397 results about "Specific model" patented technology

The invention relates to the technical field of advertisement putting and intelligent decision making, in particular to a favorite advertisement putting system for predicting the advertisement click rate, which comprises a context awareness intelligent adaptation unit and a deep enhancement decision making unit. By means of deep semantic analysis and situational inference engine processing, interest keywords are extracted by constructing a specific model, user situational portraits are constructed in combination with multi-source data, a preliminary advertisement set is screened by matching with an advertisement material library, a multi-agent architecture is constructed by a deep reinforcement decision unit, a master agent performs overall planning, slave agents are responsible for different advertisement types, and a user can perform multi-agent interaction. A multi-dimensional reward function system is designed, each agent collects feedback data such as operation of a user on an advertisement page, a reward value is calculated according to the feedback data, a strategy network is updated, a main agent integrates information to optimize an overall advertisement pushing strategy, accurate pushing of advertisements is achieved, and the advertisement click rate and the putting effect are effectively improved.

The invention belongs to the technical field of intelligent agriculture, and discloses a canopy-atmospheric environment information-based soil water content prediction method and system, and the method comprises the steps: setting a canopy and atmospheric environment monitoring unit in a field, and synchronously collecting the canopy temperature, wind speed, relative humidity, and six-dimensional environment parameters of atmospheric temperature, wind speed and relative humidity; the crown temperature difference is calculated after preprocessing; the method comprises the following steps: constructing a training set by utilizing measured data of soil relative water content, constructing a soil water content prediction model by adopting a random forest algorithm, screening and confirming an optimal prediction model through grid search and cross validation, and screening key features by combining an SHAP value; the water demand is automatically calculated according to a pre-established crop whole-growth-period drought stress threshold table and a dynamic irrigation decision formula, and meanwhile, the adaptation of a general model to a regional specific model is realized by constructing an environmental parameter-actually measured water content database and transfer learning, so that the soil water content state is accurately predicted, and the soil quality is improved. And a scientific basis is provided for precise irrigation.

The systems and methods disclosed herein relate to querying data using artificial intelligence models. A generalized model receives an output generation request and partitions it into segments mapped to specific domains, where each domain indicates associated databases and guidelines. The segments are routed to domain-specific models trained on domain-specific data, which generate query fragments by comparing performance metrics and system resource usage metrics. The query fragments are aggregated into an overall query that satisfies guidelines across domains. The systems and methods can include a feedback loop to adjust the domain-specific models using user interactions and performance metrics to dynamically adapt to a skill level or experience of the user.

In various examples, each user of a hosted application may be modeled with a corresponding user-specific resource consumption model that predicts a measure of that user's anticipated (e.g., processing, memory or storage, and / or networking) resource utilization at some future time based on an input representation of one or more features of the user's current session, application (e.g., game) setup, skill or experience level, social media activity, some time series representing feature(s) for a series of interactions, the time of day, and / or other characteristics. As such, user-specific models may be used to predict resource demands at a future time and schedule resource allocations accordingly. The present techniques may be used to manage and reallocate resources for applications such as game streaming applications, remote desktop applications, simulation applications (e.g., an autonomous or semi-autonomous vehicle simulation), virtual reality (VR) and / or augmented reality (AR) streaming applications, and / or other application types.

Examples for performing wireless sensing tasks based on foundation model are described. In one example, a described method comprises: obtaining channel information (CI) data generated based on at least one wireless channel; generating a training dataset based on the CI data, wherein the training dataset comprises: a plurality of CI pairs, original CI data and a mask; training a foundation model using the training dataset based on an aggregate of a contrastive loss function and a reconstruction loss function; training a plurality of task-specific models; and performing a plurality of wireless sensing tasks based on the foundation model and the plurality of task-specific models. Each of the plurality of task-specific models is used to perform a corresponding one of the plurality of wireless sensing tasks together with the foundation model.

An online medical question and answer dynamic retrieval enhancement generation method relates to the technical field of large language models, and by introducing semantic entropy and a semantic entropy-based dynamic threshold setting mechanism, a system determines whether to trigger retrieval according to a semantic entropy value of generated content. According to the method, excessive retrieval operations brought by a traditional static threshold method are avoided, the retrieval pertinence is improved, the professionalism and accuracy of answers are enhanced, and the possibility that the model generates'illusion 'or irrelevant answers is reduced. Medical entity extraction is carried out in combination with a special model (Med-BERT) in the medical field, and the medical elements in the text can be accurately recognized and generated in a classified mode. Based on the medical elements, a more accurate judgment basis can be provided for triggering of a system retrieval mechanism.

