26400 research outputs found
Sort by
Towards Fairer And Safe Ai: Uncovering And Interpreting Fairness Anomalies In Deep Neural Network Models
No full textThe recent advances in training deep neural networks (DNNs) have revolutionized thedevelopment of data-driven decision support software. As a result, fairness testing and verification approaches for DNNs have received considerable attention. Testing approaches, based on statistical analyses, aim to provide counterexamples to fairness, while verification approaches attempt to offer a proof of correctness. The notion of individual fairness is a well-accepted concept characterizing discrimination as the existence of a counterfactual individual, differing only in protected features, receiving a better algorithmic outcome. DNNs may encode several such counterfactual instances, and in extreme scenarios, overwhelm the analyst and hide critical instances. Moreover, the mere existence of such counterfactuals fails to provide workable information on the root cause of discrimination. We study a quantitative generalization of individual fairness, called k-unfairness, where counterexamples include the presence of k â?¥ 2 counterfactual instances. We show that this quantitative notion of individual fairness allows us to prioritize discriminatory instances, measure the sensitivity of DNNs to the protected attributes, and debug the patterns in fairness bugs with rich information. On the technical side, we propose a hybrid method that combines formal symbolic analysis (SMT and MILP solvers) to certify individual fairness with randomized search (random walks and simulated annealing) to search for instances with diverse explanations. This method brings the advantages of both techniques: it certifies the fairness requirements if no counterexample is found and quantifies discrimination, which is computationally challenging for symbolic analysis. We use random walks and simulated annealing strategies to guide the search and find inputs that maximize objectives like the sensitivity of DNNs to protected attributes. Our experiments show that some benchmarks manifest the maximum sensitivity, while others show some or no sensitivity to the protected attributes. We also find that decision trees provide intuitive explanations to understand circumstances when DNNs significantly discriminate against protected groups
Developing A Bioengineered Nanoparticle For Improving Oral Absorption Of Iron Supplements.
No full textIron deficiency (ID) and iron deficiency anemia (IDA) are widespread nutritional issues, affecting millions globally. Conventional iron supplements often suffer from low absorption rates and gastrointestinal side effects. We investigated β-glucan derivatives as potential carriers to enhance iron absorption and mitigate these drawbacks. This study explored a novel β-glucan-based carrier system loaded with ferrous sulfate heptahydrate. In-vitro studies demonstrated sustained iron stability for over six hours in simulated gastric fluids due to the carrier\u27s affinity for stomach mucin. Particle size analysis and scanning electron microscopy (SEM) images confirmed this specific binding. Additionally, drug release studies revealed a pH-dependent release profile, favoring iron delivery within the stomach. Oral administration of β-glucan-iron complexes demonstrated improved overall efficacy with minimal adverse effects, suggesting a promising strategy for improved iron supplementation. This research highlights the potential of β-glucan derivatives as carriers for oral iron delivery, offering enhanced absorption and potentially reducing gastrointestinal side effects. Further refinement holds promise for transitioning various injectable iron medications to a more patient-friendly oral form
From Normal Distribution to What? How to Best Describe Distributions with Known Skewness
No full textIn many practical situations, we only have partial information about the probability distribution -- e.g., all we know is its few moments. In such situations, it is desirable to select one of the possible probability distributions. A natural way to select a distribution from a given class of distributions is the maximum entropy approach. For the case when we know the first two moments, this approach selects the normal distribution. However, when we also know the third central moment -- corresponding to skewness -- a direct application of this approach does not work. Instead, practitioners use several heuristic techniques, techniques for which there is no convincing justification. In this paper, we show that while we cannot directly apply the maximum entropy approach to the skewness situation, we can apply it approximately -- with any approximation accuracy we want -- and get a meaningful answer to the above selection problem
Data Fusion Is More Complex Than Data Processing: A Proof
No full textEmpirical data shows that, in general, data fusion takes more computation time than data processing. In this paper, we provide a proof that data fusion is indeed more complex than data processing
Somewhat Surprisingly, (Subjective) Fuzzy Technique Can Help to Better Combine Measurement Results and Expert Estimates into a Model with Guaranteed Accuracy: Digital Twins and Beyond
No full textTo understand how different factors and different control strategies will affect a system -- be it a plant, an airplane, etc. -- it is desirable to form an accurate digital model of this system. Such models are known as digital twins. To make a digital twin as accurate as possible, it is desirable to incorporate all available knowledge of the system into this model. In many cases, a significant part of this knowledge comes in terms of expert statements, statements that are often formulated by using imprecise ( fuzzy ) words from natural language such as small , very possible , etc. To translate such knowledge into precise terms, Zadeh pioneered a technique that he called fuzzy. Fuzzy techniques have many successful applications; however, expert statements are subjective; in contrast to measurement results, they do not come with guaranteed accuracy. In this paper, we show that by using fuzzy techniques, we can translate imprecise expert knowledge into precise probabilistic terms -- thus allowing to combine this knowledge with measurement results into a model with guaranteed accuracy
