University of Illinois Urbana-Champaign
Illinois Digital Environment for Access to Learning and Scholarship RepositoryNot a member yet
123813 research outputs found
Sort by
Failure Prediction from Limited Hardware Demonstrations
Get PDFPrediction of failures in real-world autonomous systems either requires accurate model information or extensive testing. Partial knowledge of the system model makes simulation-based failure prediction unreliable. Moreover, obtaining such demonstrations is expensive, and could potentially be risky for the autonomous system to repeatedly fail during data collection. This work presents a novel three-step methodology for discovering failures that occur in the true system by using a combination of a limited number of demonstrations from the true system and the failure information processed through sampling-based testing of a model dynamical system. Given a limited budget of demonstrations from true system and a model dynamics (with potentially large modeling errors), the proposed methodology comprises of a) exhaustive simulations for discovering algorithmic failures using the model dynamics; b) design of initial demonstrations of the true system using Bayesian inference to learn a Gaussian process regression (GPR)-based failure predictor; and c) iterative demonstrations of the true system for updating the failure predictor. To illustrate the efficacy of the proposed methodology, we consider: a) the failure discovery for the task of pushing a T block to a fixed target region with UR3E collaborative robot arm using a diffusion policy; and b) the failure discovery for an F1-Tenth racing car tracking a given raceline under an LQR control policy
An Upper Bound for Interactive ECCs with Separate Budgets
Get PDFWe study interactive error-correcting codes (iECCs) in the binary communication model with separate corruption budgets for both the directions of communication. Interactive error-correcting codes (iECCs) allow two parties, Alice and Bob, to communicate over a noisy channel through an interactive protocol, with the goal of communicating a message from Alice to Bob even when a bounded fraction of the communication is adversarially corrupted. In the standard one-way communication setting, previous works [1] show that an error fraction α of bit flips can be corrected to recover the original message if and only if α is bounded by 1/4 . Berlekamp [2] showed that this fraction improves to 1/3 with interaction using noiseless feedback. Elfremenko, Kol, Saxena and Zhang [3] further construct interactive error-correcting codes (iECCs) that can uniquely decode from an error fraction of 1/4 +10 [to the power of] −5 even in the presence of noisy feedback while the best known impossibility upper bound for this setting is 13/47 [4]. In order to better understand this gap, we study the communication model with independent corruption budgets for both sides of communication and show that no iECC can be resilient to more than 7/24 fraction of errors in this model
Regionally Robust Autoencoder-Based Channel Codes
Get PDFIn this paper, we propose a novel training scheme to improve the robustness of end-to-end autoencoder-based channel codes against universal adversarial perturbations (UAP) and jamming attacks. Our training algorithm improves robustness by augmenting clean training data with uniformly sampled perturbations from a bounded l2-ball, encouraging consistent decoder outputs within these regions. Our contributions include the design of the training algorithm, a robustness comparison with standard autoencoders, and a benchmark against a traditional BPSK system with Hamming coding and maximum likelihood decoding. Results show that our approach enhances robustness to UAP and jamming attacks
How Does Antibiotic Resistance Work? The Mechanisms of Antibiotic Resistance
No full textAntibiotic resistance occurs when bacteria survive in the presence of antibiotics, which are generally used to kill bacteria. This resistance has become a growing issue in several industries, including healthcare, where bacterial infections have become more difficult to treat. With antibiotic resistance rising to be a threat to the health of humans and animals, it is essential to understand how antibiotics work and how bacteria are able to bypass their effects. Antibiotics primarily serve their purpose by breaking apart the bacteria or interrupting necessary processes for survival. Bacteria, once they acquire resistance through either mutation or uptake of foreign DNA, use many strategies to counteract antibiotics. Currently, there are various approaches being researched to prevent or work around antibiotic resistance. With the increasing prevalence of resistance, this research is crucial to reverse its impact
Evidence for Evolution: A Brief Review
