Technical Disclosure Common
Not a member yet
9157 research outputs found
Sort by
SELF-LEARNING, LLM-DRIVEN, EXPLAINABLE, PRIVACY-AWARE, LEARNING AGENT FOR INTELLIGENT MODEL RECOMMENDATIONS (SELF-LLM-XPLAINER)
No full textThe present disclosure relates to a system and method for model recommendation, more particularly, a self-learning, LLM-driven, explainable, privacy-aware, learning agent for intelligent model recommendations (SELF-LLM-XPLAINER). The present disclosure suggests ingesting a dataset provided by a user via an online interface. Thereafter, the present disclosure suggests profiling the ingested dataset to generate standardized metadata. Subsequently, assessing the standardized metadata and sampled data values to detect Personally Identifiable Information (PII) and to generate a privacy risk map. Upon generating the privacy risk map, the present disclosure suggests identifying a machine learning task corresponding to the dataset based on the standardized metadata and one or more sample values. Further, the present disclosure suggests analyzing the standardized metadata and identifying task metadata to generate a recommendation model. As a result, the present disclosure provides a self-learning, agentic system for intelligent recommendation of machine learning models and pipelines
HOLONOMIC ELECTRIC PALLET JACK FOR NARROW AISLE USE
No full textA specialized material handling apparatus provides improved efficiency and safety of moving heavy information technology (IT) equipment, such as fully populated computing racks, within constrained operational areas, such as narrow aisles found in datacenters. The weight of high-density computing racks, often exceeding 4,000 pounds, combined with hardware sensitive to vibration, contributes to significant human safety risks and increased personnel requirements for the movement of equipment. The apparatus is configured as a manually driven solution that mimics the operation of standard electric pallet jacks to improve operator learning time and handling. Moreover, the apparatus includes wheels configured to enhance maneuverability and improve cycle times and a tiller head configured to serve as a manual operator interface. The apparatus is beneficial in that it reduces personnel demands required for equipment movement and minimizes the risk of injury to people and hardware
Origin-Specific Coordinate Watermarking for Geolocation Data Leak Attribution
No full textA method is described for tracing geolocation data leaks by watermarking coordinates provided by a geolocation application programming interface. The method can involve generating a watermark, for example, by applying a cryptographic hash function to inputs such as the requesting origin\u27s domain and the most significant digits of a user\u27s latitude and longitude. This hash may then be embedded into the less significant decimal places of the coordinates, which can result in a small alteration to the location data before it is provided to the origin. A benefit of this technique is to facilitate the attribution of leaked data. By analyzing third-party datasets for the presence of these origin-specific watermarks, it may be possible to identify a potential source of a data leak, which can inform changes to improve user privacy
ADAPTIVE ON-DEVICE QUANTIZATION CONTROLLER FOR DYNAMIC WORKLOADS
No full textAn adaptive framework for edge devices is proposed herein that dynamically adjusts neural network precision (quantization) in real-time using live device telemetry. The framework ensures consistent accuracy, optimizes energy and thermal use, and seamlessly updates quantization, per layer, during inference with no downtime
System for Proactive Intervention and Dual-Mode Data Compliance
No full textEnterprise service platforms can present operational challenges due to reactive support models and compliance risks associated with managing sensitive data, such as personally identifiable information (PII). A dual-function architecture may integrate a proactive intervention engine and an augmented data compliance framework. The intervention engine can analyze real-time data signals from various enterprise systems using an inference layer, such as a large language model, to predict potential user friction points and deliver preventative interventions. Concurrently, the compliance framework can provide a proactive component for detecting potential PII for human-in-the-loop review and an on-demand utility for creating non-destructive, anonymized versions of data records. This integrated system may be used to shift enterprise support from a reactive to a more proactive model while providing a flexible, artificial intelligence-assisted framework for managing data privacy and compliance
System for Validating Medical Software Algorithms With Closed-Loop Adversarial Persona Generation