Systems and methods for domain-specific model-generated content item generation, evaluation, and selection can include generating a plurality of candidate model-generated content items that can then be evaluated based on one or more signals, which can then be leveraged for candidate model-generated content item selection. The plurality of candidate model-generated content items can be generated with a generative model that was tuned for domain-specific content item generation. The selected model-generated content item can be processed to generate an outline that may then be provided to a user for user interaction to generate an augmented outline. The augmented outline may then be processed to generate an updated model-generated content item.

A system and method for multiple fraud type detection includes anti-injection attack system that has a layered architectural approach that uses a includes combination of different specific models to detect the attacks in combination with image processing techniques, device signals and liveness checks to detect the variety of different types of fraud attacks or repeat fraud attacks. The anti-injection attack system applies the analysis of the tools used to create deepfake, face morph and face swap attacks to define the elements of its layered architecture that can these various types of attacks.

Techniques are described for scenario data generation. An example method can include processing location-specific sensor information of a traffic event at a location on a road network. The method can include determining a location-specific element based on the location-specific sensor information, the location-specific element associated with a flow of traffic at a time point. The method can further include generating a location-specific model of the traffic event based on the flow of traffic and the time point. The method can further include determining, by the computing system, control instructions for a location-specific driving maneuver for an autonomous vehicle to mitigate the traffic event based on the location-specific model. The method can further include causing the control instructions to be transmitted to an actuator of the autonomous vehicle to perform the location-specific driving maneuver thereby controlling the autonomous vehicle at the location based on the model.

The invention discloses an equipment state monitoring method based on multi-source information fusion, and the method comprises the steps: enabling a real-time sensor to collect the operation parameter data, vibration parameter data and environment parameter data of equipment, collecting the historical state data of the equipment with a recording label, and carrying out the preprocessing of the data; time domain features, frequency domain features and working condition features of the data are extracted according to a layering mode, and dynamic weight coefficients of all the features are set; calculating a reference threshold value by adopting a specific model; evaluating the health condition of the current equipment by adopting a depth measurement method, calculating an equipment health factor, and calculating a trend compensation item; and obtaining an equipment state monitoring dynamic threshold based on the reference threshold, the health correction item and the trend compensation item. The invention further discloses an equipment state monitoring device based on multi-source information fusion, corresponding equipment and a storage medium. According to the equipment state monitoring method based on multi-source information fusion provided by the embodiment of the invention, the accuracy, real-time performance and reliability of equipment state monitoring can be effectively improved.

The invention discloses a quantum adversarial attack method based on momentum, which refers to an object momentum mechanism, and designs a quantum adversarial attack method to efficiently generate adversarial samples. In order to solve the problem that an existing quantum countermeasure attack method excessively fits a specific model to cause a poor local optimal solution, the invention provides a novel quantum countermeasure attack method based on momentum (Q-MIFGSM). According to the method, a loss function of a quantum neural network classifier and momentum accumulation gradient information of a sample are defined, a momentum iteration updating mode is adopted, momentum accumulation is carried out on a loss function gradient value, the optimization direction is dynamically adjusted in combination with current and historical gradients, and a local optimal solution is effectively jumped out; the loss function value of the quantum classifier is increased to cause classification errors of the classifier, so that attack resistance is realized, and an adversarial sample is efficiently generated. Simulation experiment results show that the Q-MIFGSM algorithm can effectively execute adversarial attacks on the quantum neural network classifier and generate adversarial samples close to original data. A new strategy is provided for quantum attack resistance, and the method has potential application value in the aspects of quantum neural network security evaluation and defense strategies.

The invention relates to the field of computer systems, in particular to an operation stability testing method and system oriented to a heterogeneous computing system, and the method comprises the steps: constructing a system overall operation logic model comprising a resource topological structure, a task scheduling rule and a communication link matrix; identifying a key communication path according to the communication link matrix, analyzing a resource competition hotspot in combination with a task scheduling rule, and establishing an interference model library; calling a specific model instance in the interference model library according to a test target, configuring injection parameters, and dynamically applying interference in a system operation process; key performance indexes in the system operation process are collected in real time; based on the real-time monitoring data, a multi-index weighted scoring algorithm is adopted to calculate an overall stability coefficient, a robustness grade and a self-healing capability index of the system, and a stability evaluation result is generated; and generating a visual report according to the topological structure of the operation logic model, and supporting comparison and analysis of historical versions. Unified centralized management is realized, and test efficiency and evaluation accuracy are improved.

