Search for: All records

Active Learning (AL) has gained prominence in integrating data-intensive machine learning (ML) models into domains with limited labeled data. However, its effectiveness diminishes significantly when the labeling budget is low. In this paper, we first empirically observe the performance degradation of existing AL algorithms in the low-budget settings, and then introduce Direct Acquisition Optimization (DAO), a novel AL algorithm that optimizes sample selections based on expected true loss reduction. Specifically, DAO utilizes influence functions to update model parameters and incorporates an additional acquisition strategy to mitigate bias in loss estimation. This approach facilitates a more accurate estimation of the overall error reduction, without extensive computations or reliance on labeled data. Experiments demonstrate DAO's effectiveness in low budget settings, outperforming state-of-the-arts approaches across seven benchmarks. 

The current paper discusses conversation-based assessments (CBAs) created with prompt engineering for LLMs based on Evidence-Centered Design (ECD). Conversation-based assessments provide students the opportunity to discuss a given topic with artificial agent(s). These conversations elicit evidence of students’ knowledge, skills and abilities that may not be uncovered by traditional tests. We discuss our previous method of creating such conversations with regular expressions and latent semantic analysis in an expensive methodology requiring time and various expertise. Thus, in this novel work, we created a prompt-engineered version of CBAs based on evidence-centered design that remains on the domain topic throughout the conversation as well as provides evidence of the student knowledge in a less expensive way. We present the methodology for creating these prompts, compare responses to various student speech acts between the previous version and the prompt engineered version, and discuss the evidence gleaned from the conversation and based on the prompt. Finally, limitations, conclusions and implications of this work are discussed. 

In this work, we develop an open-source surgical simulation environment that includes a realistic model obtained by MRI-scanning a physical phantom, for the purpose of training and evaluating a Learning from Demonstration (LfD) algorithm for autonomous suturing. The LfD algorithm utilizes Dynamic Movement Primitives (DMP) and Locally Weighted Regression (LWR), but focuses on the needle trajectory, rather than the instruments, to obtain better generality with respect to needle grasps. We conduct a user study to collect multiple suturing demonstrations and perform a comprehensive analysis of the ability of the LfD algorithm to generalize from a demonstration at one location in one phantom to different locations in the same phantom and to a different phantom. Our results indicate good generalization, on the order of 91.5%, when learning from more experienced subjects, indicating the need to integrate skill assessment in the future. 

Double heterojunction nanorods enable both electroluminescence and light harvesting capabilities within the same device structure, providing a promising platform for energy-scavenging displays and related applications. However, the efficiency of the photovoltaic mode remains modest for useful power conversion and may be challenging to improve without sacrificing performance in electroluminescence. Through a facile on-film partial ligand exchange with benzenethiol integrated into the device fabrication step, we achieve an average of more than threefold increase in power conversion efficiency while maintaining the maximum external quantum efficiency and the maximum luminance in the LED mode. The improved photovoltaic performance is mainly due to the increase in the short circuit current, which we attribute to the enhanced charge separation afforded by the partial ligand exchange. The recovery of the photoluminescence lifetime under the forward bias suggests that the hole traps introduced by benzenethiols are filled prior to reaching the voltage at which light emission begins, allowing LED performance to be maintained and possibly improved. 

Collaborative learning can improve student learning, student persistence, and the classroom climate. While work has documented the tradeoffs of face-to-face collaboration and asynchronous, online learning, the trade-offs between asynchronous (student-scheduled) and synchronous (instructor-scheduled) collaborative and online learning have not been explored. Structured roles can maximize the effectiveness of collaborative learning by helping all students participate, but structured roles have not been studied in online settings. We performed a quasi-experimental study in two courses—Computer Architecture and Numerical Methods—to compare the effects of asynchronous collaborative learning without structured roles to synchronous collaborative learning with structured roles. We use a data-analytics approach to examine how these approaches affected the student learning experience during formative collaborative learning assessments. Teams in the synchronous offering made higher scoring submissions (5-10% points better on average), finished assessments more efficiently (11-16 minutes faster on average), and had greater equality in the total number of submissions each student made (for example, significant increase of 13% in the mean equality score among all groups).