Abstract: Computer-implemented systems and methods dynamically verify the trustworthiness of artificial-intelligence-generated content. A computing system receives a document generated by a large language model and extracts a reduced set of terms representing key truth-related arguments using statistical or language-model-based techniques. The reduced set is compared to corpus-derived terms obtained from related documents to identify matched, omitted, and additional arguments. Quantitative explainability metrics are computed based on relationships among the terms within the corpus, and a multidimensional trust representation is generated from the metrics. The trust representation may be stored as a structured data object and used to control subsequent content generation, filtering, or presentation by the computing system.

Abstract: A system develops classifiers that can aid medical professionals by diagnosing whether or not a patient has a medical condition like pneumothorax. The system breaks the task into multiple steps, using YOLOv4 to extract relevant regions of the video and a 3D sparse coding model to represent video features. Given the difficulty in acquiring positive training videos, the inventors trained a small-data classifier with a maximum of 15 positive and 32 negative examples. To counteract this limitation, the inventors leveraged subject matter expert (SME) knowledge to limit the hypothesis space, thus reducing the cost of data collection. The inventors present results using two lung ultrasound datasets and demonstrate that the inventors' model is capable of achieving performance on par with SMEs in pneumothorax identification.

Abstract: The present invention relates to an implantable device that may be used for bone fixation comprising a combination of a solid swelling copolymer and a porous swelling copolymer. The present invention also relates embodiments where the implantable device is a suture anchor, and to osteochondral defect repair systems comprising the implantable device. The anchor may be configured for anchoring a repair system to bone; and a joint interface component can be affixed to the anchoring component and may be provided with a smooth surface adapted for articulation of natural cartilage of opposing bone along said surface.

Abstract: The disclosure provides compounds of Formula (I) having the structure: Compounds of Formula (I) are fractalkine receptor agonists and useful in treating, preventing or minimizing metastasis in a subject diagnosed with cancer. The compounds of the disclosure are further useful in treating central nervous system diseases (such as, but not limited to, HIV Associated Neurocognitive Disorders (HAND), and/or Alzheimer's disease), pain, inflammation (such as, but not limited to, arthritis), cardiovascular disease (such as, but not limited to, undesired vascular smooth muscle proliferation, atherosclerosis, coronary vascular endothelial dysfunction, and/or coronary artery disease), and/or multiple sclerosis.

Abstract: The present disclosure is directed to methods for scrubbing low levels of urea from aqueous solutions such as a dialysate from dialysis, and including blood and blood products, and devices capable of employing these methods.

Abstract: The present invention(s) is directed to novel conductive Mn+1Xn(Ts) compositions exhibiting high volumetric capacitances, and methods of making the same. The present invention(s) is also directed to novel conductive Mn+1Xn(Ts) compositions, methods of preparing transparent conductors using these materials, and products derived from these methods.

Abstract: The present application relates to a multifunctional-damage responsive biofiber and methods of forming thereof, comprising: a core comprising a polymeric fiber; a crosslinked endospore loaded hydrogel layer coating the core, wherein the hydrogel layer is formed with: a solution comprising endospores, one or more anionic polymers, and one or more crosslinking agents; and a (co)polymer shell encapsulating the hydrogel layer, wherein the shell has one or more layers which may be the same or different, and each of the one or more layers is formed with a (co)polymer selected from the group consisting of nitrocellulose (NITR), epoxy Resin (ER), polymethylmethacrylate (PMMA), polyvinylidene fluoride (PVDF), cyanoacrylate adhesive (CYA), polystyrene (PS), polylactic acid (PLA), and combinations thereof. This application also relates to methods for self-healing a matrix material including impregnating a matrix material with the multi-functional damage-responsive biofiber.

Abstract: Systems and methods for detecting malware on routers use transformer-based anomaly detection. System call data and network packet data are collected from a router and transformed into vector embeddings using embedding models such as sys2vec or net2vec. The embeddings are processed by a transformer-based encoder, such as a BERT model, trained with contrastive or contrastive adversarial learning using anchor, positive, and negative samples derived from benign behavior. The resulting model distinguishes anomalous activity, including zero-day and network-intensive malware, from benign behavior with low false positive rates. Detection occurs at the router or hub level, reducing computational overhead on resource-constrained Internet of Things (IoT) devices and enabling real-time protection without reliance on cloud-based solutions.

Abstract: Provided are methods of effecting cation exchange in MXene materials so as to stabilize the materials. Also provided are compositions, comprising layered MXene materials that comprise an organic cation between layers. Also provided are MXene compositions comprising a chalcogen disposed thereon, the MXene composition further optionally comprising a quaternary ammonium halide disposed thereon.

Abstract: Backpropagation enabled neural network optimization, but is still expensive at scale. An efficient alternative for neural networks must reduce scaling costs and provide high-efficiency optimizations. The presented solution for feed-forward layers is extended compositionally, for transformers containing feed-forward and self-attention layers. These train complex models modularized by self-attentive feed-forward units (SAFFUs), defining efficient architectures that generalize over much less data. Results demonstrate outperformance of backpropagation (alone), compared to application of backpropagation after explicit solutions, which uncover better optima from less data. Ablations train a roadmap of 245 transformers to determine specifications for the SAFFU-transformer, showing several architectural variants appear quite performant. The most performant models appear to not be the most parameterized.