A hierarchical language model system is disclosed for generating optimal responses to user queries. The system comprises multiple lower-level large language models (LLLMs) that are domain-specific models trained on particular datasets to generate proposed responses. At least one LLLM is a finetuned model for a specialized domain. The system also includes a higher-level language model (HLLM) that analyzes the proposed responses from the LLLMs based on criteria such as relevance, accuracy, redundancy, and domain relevance. The HLLM uses algorithms to select the best response or synthesize an optimal response by combining elements of multiple proposals. This allows customizing the response by identifying and integrating the most salient information from different LLLMs. An adaptive learning component allows refining the HLLM's selection strategy over time based on user feedback. The hierarchical ensemble approach balances the breadth of general models with the depth of domain-specific understanding. By orchestrating specialized lower-level models with an overarching higher-level model, the system delivers accurate, nuanced, and consistent responses across a wide spectrum of user queries.

The invention relates to the field of large language models, and provides a safety guarantee method for text generation of a large language model. The existing watermarking technology can detect whether a text comes from a specific model or not, but cannot trace the identity of a user, so that responsibility can not be traced when the text is spread or tampered for the second time. According to the main scheme, a statistical signal is embedded in a generated text based on a relaxed red-green dictionary method, and a model tends to select a green list token by adjusting the token sampling probability when the language model is generated; the user information is a mobile phone number or an identity card number, Base4 coding is carried out, and offset information is obtained; in the basic watermark text, starting from the first green token, generating a synonym list based on the context every two continuous token positions, and selecting corresponding synonyms for replacement according to the offset so as to implicitly embed user information; the method comprises the following steps: detecting a basic watermark, recovering a text by using a Transform-based sequence denoising auto-encoder, and extracting user information according to a replacement position and an offset.

The systems and methods disclosed herein relate to querying data using artificial intelligence models. A generalized model receives an output generation request and partitions it into segments mapped to specific domains, where each domain indicates associated databases and guidelines. The segments are routed to domain-specific models trained on domain-specific data, which generate query fragments by comparing performance metrics and system resource usage metrics. The query fragments are aggregated into an overall query that satisfies guidelines across domains. The systems and methods can include a feedback loop to adjust the domain-specific models using user interactions and performance metrics to dynamically adapt to a skill level or experience of the user.

In various examples, self-supervised learning may be used to pre-train an encoder network of a masked prediction model to reconstruct masked regions of an input representation of 3D detections such as LiDAR point cloud(s). Spatial and / or temporal masking may be applied to a projected representation of 3D detections (e.g., a two-dimensional (2D) projection image), and the masked prediction model (e.g., a masked auto-encoder or joint-embedding predictive architecture) may be used to reconstruct a representation of the masked regions (e.g., reflection characteristic(s) stored in corresponding pixels or cells of the projected representation, a latent representation of the reflection characteristic(s)) during iterations of self-supervised learning. As such, the pre-trained encoder network of the masked prediction model may be used as a foundation model and fine-tuned with a task-specific output head or its pre-trained weights may be used to initialize a task-specific model.

The invention relates to the technical field of soil detection, and discloses a soil property quantitative analysis method of a spectrum-driven and greedy optimization integrated model, which comprises the following steps: preprocessing data; defining a basic model; selecting a greedy optimization model; integrating learning models; evaluating model performance; according to the method, the optimal model combination is selected through the greedy optimization method, the advantages of multiple models are integrated, and the accuracy of soil property prediction is improved. The greedy algorithm is simple and efficient, model selection and integration can be completed in a short time, and the method is suitable for processing of large-scale data sets. The method does not depend on a specific model or data set, and can be expanded and applied to prediction tasks in other fields.

An iterative method for monitoring a computing device characterized by metric data to be monitored, including, for each iteration, of collecting metric data over a predetermined interval of time, detecting a seasonality pattern of said metric data over said predetermined interval of time, determining an interval-specific model representing the detected seasonality pattern, calculating modelled data using said determined model and the collected metric data, comparing the calculated modelled data with the collected metric data to calculate a score characterizing the difference between the calculated modelled data and the collected metric data, calculating an anomaly likelihood for each data of the collected metric data using the calculated score, detecting an anomaly on a data when probability that the value of said data is an anomaly is greater than a predetermined threshold.

