17 results about "Selection criterion" patented technology

A search unit performs a search process. In the search process, a first process and a second process are iteratively performed. The first process is a process of selecting a change-target state variable from a plurality of state variables included in an evaluation function, based on the amounts of change in the value of the evaluation function caused by updating the values of the individual state variables and a first function that is a function representing a selection criterion for the change-target state variable according to the amounts of change. The second process is a process of updating the value of the selected change-target state variable. A control unit switches the function used by the search unit in the search process to a second function different from the first function.

The method includes receiving spectral data of a substrate and measurement data corresponding to the spectral data of the substrate. The method includes determining multiple feature model configurations for each of a plurality of feature models, further including determining that each of the multiple feature model configurations includes one or more feature model conditions. The method includes determining multiple feature model combinations, further including determining that each of the multiple feature model combinations includes a subset of the multiple feature model configurations. The method includes generating multiple input datasets, each of which is generated based on applying spectral data to each of the multiple feature model combinations. The method includes training multiple machine learning models, each of which is trained to produce an output using an input dataset from a plurality of input datasets and measurement data. The method further includes selecting from the trained machine learning models a trained machine learning model that satisfies one or more selection criteria.

This invention belongs to the field of speech technology, specifically relating to a method for selecting audio training datasets based on phoneme matching pronunciation tables. Addressing the problems of high redundancy, high acquisition and annotation costs, and difficulty in verifying content integrity in existing technologies, the following solution is proposed: A target phoneme set is constructed according to the needs of the target task or domain; a pronunciation table is constructed or obtained and stored in key-value pair format; the original audio-text pair dataset is obtained, and the text is preprocessed; the preprocessed text is converted into corresponding phoneme sequences in the pronunciation table, and each phoneme sequence is analyzed; based on preset selection criteria, sample pairs in the original audio-text pair dataset are selected, retaining those that meet the selection criteria to form a reduced dataset; the reduced dataset is output. This solution reduces annotation costs while ensuring data comprehensiveness, and is suitable for processing datasets for lightweight edge speech synthesis models and fine-tuning of speech models.

The present application provides a kind of ecological sensitive fire point identification and false fire point elimination method based on satellite fire point product data, comprising: obtaining satellite fire point product data and carrying out standardization integration;With each fire point longitude and latitude coordinate as center, buffer zone is constructed according to preset radius, and the proportion of each land use type in each buffer zone range is counted;According to the quantitative data of land use structure around fire point, selection criteria is set, ecological sensitive fire point is retained, and non-ecological sensitive fire point is removed;Using spatial clustering model, clustering operation is carried out on ecological sensitive fire point;The core quantitative index of each fire point cluster is counted, and the key time sequence parameter of each fire point cluster is extracted based on the observation time sequence information of each fire point in cluster;The key time sequence parameter of each fire point cluster is grouped and counted according to date-hour with fire point cluster as unit, and fire point cluster hour-level time sequence distribution data set is generated;According to the core quantitative index of each fire point cluster and fire point cluster hour-level time sequence distribution data set, multidimensional identification is carried out on fire point cluster, and false positive point cluster is removed.

Various embodiments of the present disclosure provide machine learning architectures and optimization techniques for improving predictive functionality of a computer. The techniques comprise generating a cohort-level optimization dataset with a plurality of entity-level predictive features and a plurality of feature-level predictive features for a plurality of entity data objects using a machine learning ensemble model. The techniques comprise identifying a plurality of iterative candidate outputs through a series of optimization iterations. During each optimization iteration, an iterative candidate output may be generated by applying optimization model and a constraint set combination to the cohort-level optimization dataset. The techniques comprise selecting a target output from the plurality of iterative candidate outputs based on selection criteria.

This invention discloses a method and system for simplifying indicators for assessing the resilience of aquatic ecosystems. The method generates a multimodal retrieval strategy and obtains diverse text by parsing the assessment scenario instructions using thought chain analysis. It then extracts indicator relationships from the text, analyzes the coupling relationships between indicators, constructs a knowledge graph of aquatic ecological indicators, and forms a pool of all candidate indicators. Semantic and correlation analyses are used to calculate the correlation between indicators, eliminating highly correlated redundant indicators to form a preliminary simplified indicator set. The preliminary simplified indicator set and selection criteria are output. After verification, the final simplified indicators are output. This invention solves the problems of complex indicator systems, high collinearity and redundancy, and strong subjectivity in traditional aquatic ecological assessments by integrating the logical constraints of knowledge graphs with the semantic understanding capabilities of large-scale models. It achieves scientific dimensionality reduction and dynamic optimization of assessment indicators, possessing simplicity, interpretability, and engineering application value.

The application provides a risk prediction window selection method and device based on multi-target cooperative evaluation and electronic equipment, relates to the technical field of machine learning and data modeling, and comprises the following steps: obtaining time sequence historical data of a target object, and dividing the time sequence historical data into a training subset and a test subset used for simulating future prediction according to time sequence; a candidate prediction window set containing at least two different time lengths is set, for each window length in the set, a corresponding binary classification prediction task label is defined based on the length, and a corresponding prediction model is trained by using the training subset; taking the model performance corresponding to the longest reference window in the candidate window as a reference, the optimal window length is selected from the candidate window set according to a preset selection criterion. The application ensures that the selected window length directly serves the accuracy target of the final prediction model, so that the shortest effective window is automatically found under the premise of ensuring that the prediction performance of the long window reference model is equivalent.