Study Of Morphology, Structure, And Magnetic Properties Of High Entropy Alloy Systems
Get PDFAbstract:This thesis presents an investigation into the magnetic properties of high entropy alloy (HEA) samples, focusing primarily on compositions within the FeNiCoMn system and their gallium/chemically doped derivatives. High entropy alloys represent a promising class of materials with unique properties resulting from their complex composition and structural arrangements. The behavior of these alloys, particularly in their magnetic properties, remains an area of significant interest due to their potential applications in magnetic storage, sensing, and other technological domains. The study encompasses an examination of seven distinct HEA samples: FeNiCo, FeNiCoMn0.2, FeNiCoMn0.4, FeNiCoMn0.6, FeNiCoMn0.2Ga0.1, FeNiCoGa0.2, FeNiCoMn0.2Ga0.3. Furthermore, three samples of FeNiCoMn0.2 and three samples of FeNiCoMn0.2Ga0.1 are examined after undergoing chemical dopping using Fe, Ni, and Co. Through the combination of X-ray diffraction (XRD), vibrating sample magnetometry (VSM), and electron microscopy, the magnetic properties and microstructural characteristics of these samples are analyzed. The FeNiCoMn0.2-0.6 series, with varying manganese content, serves as a baseline for understanding the influence of composition on magnetic behavior. Subsequently, gallium (Ga) is introduced into the system, both independently and in conjunction with manganese, in the FeNiCoMn0.2Ga0.1-0.3 and FeNiCoGa0.2 samples, respectively. The effects of gallium incorporation on the magnetic properties are examined, shedding light on the diminished or enhanced magnetic properties of the sample. Introducing higher concentrations of Fe, Ni, and Co through the use of chemical doping is performed to further study how the samples magnetic properties can be influenced and/or modified to suit the desired need of the researcher
Mav Localization In Gps-Denied Environments And Synthetic Data Collection In Challenging Simulated Conditions
Get PDFThe development of unmanned aerial systems presents an opportunity for conducting industrial inspections in environments where traditional navigation systems, such as the Global Navigation Satellite System (GNSS), are compromised. This dissertation investigates the implementation of a micro aerial vehicle (MAV) capable of autonomous data acquisition in complex, GNSS-degraded industrial settings. The primary challenge addressed is the robust localization of MAVs, a critical aspect in ensuring reliable operation under varying and uncertain environmental conditions.
The work is divided into two main parts. The first part focuses on the design and integration of a MAV system specifically for power plant inspections in GPS-denied environments. This system leverages vision-based sensors to achieve reliable and accurate pose tracking, overcoming the limitations encountered in GPS-degraded environments of real-world industrial settings. The physical system enabled the acquisition of inspection data from hard-to-reach locations, mitigating the risks associated with manual inspections and enhancing safety.
The second part of this dissertation develops a realistic simulation framework that allows for the collection of large synthetic datasets. These datasets are required for training a deep neural network (DNN) for environment classification and sensor failure detection, addressing the challenges posed by unreliable perception data in unstructured environments. The resulting data was collected in varied locations under different challenging simulated conditions, such as low light, dusty environments, or settings with lighting that creates sudden contrast changes.
Key contributions of this work include the successful deployment of an autonomous MAV for power plant inspections in GPS-denied environments and the creation of a significant dataset with the potential to support the development of AI-driven localization and inspection techniques. This dataset, comprised of simulated perceptual data, is packed in a format that simplifies distribution in order to further research and development in the field.
By focusing on the robustness of MAV localization for safe and efficient application in challenging environments, this research confirms the viability of MAVs as tools for autonomous data collection and validates the effectiveness of realistic simulations in creating synthetic data that can be used to improve machine learning models for autonomous systems. The outcomes of this study are expected to significantly advance the capabilities of autonomous drones in industrial applications, particularly in settings where GPS is unreliable or unavailable
How to Propagate Uncertainty via AI Algorithms
Get PDFAny data processing starts with measurement results. Measurement results are never absolutely accurate. Because of this measurement uncertainty, the results of processing measurement results are, in general, somewhat different from what we would have obtained if we knew the exact values of the measured quantities. To make a decision based on the result of data processing, we need to know how accurate is this result, i.e., we need to propagate the measurement uncertainty through the data processing algorithm. There are many techniques for uncertainty propagation. Usually, they involve applying the same data processing algorithm several times to appropriately modified data. As a result, the computation time for uncertainty propagation is several times larger than data processing itself. This is a very critical issue for data processing algorithms that take a lot of computational steps -- such as modern deep learning-based AI techniques, for which a several-times increase in computation time is not feasible. At first glance, the situation may seem hopeless. Good news is that there is another problem with modern AI algorithms: usually, once they learn, their weights are frozen, and they stop learning -- as a result, the quality of their answers decreases with time. This is good news because, as we show, solving the second problem -- by allowing at least one learning step for each new use of the model -- helps to also come up with an efficient uncertainty propagation algorithm