The invention discloses an intelligent automatic generation system and method for coral species semantic samples, relates to the technical field of computer vision and marine ecological monitoring, and aims to efficiently generate high-precision semantic annotation samples from underwater coral images. The system adopts a hierarchical feature extraction architecture, a common model and a field-specific model CoralSCOP are collaboratively optimized, an SAM series provides a high-precision geometric segmentation region through zero sample learning, and the CoralSCOP realizes multi-granularity label distribution through parallel semantic branches. The method innovatively introduces a two-stage mask optimization mechanism: in the initial stage, a semantic segmentation model is utilized to generate a 50 + category semantic mask; in the post-processing stage, boundary details are extracted through SAM / CoralSCOP, semantic and geometric features are matched through a region-level fusion strategy, and the segmentation precision and efficiency are improved. The system provides a high-quality marking benchmark for coral reef monitoring, the expert dependence is remarkably reduced through the zero-sample and multi-granularity characteristics, and the system is suitable for marine ecological research and AI model training.

The embodiment of the invention provides a multi-modal general-purpose model collaborative reasoning method based on a dynamic routing mechanism, and the method comprises the steps: receiving a task instruction and multi-modal data corresponding to the task instruction, determining a task type corresponding to the task instruction, and extracting the data features of the multi-modal data, fusing the data features through a task general model to generate general fusion features, matching a task special model of the current task type, a modal processing flow and a fusion feature weight according to a preset processing mapping relation, and transmitting the data features and the general fusion features to the task special model through a routing node, and enabling the task special model to process the data features and the general fusion features according to the modal processing flow and the fusion feature weight to obtain special fusion features, and processing the special fusion features through a task execution component to obtain a task result corresponding to the task instruction. According to the method, the defects of lack of model flexibility and the like during multi-modal data processing can be effectively solved, and the convenience degree of semantic fusion is remarkably improved.

A method, system, device, and medium for generating 3D model of industrial scene are disclosed. The method comprises: acquiring point cloud data for an industrial scene; determining a domain of the industrial scene; acquiring a trained semantic segmentation model associated with the domain from a model library, wherein the model library comprises multiple trained semantic segmentation models associated with respective domains; performing semantic segmentation on the point cloud data to obtain multiple segmented objects, based on the trained semantic segmentation model associated with the domain; meshing the multiple segmented objects; and generating a 3D model of the industrial scene based on the meshed multiple segmented objects. A flexible domain-specific model library is provided to meet the needs from different industrial scenes, thus improving modeling efficiency.

The invention discloses a database query plan optimization method and system, and belongs to the technical field of databases and artificial intelligence. AI model training: training the model by using the preprocessed historical query data according to a specific model architecture, a training mode and resource and time arrangement, and dynamically adjusting the model according to the change of the database; query analysis: evaluating the complexity of the query and the possibly involved data volume, and allocating a complexity level and a data access volume range for the query; query plan generation: generating and evaluating a plurality of candidate query plans based on prediction and query analysis results of the AI model, and selecting an optimal query plan as a final execution scheme; and performing query execution and monitoring. According to the method, the query execution time can be greatly shortened, and the database response speed is increased; system resources are utilized more effectively; dynamic changes of the database can be adapted, and the query plan is automatically updated according to structure and data distribution adjustment.

The present disclosure relates to a system for dynamic task processing using a core language model and task-specific models. The system comprises a memory, a persistent storage, and a core model that resides in memory. A plurality of task-specific models are stored in persistent storage and configured to be dynamically loaded into and unloaded from memory in real-time based on incoming task requests. Upon receiving a task request, a corresponding task-specific model is loaded into memory and integrated as an additional layer of the core model to fine-tune its output for the specified task.

A system includes a hardware processor configured to execute a machine learning (ML) model training pipeline to train an ML model using data relevant to a world of a digital persona to provide a dialogue model, generate, using the dialogue model, first conversational outputs, train the dialogue model, based on the first conversational outputs, to avoid hallucinations and / or undesirable expressions to provide a guardrailed dialogue model, generate, using the guardrailed dialogue model, second conversational outputs, train the guardrailed dialogue model, based on the second conversational outputs and persona data identifying interaction characteristics of the digital persona to provide a persona-specific model, generate, using the persona-specific model, a response to a scripted question, determine a quality score for the response, and further train the persona-specific model or validate the persona-specific model for human interaction, depending upon whether the quality score fails to satisfy or satisfies a quality criterion.

