84 results about "Validator" patented technology

Various embodiments of the present disclosure provide methods, apparatus, systems, computing devices, computing entities, and / or the like for identifying one or more evidence text portions comprising one or more bases relied on by a generative machine learning model for assigning a plurality of model-assigned categorical identifiers to a plurality of text segment data objects associated with a document data object, and verifying the one or more evidence text portions with a verifier machine learning model to generate one or more classifications of the document data object and provide the one or more verified evidence text portions along with the one or more classifications.

A diversified validation test suite application (DVTSA) for a generative AI powered tool or large language model (LLM) tool includes at least first, second, and third control logics. The first control logic receives a seed test input or user input, analyzes the seed test input or user input, and extracts key elements for variations. The second control logic performs a coverage measurement of outputs of the first control logic. The third control logic causes a human validator to evaluate outputs of the second control logic relative to predefined coverage metrics and selectively and continuously iterate to cause outputs of the second control logic to increase LLM tool input robustness and output robustness from a first level to a second level greater than the first level, in both production and pre-production LLM tool processes.

A system, method, and computer-readable medium enforce compliance, policy governance, and jurisdiction-aware control of digital artifacts generated by AI models. A deterministic digest of each artifact is stored in a registry, and a compliance token is generated encoding policy predicates, consent scope, jurisdictional lineage, monetization parameters, and a zero-knowledge proof component including a proof statement ID, cryptographic commitment, and verification key. A gateway verifier receives the artifact and token, verifies the signature, performs zero-knowledge proof validation without revealing regulated data, evaluates encoded policies, and renders a permit-or-deny decision. Verification results and compliance decisions are recorded via a dual-ledger atomic commit engine that updates both compliance and financial ledgers. A settlement engine calculates royalty values based on validated performance metrics and appends the results to the commit entry. Upon jurisdictional transitions, a transition engine detects boundary crossings and issues a forked compliance token with updated predicates. The system ensures verifiable, policy-aware, and jurisdictionally adaptive governance.

The invention relates to a large language model unified verification method, which comprises the following steps of synthesizing an assembly line based on evolutionary verification data, and outputting to obtain synthesized training data; in combination with an improved GRPO algorithm, the current large language model is trained by using the synthetic training data, and a verifier model is constructed; and evaluating the output of the current large language model by using the verifier model to obtain an evaluation result. Compared with the prior art, the method has the advantages that comprehensive and universal verification capability is injected into a compact and efficient model, and objective fact verification and subjective quality judgment tasks can be processed in a single model at the same time by constructing a unified, universal and customizable verification model; fine-grained verification instructions and output format requirements which are issued by a user in a natural language and are in any form can be understood and followed, and detailed and rational explanation can be provided for verification results.

Reinstating access to a system of an admin whose certificate is invalid or expired is disclosed using a secret that is remapped. When the admin's certificate is expired, the admin may send a request for reinstatement to tenant admins. The voting operation is based on shares of a secret that have been distributed to validators. When the shares are successfully reconstructed to obtain the secret, the voting operation is affirmed or allowed to proceed. The secret is remapped after a period of time. Remapping the secret includes creating a new secret and distributing new shares of the new secret to a different set of validators.

The invention discloses a website data analysis method and device based on a large model technology, and belongs to the field of fusion of web crawlers and artificial intelligence. According to the method, a local knowledge base drives a large model to generate a precise XPath rule, and the precise XPath rule is packaged into a decoupling plug-in and integrated into an existing crawler framework; a dynamic feedback optimization mechanism is innovatively designed, rule validity is monitored in real time by using a composite verifier, and XPath reconstruction and knowledge base updating are automatically triggered. According to the method, the high concurrency performance of a traditional crawler is reserved, the rule generation efficiency and the system robustness are remarkably improved, the manual maintenance cost is reduced, and the method is suitable for a website data collection scene with a complex structure or frequent updating.

This invention provides a method for generating complex functions on a low-code platform based on large model code interpretation and generation. By capturing and standardizing user input, a structured intent graph containing business entities, operational relationships, parameter constraints, and control flow topology is constructed as a unified semantic anchor. Combined with a pre-defined domain rule base, a semantic alignment validator is designed to achieve multi-dimensional semantic comparison and conflict detection between code and requirements. Code snippets are generated in parallel by multiple models and verified in real time, and semantic conflicts are resolved through local regeneration. By fusing the results of multiple models to construct a joint abstract syntax tree, lightweight symbolic execution is used to simulate the correctness of cross-model code, locate and correct core logical conflicts. This invention significantly improves the consistency, accuracy, and automatic correction capabilities of low-code platform code generation.