No full textEvolutionary theory offers a unifying framework for understanding the diversity of life on Earth. Formalized by Charles Darwin as "descent with modification" through natural selection, the concept has since been supported by extensive evidence across multiple scientific disciplines. Paleontological records reveal transitional forms and chronological patterns of increasing complexity. Comparative anatomy highlights homologous structures that point to common ancestry, while embryology demonstrates shared developmental pathways among diverse species. Molecular biology confirms genetic relationships through DNA sequence comparisons, and biogeography illustrates the role of geographic isolation in speciation. Real-time evolutionary changes, such as the emergence of antibiotic resistance, provide observable instances of natural selection. Additionally, vestigial structures and pseudogenes offer insight into evolutionary history and inherited traits that have lost their original function. Collectively, these lines of evidence converge to substantiate evolution as a foundational principle in modern biology
The Critical Window for Estrogen Replacement Therapy in Menopausal Women: Exploring the Neuroprotective Effects of Estrogen in Reducing Dementia Risk
No full textMenopause, characterized by a significant decline in estrogen levels, profoundly impacts brain health, influencing cognitive functions and neurobiological integrity. Estrogen replacement therapy (ERT), especially when administered within a "critical window" near the onset of menopause, has shown promise in mitigating cognitive decline and reducing dementia risk. This paper explores the neurobiological mechanisms underlying estrogen’s protective role, including its effects on mitochondrial health, synaptic plasticity, neurogenesis, and amyloid-beta plaque clearance. By synthesizing current literature, we underscore the importance of early ERT initiation and highlight the implications of the critical window hypothesis for optimizing cognitive outcomes in aging women
Digital Disconnection in Public Spaces
No full textThis study explores how people use digital devices and negotiate social presence in everyday public spaces. Four naturalistic observations were conducted in Champaign and Urbana at a fast-food restaurant, two cafés, and a transportation terminal. Using a non-participant observer role and detailed field notes, the study documented patterns of device use, attention, and interaction across diverse settings. Data were analyzed inductively through open coding, which generated categories describing different forms of digital engagement. The findings show that device use was common in all locations but shaped interaction in different ways. At McDonald’s, screens often replaced conversation and contributed to quiet routines. At Espresso Royale, technology supported individual work and quiet coexistence. On the university campus, students integrated laptops and phones into group study, producing flexible patterns of multitasking and shared engagement. At the Illinois Terminal, device use offered comfort during long waits and occasionally created brief shared moments among strangers. These observations suggest that digital disconnection varies across settings and depends on the social expectations and purposes of each environment. The study highlights the value of observing real-world behavior to better understand how technology shapes interaction, presence, and public life
OSTINATOS IN JAZZ DRUM SET IMPROVISATION: A PEDAGOGICAL AND CONCEPTUAL GUIDE FROM MAX ROACH TO MARCUS GILMORE
No full textThis study undertakes a pedagogical exploration of ostinatos in jazz drum set improvisation, tracing their evolution from the swing era to contemporary performance practice. Through the analysis of ostinato applications from drummers Max Roach, Bill Stewart, Joseph Farnsworth, and Marcus Gilmore, this paper categorizes ostinatos into three types—simple timekeeping, layered and polyrhythmic, and “created clave”—to examine their distinct functions
within improvisational contexts. Rooted in jazz’s intercultural rhythmic heritage, the work argues for the ostinato as a powerful improvisational tool, capable of enhancing accessibility and danceability without necessitating a reduction in complexity. Through transcription, analysis, and original exercises, this guide offers performing drummers both technical and conceptual
frameworks to integrate ostinatos into modern jazz improvisation. Ultimately, it reframes the ostinato not as a clichéd function of specific styles, but as a flexible and expressive device that encourages rhythmic exploration and audience connection
Advanced Sustainable Manufacturing: Free Training
No full textThis flier advertised ISTC's Collaborative Training and Assessments in Pollution Prevention (CTAP2), funded through a grant from the U.S. Environmental Protection Agency. The program was designed to train individuals on best practices in assessing their operations to uncover areas that may be improved by implementing sustainability methods, which in turn result in cost reduction and reduced environmental impacts. ISTC staff offered pollution prevention training guided them through the assessment process at their facilities. The program focused on continuous improvement using tools to implement changes over time.U.S. Environmental Protection Agency ; Assistance Agreement No. 4U – 00E03255 -