various community members, leverage strategic partnerships to increase impact, and cultivate a supportive community of practice with other change agents. 

Anomaly/Outlier detection (AD/OD) is often used in controversial applications to detect unusual behavior which is then further investigated or policed. This means an explanation of why something was predicted as an anomaly is desirable not only for individuals but also for the general population and policy-makers. However, existing explainable AI (XAI) methods are not well suited for Explainable Anomaly detection (XAD). In particular, most XAI methods provide instance-level explanations, whereas a model/global-level explanation is desirable for a complete understanding of the definition of normality or abnormality used by an AD algorithm. Further, existing XAI methods try to explain an algorithm’s behavior by finding an explanation of why an instance belongs to a category. However, by definition, anomalies/outliers are chosen because they are different from the normal instances. We propose a new style of model agnostic explanation, called contrastive explanation, that is designed specifically for AD algorithms which use semantic tags to create explanations. It addresses the novel challenge of providing a model-agnostic and global-level explanation by finding contrasts between the outlier group of instances and the normal group. We propose three formulations: (i) Contrastive Explanation, (ii) Strongly Contrastive Explanation, and (iii) Multiple Strong Contrastive Explanations. The last formulation is specifically for the case where a given dataset is believed to have many types of anomalies. For the first two formulations, we show the underlying problem is in the computational class P by presenting linear and polynomial time exact algorithms. We show that the last formulation is computationally intractable, and we use an integer linear program for that version to generate experimental results. We demonstrate our work on several data sets such as the CelebA image data set, the HateXplain language data set, and the COMPAS dataset on fairness. These data sets are chosen as their ground truth explanations are clear or well-known.