One or more systems, devices, computer program products and / or computer-implemented methods of use provided herein relate to assessing certainty of artificial intelligence models used for detection or segmentation of pathologies. Accordingly, a system can comprise a memory that can store computer executable components. The system can further comprise a processor that can execute at least one of the computer executable components. The computer executable components can comprise a synthetic data generation component that generates biologically-inspired synthetic data that approximates a task-specific data manifold of a medical image from a radiomic features perspective; an artificial intelligence component that uses an artificial intelligence model to learn relevant representations of the synthetic data for an at least one image task; and a training component that utilizes the relevant representations and the artificial intelligence model to generate a task-specific model for the at least one image analysis task.

The invention discloses a cross-domain structural damage identification method based on digital twinning and time-space feature fusion, belongs to the technical field of structural health monitoring, and provides a cross-domain transfer learning damage identification technical scheme fusing a digital twinning model and structural multi-sensor data time-space features. A measurement point connection relation of a digital twinborn model and an actual physical structure is abstracted into a graph structure, and a spatial-temporal feature learning model for simultaneously processing multi-sensor time domain data and spatial feature information is designed based on classic time domain convolution and graph convolution network. A specific model architecture is provided for deep fusion of a structural digital twin model and spatial-temporal features of multi-sensor data of an actual structure to realize structural damage detection in a cross-data-domain scene, so that the problems of insufficient actual damage data, insufficient fusion of structural multi-dimensional feature data and the like are solved.

The invention discloses a dynamic modeling tracking control method for a self-adaptive fiber placement compaction mechanism, and belongs to the technical field of specific model calculation systems, and the method comprises the following steps: S1, building a kinematic model of the self-adaptive fiber placement compaction mechanism; s2, establishing a workpiece surface model, and solving tangent point coordinates and position coordinates of the driving wheel, the driven wheel and the ground; s3, according to the force balance equation and the moment balance equation of each rod, the relation between each bearing reaction and the external force is solved; s4, establishing a dynamic model of the driving device; and S5, error analysis and optimization are carried out, and it is ensured that the main pressure is within the fluctuation range. According to the device, it can be ensured that the rollers output constant compaction force when the abrupt change molded surface structure is laid, and the forming defect caused by uneven compaction force is avoided.

The invention discloses an industry exclusive customization model generation method based on a large model, and the method comprises the following key steps: firstly, collecting structured and semi-structured multi-modal data containing key entities, concepts and relationships in a target industry; then, constructing a knowledge graph through a deep learning method; in a model construction stage, carefully selecting a pre-trained large language model, and carrying out personalized adjustment on feature engineering and a model architecture according to industry characteristics; on the basis of the constructed knowledge graph, a retrieval enhancement technology based on the knowledge graph is adopted, the answer content of the large language model is optimized, and target industry domain knowledge is injected; and finally, by using a reinforcement learning algorithm, optimizing model output by training a reward model. According to the method, specialized customization of the industry exclusive model is realized, the question and answer ability of the large language model in the target industry field is remarkably improved, and high accuracy and high correlation of the model when the model answers related questions of the target industry are ensured.

The invention discloses a remote sensing image small target feature enhancement refined classification method based on dynamic mixed knowledge distillation, and the method comprises the steps: inputting a low-resolution remote sensing image of a to-be-classified small target into a pre-trained classification model, and obtaining the specific model information of the target; the classification model comprises a trunk low-resolution image student network branch, and in the training stage of the classification model, an auxiliary high-resolution image teacher network branch consistent with the student network branch in structure and a dynamic mixed knowledge distillation module are introduced, the dynamic mixed knowledge distillation module is deployed between a trunk low-resolution image student network branch and an auxiliary high-resolution image teacher network branch, and high-resolution image intermediate features and output features learned by the teacher network branch are used as priori knowledge to be distilled to the student network branch through the dynamic mixed knowledge distillation module; and training of the student network branches is guided, so that the feature representation capability of the student network branches is enhanced, and a trained classification model is obtained.

Embodiments of the invention provides a sticker generation method and device, an electronic device, a computer readable storage medium, a computer program product and a computer program. The method comprises: obtaining a material image of a target component on an avatar, the target component being in motion in a sticker comprising the avatar; determining a global position of the target component according to the material image; determining a periodic motion amplitude of the target component in the sticker; generating the sticker according to the material image, the global position, and the periodic motion amplitude. Therefore, a dynamic effect of components in the sticker is achieved, a user does not need to master a specific software and use complex skills to perform vertex layout and movement, and a specific model file does not need to be involved. Thus, the sticker making difficulty is reduced, the sticker making efficiency is improved, and better user experience is achieved.