Abstract: Provided are 5-layered MXene materials having the formulas M5X4Tx; (M?aM?b)X4Tx (where a+b=5); and (M?aM?b)5X4Tx (where a+b=1). Also provided are related methods, compositions, and applications.

Abstract: Methods of making expanded hexagonal boron nitride including steps of: a) intercalating a hexagonal boron nitride with one or more guest molecules to form an intercalated hexagonal boron nitride; and b) thermally shocking the intercalated hexagonal boron nitride to form the expanded hexagonal boron nitride. An expanded hexagonal boron nitride formed by the method.

Abstract: A framework and associated data schema is used for machine learning-based prediction of performance metrics in digital integrated circuit (IC) design. In one aspect, a computer-implemented method converts design data from an initial design stage into one or more graph representations comprising nodes and edges, each annotated with structural, spatial, and performance-related features. A graph neural network, selected from a graph convolutional neural network (GCN), a spatial graph convolutional neural network (SGCN), or a hybrid thereof, processes the graph to generate embeddings for predicting one or more downstream performance metrics, including arrival time, interconnect parasitic impedance, total power, total area, or slack violations. Graph types include netlist, timing path, interconnect, clock network, and fused multi-graph representations. The method supports multi-stage and iterative predictions, feature importance analysis, ensemble models, and composite loss functions.

Abstract: In absence of electrical approaches for realization of highly stable RF oscillator, opto-electronic oscillators (OEO) techniques are provided, where self-forced oscillation techniques using long optical delays demonstrate significant short-term and long-term frequency stability. Fully integrated opto-electronic oscillator chip (IOEC) may be the most efficient realization of an RF frequency synthesizer in terms of operation frequency (covering microwave and millimeter wave), size (<10 cm3), ruggedness to environmental effects of temperature (?40 to 80 C), vibration (up to 40 g), low timing jitter (<5 fs for 40 GHz carrier), and wall-plug efficiency (output power >10 dBm from under 1 W power). A free-running III-V (primarily InP) based multi-mode laser (MML) diodes is designed with large mode number (e.g., over 60 modes) and intermodal oscillation frequency compatible with desired RF carrier signal (e.g., 1-40 GHz).

Abstract: Described herein is a method of diagnosing Lyme disease in a subject. The method includes determining a glycosylation profile of a protein whose glycosylation profile is associated with Lyme disease in a subject, and comparing the glycosylation profile of the protein with a predetermined glycosylation profile of the protein indicating free of Lyme disease or a predetermined glycosylation profile of the protein indicating Lyme disease. Further described herein is a method of treating and/or ameliorating Lyme disease in a subject. The method includes diagnosing Lyme disease in a subject and, if the subject is diagnosed to have Lyme disease, administering to the subject an effective amount of compound effective for treating Lyme disease. Further described herein is a method of evaluating a treatment for Lyme disease. The method includes comparing a glycosylation profile of the patient with glycosylation profiles associated with successful/unsuccessful treatments.

Abstract: The present disclosure describes a strained dielectric material comprising at least one type of component containing a domain wall variant pattern, or superdomain structure, that is in phase-co-existence with, or in close phase proximity to, a paraelectric state achieved at zero electric field or over a finite range of non-zero electric field, wherein the at least one type of component comprises one or more of an in-plane sub-domain polarization component, a plane-normal sub-domain polarization component, or a solid solution of a ferroelectric.

Abstract: An alignment guide for total ankle replacement surgery and a method of creating and using the alignment guide. The alignment guide has a surface that interfaces with a corresponding surface of a superior aspect a talus of a patient, and as second surface portion that interfaces with a corresponding distal surface of a section of a tibia of the patient. When the alignment guide is in position, it maintains the ankle of the patient in a preselected position.

Abstract: A field-programmable analog array (FPAA) fabric includes a 6×6 matrix of configurable analog blocks (CABs). The implementation of programmable CABs eliminates the use of fixed analog subcircuits. A unique routing strategy is developed within the CAB units that supports both differential and single-ended mode circuit configurations. The bandwidth limitation due to the routing switches of each individual CAB unit is compensated for through the use of a switch-less routing network between CABs. Algorithms and methodologies facilitate rapid implementation of analog circuits on the FPAA. The proposed FPAA fabric provides high operating speeds as compared to existing FPAA topologies, while providing greater configuration in the CAB units as compared to switch-less FPAA. The FPAA core includes 498 programming switches and 14 global switchless interconnects, while occupying an area of 0.1 mm2 in a 65 nm CMOS process.

Abstract: In this paper, a novel field-programmable analog array (FPAA) is proposed to secure the intellectual property (IP) of analog and mixed-signal circuits. A obfuscation technique is developed to efficiently mask the topology of both differential mode and single-ended mode analog circuits.

Abstract: Methods, systems, and computer-readable media are used in graph-based machine learning of analog integrated circuits (ICs). In a first aspect, functional device pairs in transistor-level circuits are detected and classified using a hybrid approach combining a subgraph isomorphism algorithm, such as VF2, with a trained relational-graph convolutional network (RGCN). The VF2 algorithm identifies candidate pairs and initial categories, while the RGCN filters false positives using link prediction scores, preserving category labels for validated pairs. In a second aspect, a relational GraphSAGE model performs multi-class link prediction on a heterogeneous graph with netlist and functional relation edge types, labeling device pairs into analog primitive categories without technology-dependent features and merging overlapping pairs into larger functional groups.