No full textValidating conversational artificial intelligence (AI) for regulated medical software applications may present challenges, as static test datasets and manual review may be limited in identifying emergent, conversational anomalies. A multi-agent AI system may be configured in a closed-loop for automated validation. The system can, for example, utilize an end user persona simulator agent to generate prompts for a target model and a domain /regulatory expert adjudicator agent to evaluate the target model’s responses against a configurable rubric. A meta-analysis agent can analyze anomalies to identify underlying vulnerabilities, which may then be used to programmatically synthesize new adversarial personas. This adaptive process can generate evidence to support regulatory compliance and continuous performance monitoring for medical software algorithms systems
Portable Compute Device with Top Surface Ventilation and a Retention Mechanism
No full textHigh-performance portable compute devices, such as those for mixed reality applications, smartphones, or wearable computers, may require consideration for heat generation when operated in confined spaces like a pocket, where air vents may become obstructed. To mitigate this, a device architecture can position both air intake and exhaust vents for an active cooling system on a single surface of the device\u27s housing. The device can also incorporate an integrated retention mechanism, for example, a clip, designed to engage with an edge of a pocket. This configuration may suspend the device, helping to keep the vent-bearing surface exposed to ambient air. This can facilitate more consistent airflow for thermal management, which may enable the device to sustain higher computational performance for extended periods while stowed by helping to reduce the likelihood of vent obstruction
Large Language Model-Based Data Extraction from Web Pages Using a Compact Content Representation
No full textSystems for extracting specific data, such as a transaction total from a web page, may face challenges with accuracy and scalability when using heuristic-based methods like regular expressions. A described technique may utilize a client-server system where a client application on a computing device (e.g., a smartphone, smart watch, or laptop) can generate a compact, text-based representation of a web page’s rendered content. This representation can be transmitted to a remote server where a large language model, potentially guided by an engineered prompt, can analyze the content to semantically identify and extract desired data, such as a checkout amount. The system can return this data in a structured format to the client. This approach may improve data extraction accuracy and scalability across various websites by using contextual understanding rather than more rigid, site-specific rules, potentially reducing the maintenance burden associated with some rule-based systems
Research Report and Business Case: “VCP-V1” Device for Detecting Non-Coherent Photonic Projections
No full textThis document, produced with the assistance of GPT-5.2 Thinking and Gemini 3 Reasoning, is released under the Apache License 2.0. It is a voluntary defensive publication (prior art) and therefore enters the prior art upon release under the applicable patent statutes : EPC Art. 54(2) (European Patent Convention), French IPC Art. L 611-11 (French Intellectual Property Code), 35 U.S.C. §102(a) (United States Patent Act), Chinese Patent Law Art. 22(5) (中华人民共和国专利法), and Japanese Patent Act Art. 29(1) (特許法). This report discloses an optronic + software system (VCP-V1) designed to discriminate real objects from synthetic wavefronts/photonic projections using second-order intensity correlations g(2), speckle analysis, polarimetric imaging (Stokes), topological signatures (Berry phase, OAM), spectral filtering (silica Bragg gratings, tunable filters), optional active probes (ToF/LiDAR), and edge+cloud AI fusion, including calibration/QA, traceability, and operational workflows.
Original Zenodo url: https://zenodo.org/records/1811897
Strategic Report: Pt@C60
No full textThis document, produced with the assistance of GPT-5.2 Thinking and Gemini 3 Raisonnement, is released under the Apache License 2.0. It is a voluntary defensive publication (prior art) and therefore enters the prior art upon release under the applicable patent statutes : EPC Art. 54(2) (European Patent Convention), French IPC Art. L 611-11 (French Intellectual Property Code), 35 U.S.C. §102(a) (United States Patent Act), Chinese Patent Law Art. 22(5) (中华人民共和国专利法), and Japanese Patent Act Art. 29(1) (特許法). It discloses enabling synthesis (arc/laser), purification, QA/metrology, EHS containment, REACH/CLP compliance workflows, data interoperability, and health & electronics applications for Pt@C60 as claim-ready inventions and embodiments.
Original Zenodo url: https://zenodo.org/records/1811649