SYSTEM AND METHOD FOR OPTIMIZING PHYSICAL CELL IDENTITY CONFIGURATION IN WIRELESS COMMUNICATION NETWORK 5 Disclosed is a system (240) and method (400) for optimizing Physical Cell Identity (PCI) configuration in a wireless communication network. The method comprises collecting configuration data and performance metrics from one or more cells within a defined coverage area. The method further comprises processing the collected 10 configuration data and the performance metrics to identify at least one cell among the one or more cells exhibiting at least one of a PCI collision or a PCI confusion. The method further comprises determining a target PCI value for each identified cell based on a predefined selection criteria and executing a PCI update operation based on the determined target PCI value for each identified cell to optimize the PCI 15 configuration in the wireless communication network.

This invention relates to a distribution network fault detection method based on the generalized S-transform with variable factors, belonging to the field of distribution network fault detection technology. This invention obtains data such as time-frequency spectrum and transient energy through the generalized S-transform with variable factors, and constructs three line selection criteria: a comprehensive similarity coefficient of time-frequency spectrum, peak transient energy, and relative entropy of transient energy. Combining this with D-S evidence theory, it achieves multi-criteria fusion, enabling comprehensive analysis of line selection results based on multiple criteria, resulting in more accurate line selection. This invention can accurately select lines under different fault distances, fault lines, fault initial angles, fault grounding resistances, and noise environments, exhibiting good adaptability, sensitivity, and anti-interference capabilities. This invention converts the calculated transient energy characteristics into a confidence allocation function to construct evidence combinations, using different fault feature quantities to measure the confidence level and construct line selection criteria, thereby complementing and fusing them to obtain more accurate line selection results.

This invention relates to the fields of artificial intelligence and natural language processing, and discloses a method and system for fine-tuning a large-scale language model based on professional corpus of drilling acceleration scenarios. The method includes: constructing a fine-tuning dataset for drilling acceleration scenarios, the fine-tuning dataset including original professional corpus data of drilling acceleration scenarios and label information representing the acceleration effect corresponding to the original professional corpus data; obtaining a pre-trained model that meets preset base language model selection criteria as a base model; and fine-tuning the base model using the fine-tuning dataset to obtain a target large-scale language model. The large-scale language model fine-tuning method and system based on professional corpus of drilling acceleration scenarios disclosed in this application solves the problem that in the current field of drilling acceleration, large-scale language models often struggle to capture key features when processing professional corpus related to drilling acceleration due to a lack of targeted training, thus improving the application effect of large-scale language models in the field of drilling acceleration.

A computer-implemented method for monitoring and notifying a user of changes to characteristics of interest list items. The method includes recording within a memory of a user device a specification of a user's interest list items and characteristics thereof, and a specification of the user's criteria for notification of changes to the characteristics. The method includes receiving at the user device update information including updated characteristics of interest list items, comparing the update information with the recorded specifications, and determining that the user's criteria for notification of changes to characteristics of interest list items are satisfied. In response, a notification of the changes is generated for the user. The method further includes replacing the recorded characteristics with the updated characteristics of the interest list items.

Methods and systems for assessing compliance of aircraft engines are described. The method comprises selecting a test population from a set of deployed aircraft engines based on one or more selection criteria, each of the deployed aircraft engines having associated thereto design requirements and field data; selecting at least one parameter that is present in both the design requirements and the field data of the test population, the design requirements comprising design time-series profiles, the field data comprising field time-series profiles; comparing the field time-series profiles to the design time-series profiles for the at least one parameter using a similarity metric to obtain field statistics; comparing the field statistics to expected statistics from the design requirements; and generating a compliance assessment output of the test population based on a difference between the field statistics and the expected statistics.

Disclosed aspects include orchestration of machine learning and / or other rule-based indicators across elements of a webpage, including text, images, URLs, phone numbers, audio, video, to determine overall fraudulent intent. A set of dynamic web site data is received via a web browser extension installed and opted-in by the end user. The orchestration technique may be used to process the set of dynamic web site data. Accordingly, a fraud event is detected using the orchestration technique. Based on a selection criterion with respect to the fraud event, it is determined to provide an opportunity to warn the end. In response to making the determination, the opportunity can be provided to produce an alert injected into the visited website by the end user.

This invention discloses a method for optimizing drinking water taste questionnaire scoring data and selecting evaluator weights, comprising the following steps: Step S1, obtaining taste questionnaire scoring data of multiple drinking water samples from random evaluators, constructing a scoring matrix, where represents the water sample number and represents the evaluator number; the differences in taste sensitivity among the evaluators are objectively reflected in their scoring data, and are quantified and distinguished through subsequent objective weighting steps; Step S2, preprocessing the scores to obtain a standardized matrix for weight calculation. This invention, through consistency assessment and selection criteria, makes the weighting scheme selection interpretable and verifiable, avoiding the uncertainty caused by empirical selection, and the correlation coefficient between the information content weighting method and other methods is greater than 0.9, verifying the effectiveness and stability of the consistency selection mechanism of this invention.

The present disclosure relates to a computer-implemented method for assessing reliability of outputs from machine learning models, comprising: - generating, by a first machine learning model, a sequence of items, wherein each item is selected from a corresponding set of candidate items based on at least one selection criterion, wherein each candidate item of a set of candidate items is associated with probability information; and - performing, by an assessment model, a factuality check on the generated sequence of items based on the probability information of the sets of candidate items associated with the items of the sequence of items.