Improved Information-Theoretic Generalization Bounds for Distributed,Federated, and Iterative Learning

We consider information-theoretic bounds on the expected generalization error for statistical learning problems in a network setting. In this setting, there are K nodes, each with its own independent dataset, and the models from the K nodes have to be aggregated into a final centralized model. We consider both simple averaging of the models as well as more complicated multi-round algorithms. We give upper bounds on the expected generalization error for a variety of problems, such as those with Bregman divergence or Lipschitz continuous losses, that demonstrate an improved dependence of 1/K on the number of nodes. These “per node” bounds are in terms of the mutual information between the training dataset and the trained weights at each node and are therefore useful in describing the generalization properties inherent to having communication or privacy constraints at each node.     

Misuse of Cardiac Lipid upon Exposure to Toxic Trace Elements—A Focused Review

Heavy metals and metalloids like cadmium, arsenic, mercury, and lead are frequently found in the soil, water, food, and atmosphere; trace amounts can cause serious health issues to the human organism. These toxic trace elements (TTE) affect almost all the organs, mainly the heart, kidney, liver, lungs, and the nervous system, through increased free radical formation, DNA damage, lipid peroxidation, and protein sulfhydryl depletion. This work aims to advance our understanding of the mechanisms behind lipid accumulation via increased free fatty acid levels in circulation due to TTEs. The increased lipid level in the myocardium worsens the heart function. This dysregulation of the lipid metabolism leads to damage in the structure of the myocardium, inclusive fibrosis in cardiac tissue, myocyte apoptosis, and decreased contractility due to mitochondrial dysfunction. Additionally, it is discussed herein how exposure to cadmium decreases the heart rate, contractile tension, the conductivity of the atrioventricular node, and coronary flow rate. Arsenic may induce atherosclerosis by increasing platelet aggregation and reducing fibrinolysis, as exposure interferes with apolipoprotein (Apo) levels, resulting in the rise of the Apo-B/Apo-A1 ratio and an elevated risk of acute cardiovascular events. Concerning mercury and lead, these toxicants can cause hypertension, myocardial infarction, and carotid atherosclerosis, in association with the generation of free radicals and oxidative stress. This review offers a complete overview of the critical factors and biomarkers of lipid and TTE-induced cardiotoxicity useful for developing future protective interventions.     

Fluorescent Molecular Logic Gates and Pourbaix Sensors in Polyacrylamide Hydrogels

Polyacrylamide hydrogels formed by free radical polymerisation were formed by entrapping anthracene and 4-amino-1,8-naphthalimide fluorescent logic gates based on photoinduced electron transfer (PET) and/or internal charge transfer (ICT). The non-covalent immobilisation of the molecules in the hydrogels resulted in semi-solid YES, NOT, and AND logic gates. Two molecular AND gates, examples of Pourbaix sensors, were tested in acidic aqueous methanol with ammonium persulfate, a strong oxidant, and displayed greater fluorescence quantum yields than previously reported. The logic hydrogels were exposed to aqueous solutions with chemical inputs, and the fluorescence output response was viewed under 365 nm UV light. All of the molecular logic gates diffuse out of the hydrogels to some extent when placed in solution, particularly those with secondary basic amines. The study exemplifies an effort of taking molecular logic gates from homogeneous solutions into the realm of solid-solution environments. We demonstrate the use of Pourbaix sensors as pE-pH indicators for monitoring oxidative and acidic conditions, notably for excess ammonium persulfate, a reagent used in the polymerisation of SDS-polyacrylamide gels.     

Coordination-induced bond weakening of water at the surface of an oxygen-deficient polyoxovanadate cluster

Hydrogen-atom (H-atom) transfer at the surface of heterogeneous metal oxides has received significant attention owing to its relevance in energy conversion and storage processes. Here, we present the synthesis and characterization of an organofunctionalized polyoxovanadate cluster, (calix)V₆O₅(OH₂)(OMe)₈ (calix = 4-tert-butylcalix[4]arene). Through a series of equilibrium studies, we establish the BDFE(O–H)_avg of the aquo ligand as 62.4 ± 0.2 kcal mol⁻¹, indicating substantial bond weaking of water upon coordination to the cluster surface. Subsequent kinetic isotope effect studies and Eyring analysis indicate the mechanism by which the hydrogenation of organic substrates occurs proceeds through a concerted proton–electron transfer from the aquo ligand. Atomistic resolution of surface reactivity presents a novel route of hydrogenation reactivity from metal oxide surfaces through H-atom transfer from surface-bound water molecules.

Coordination-induced bond weakening facilitates H-atom transfer from polyoxovanadate surface.     

