93887 research outputs found
Sort by
Ensemble Feature Selection for Network Intrusion Detection Systems Using Explainable AI: A Frequency-Based Approach
Full text linkWe propose an XAI-based ensemble feature selection method combining XAI methods to reduce bias in IDS. A frequency-based approach identifies key features, validated on CICIDS-2017 with improved accuracy and efficiency. Our framework enhances robustness and interpretability in intrusion detectio
Using Speakers as Sensors: Detecting Acoustic Loads with Dense Neural Networks and Impedance Features
Full text linkThe use of speakers as sensors to detect ear canal conditions has been previously demonstrated using variable syringe lengths attached to earphones. Building on this foundation, our research explores the potential of a single speaker to function as both an actuator and sensor by leveraging electrical impedance measurements under varying acoustic loads, analyzed with Dense Neural Networks (DNN). Electrical impedance data, including magnitude and phase, were collected from four speakers across fourteen distinct acoustic load conditions, yielding a dataset of 5,600 samples. The raw data were processed through the DNN model, achieving an 87% accuracy in length prediction independent of speaker types, which improved to 91% when incorporating speaker-specific characteristics
Charging Indiana’s Crossroads
Full text linkJoin us for an insightful update on INDOT’s Charging the Crossroads Program. Speakers will highlight the latest program activities, key milestones and challenges with Round 1 and Round 2 projects, and offer a glimpse into activities beyond the stations themselves. Learn how Indiana is paving the way for a sustainable, accessible, and equitable deployment of DC fast charging infrastructure.
Presentation Speakers/Participants George McCue, INDOT, [email protected] Jack Sinton, HNTB Corporation, [email protected] Kerri Garvin, HNTB Corporation, [email protected]
The Lloyd4U: Taking Innovative Intersections From Concept to Construction
Full text linkTheLloyd4U includes more than a dozen improvement projects along the Lloyd Expressway in Evansville to make the busy roadway safer and more efficient. The first displaced left turn intersection in Evansville opened in 2024 with innovative outreach strategies to prepare motorists. This talk will walk you through practices that have worked well and what’s still to come with a project that stretches across Vanderburgh County
National Wildlife Crossings Program: Enhancing Roadway Safety
Full text linkWildlife crossings are an emerging creative infrastructure solution that enhance road safety by providing safe passage for wildlife across busy highways, which significantly lowers the risk of severe wildlife-vehicle collisions. Join us as Lochner’s Roadway Ecologists share their nationwide experience with wildlife crossing projects and the funding sources available to bring these solutions to Indiana. By learning from successful models across the country, we can protect ecosystems while improving roadway safety
CSGP/Rule 5 Erosion Control Tips and Tricks
Full text linkThis presentation will walk through the revised erosion control rule, highlight commonly overlooked items, and provide tips for working through this permit
Writing Confidence: Tutoring, Identity, and Race—A Mixed-Methods Approach
Full text linkThis mixed-methods study sought to better understand how confidence in writing and race interact as factors within writing centers. Students utilizing our writing center were asked to provide data about racial identity and writing confidence both when registering with the writing center and when completing postsession surveys. From this data, we interviewed a racially representative pool of respondents to better understand their definitions of confidence and the identity factors that have shaped their confidence in writing. Our survey data showed that students’ confidence increased significantly as a result of a writing center session, replicating previous writing center research. Furthermore, we found that improvements in confidence were consistent across racial identities, with students from different racial backgrounds reporting comparable gains. Our qualitative interview results revealed how students struggle with both identity-and non-identity- based factors that lower their confidence in academic writing. Results offer a more nuanced picture of how student identity impacts writing confidence both within and outside the writing center
Probing Surface Characteristics of Plant-based Protein Powders
Full text linkPlant-based protein powders are expected to take-on a larger role in human diet as concerns regarding animal-based protein production grow. Unfortunately, not much physical property information about these proteins is available. This study was an attempt by researchers to perform an initial investigation of five powders: Mung bean, peas, pumpkin seed, Fava bean, and Soybeans. Typical particle sizing was determined, along with the surface morphology. Shear testing on the bulk materials was performed to establish the powders’ potential flowability. Inverse gas chromatography provided surface energies of the five protein types, and an X-ray photoelectron spectrograph was used to determine the specific chemical composition of each particle. Further work refining these properties is expected, but this study has provided a baseline of material values for benchmarking new results
The N-Alternative Optional Choice Experiment
Full text linkWhen testing visual function in clinical settings, one encounters test-takers who have difficulty following a forced-choice instruction when stimulus intensity is low. “I didn’t see anything, why do I have to guess?” An optional-choice paradigm makes it easier for these test-takers to provide data. Here we describe a theory of signal detection for N-alternative optional-choice (N-AOC) psychophysical tasks. The theory generalizes two well-known paradigms into a single framework: the theory for yes-no tasks is a degenerate case in which N=1, and the theory for N-alternative forced choice tasks (N-AFC) is a degenerate case in which the decision criterion is liberal. A different generalization, that can be applied either to yes-no or to N-AFC (and that we were reminded of by Jeffrey Mulligan), is to collect the observer’s confidence judgment on each trial. Confidence judgments are easily incorporated into N-AOC as well
Attention Improves Classification Performance of Locations in Ventral and Dorsal Visual Pathways
Full text linkVisual processing is divided into a ventral and dorsal pathway, thought to be important for object recognition and spatial cognition, respectively. Yet both streams encode shape and spatial information, and both are modulated by attention and task demands. This study compared how attention influences fMRI bold responses in a ventral region (fusiform gyrus, FG) and dorsal region (superior parietal gyrus, SPG). Specifically, using multivariate pattern analysis within each region, we examined whether attention modulated spatial location classification performance. During fMRI scans, participants performed one of 3 possible tasks: a 1-back task under attend-to-shape (detect a shape repetition); a 1-back task under attend-to-location (detect a location repetition); and a passive task (detect a color change of the fixation point). Each task had identical stimulus presentations, in order to isolate attentional effects. Spatial location classification was performed using support vector machine (SVM) across the regions of interest, including FG in the ventral pathway and SPG in the dorsal pathway. FG and SPG each showed higher spatial classification accuracies in both the attend-to-shape and attend-to-location conditions compared to chance levels for SVM classification approach. In addition, FG and SPG each showed higher spatial classification accuracies in both attention conditions compared to the passive condition for SVM classification algorithm. These results suggest that attention, either to the shape or location of a stimulus, enhances spatial representations not only in the dorsal pathway but also in the ventral pathway, supporting the view that spatial representations are modulated by attention along both streams. The findings, including why attention to shape improves spatial classification, will be discussed