Mode Localization Induced by a Nonlinear Control Loop

We present the analysis of a nonlinear controlsystem that is used to excite and maintain a specified amplitude ofoscillation in the Jet Propulsion Laboratory vibratory gyroscope.This experimental application shows that nonlinear localizationthrough active means can be implemented in a practical system when itis desirable to confine the response to a favorable mode. Theclosed-loop system response predicted by the model shows very closeagreement with the experimental results for a significant range ofcontroller parameters. We also experimentally demonstrate that theactively localized motion is eliminated through bifurcation, similarto what was observed in previous passive localization studies appliedto extended flexible oscillators.     

Chess AI: Competing Paradigms for Machine Intelligence

Endgame studies have long served as a tool for testing human creativity and intelligence. We find that they can serve as a tool for testing machine ability as well. Two of the leading chess engines, Stockfish and Leela Chess Zero (LCZero), employ significantly different methods during play. We use Plaskett’s Puzzle, a famous endgame study from the late 1970s, to compare the two engines. Our experiments show that Stockfish outperforms LCZero on the puzzle. We examine the algorithmic differences between the engines and use our observations as a basis for carefully interpreting the test results. Drawing inspiration from how humans solve chess problems, we ask whether machines can possess a form of imagination. On the theoretical side, we describe how Bellman’s equation may be applied to optimize the probability of winning. To conclude, we discuss the implications of our work on artificial intelligence (AI) and artificial general intelligence (AGI), suggesting possible avenues for future research.     

Purification and Characterization of Gum-Derived Polysaccharides of Moringa oleifera and Azadirachta indica and Their Applications as Plant Stimulants and Bio-Pesticidal Agents

Plant gums are bio-organic substances that are derived from the barks of trees. They are biodegradable and non-adverse complex polysaccharides that have been gaining usage in recent years due to a number of advantages they contribute to various applications. In this study, gum was collected from Moringa oleifera and Azadirachta indica trees, then dried and powdered. Characterizations of gum polysaccharides were performed using TLC, GC-MS, NMR, etc., and sugar molecules such as glucose and xylose were found to be present. Effects of the gums on Abelmoschus esculentus growth were observed through root growth, shoot growth, and biomass content. The exposure of the seeds to the plant gums led to bio stimulation in the growth of the plants. Poor quality soil was exposed to the gum polysaccharide, where the polysaccharide was found to improve soil quality, which was observed through soil analysis and SEM analysis of soil porosity and structure. Furthermore, the plant gums were also found to have bio-pesticidal activity against mealybugs, which showed certain interstitial damage evident through histopathological analysis.     

Isolation of a Novel Polyketide from Neodidymelliopsis sp.

Fungi have become an invaluable source of bioactive natural products, with more than 5 million species of fungi spanning the globe. Fractionation of crude extract of Neodidymelliopsis sp., led to the isolation of a novel polyketide, (2Z)-cillifuranone (1) and five previously reported natural products, (2E)-cillifuranone (2), taiwapyrone (3), xylariolide D (4), pachybasin (5), and N-(5-hydroxypentyl)acetamide (6). It was discovered that (2Z)-cillifuranone (1) was particularly sensitive to ambient temperature and light resulting in isomerisation to (2E)-cillifuranone (2). Structure elucidation of all the natural products were conducted by NMR spectroscopic techniques. The antimicrobial activity of 2, 3, and 5 were evaluated against a variety of bacterial and fungal pathogens. A sodium [1-¹³C] acetate labelling study was conducted on Neodidymelliopsis sp. and confirmed that pachybasin is biosynthesised through the acetate polyketide pathway.     

Comparison of model predictions with measured proton-induced production of nickel and iridium isotopes

Journal of Radioanalytical and Nuclear Chemistry - Calculations were performed to support method development for simultaneous production of Ni and Ir isotopes. The work scope included development...     

Physicochemical properties and bio-interfacial interactions of surface modified PDLLA-PAMAM linear dendritic block copolymers

Here, we demonstrate the applicability of self-assembling linear-dendritic block copolymers (LDBCs) and their nanoaggregates possessing varied surfaces as therapeutic nanocarriers. These LDBCs are comprised of a hydrophobic, linear polyester chemically coupled to a hydrophilic dendron polyamidoamine (PAMAM)—the latter of which acts as the surface of the self-assembled nanoaggregate in aqueous media. To better understand how surface charge density affects the overall operability of these nanomaterials, we modified the nanoaggregate surface to yield cationic (NH₃⁺), neutral (OH), and anionic (COO⁻) surfaces. The effect of these modifications on the physicochemical properties (i.e., size, morphology, and surface charge density), colloidal stability, and cellular uptake mechanism of the polymeric nanocarrier were investigated. This comparative study demonstrates the viability of nanoaggregates formed from PDLLA-PAMAM LDBCs to serve as nanocarriers for applications in drug delivery.