The invention relates to a semantic abstract generation and security verification method and system for multi-modal data, and belongs to the technical field of data security and information process.The method comprises the steps that feature extraction is carried out, specifically, a client processes a multi-modal plaintext data stream, and key features of the multi-modal plaintext data stream are extracted; vector mapping: the client performs vector space transformation on the extracted features to generate a unique corresponding data abstract; local persistence: the system stores the generated data abstract back to a user local database for subsequent use; abstract collection, wherein the detector obtains generated data abstract information from the data stream; and integrity verification: carrying out consistency comparison on the data stream with the data abstract by the detection and verification device, and judging whether the data stream is complete and reliable or not. The data abstract generation system comprises a client module, a server module and a middleware module. According to the method, efficient and reliable data stream abstract generation and credible management are realized on the premise of fully guaranteeing user data privacy.

A cache validation system receives Content Delivery Network (CDN) caching rules and a request log, such as a log of HTTP requests received by a CDN. The cache validation system generates test requests from the CDN caching rules and the request log. In a testing environment, the cache validation system determines one or more cache validation indicators, such as, but not limited to, a hit-to-miss value, from the test requests and the CDN caching rules before application of a proposed change to a CDN caching rule. The cache validation system applies the proposed change to the CDN caching rules in the testing environment. The cache validation system determines one or more updated cache validation indicators from the test requests for the CDN caching rules after application of the proposed change. If the initial and updated cache validation indicator(s) satisfy a threshold, then the cache validation system can provide a notification.

An identity verification-authentication system and method are used for verifying and authenticating at least a first user. The system includes a first prover for acquiring a first unique feature provided by the first user; and a verifier in communication with the first prover for generating a first proof request. The verifier sends a first variant data included in the first proof request to the first prover. The first prover generates a first to-be-compared answer data according to the first unique feature and the first variant data. The first prover or the verifier then generates a determination result according to a comparison of the first to-be-compared answer data and a first answer data included in the first proof request, and determines whether the identity verification and authentication of the first user is successfully according to the determination result.

Embodiments of this disclosure provide, but are not limited to, solutions for protecting, sharing, verifying, and / or generating data resources such as documents. In exemplary embodiments, a data resource is represented, expressed, or defined as a plurality of hashed constituent segments. Using the segment hashes as leaves, a Merkle tree is generated representing the complete original version of the data resource. An edited version of the data resource can be shared, omitting one or more segments that the data administrator wishes to keep secret or does not wish to share in its original form. To verify the authenticity and / or completeness of the edited data resource, a validator is provided with sufficient data (e.g., the hash of each of the non-shared segments) to verify the presence of non-shared segments in the tree representing the complete original version, and an authorized signature that signed the root. In some embodiments, the data resource is a code library, and the segments are parts of computer code that perform specific subtasks. Shared code segments, when executed on a processor, may be selected to combine to perform larger tasks.

The disclosed techniques pertain to training large language models (“LLMs”) using table data. Specifically, the disclosed techniques pertain to training LLMs for table-related tasks using two models, each model reserved for different functions. A first model is reserved for generator functions and a second model is reserved for validator functions. The first model receives table data and generates training data. The training data is fed to the second model, which identifies instances of training data meeting or exceeding at least one validity threshold. Instances of training data meeting or exceeding the at least one validity threshold are output as validated training data. The validated training data is used to iteratively fine-tune the two models by increasing or decreasing one or more numeric weight parameters in each of the models that control how the models process input data and produce outputs.

A system for computing confidence scores for large language model (LLM) based retrieval augmentation (RAG) tools includes a host device having a controller. The controller executing programmatic control logic including an algorithm for computing confidence scores for LLM based RAG tools (CLR application). The CLR application receives an input to the LLM from a user and engages an ensemble retriever. The ensemble retriever determines a similarity between the input and predetermined data in the one or more databases. The ensemble retriever generates an output and an output confidence score. The CLR application then causes a human validator to prioritize and review the output in accordance with the output confidence score, where the output is a command to one or more systems of the host device. The system progressively reduces computational resource utilization, progressively increases computational efficiency, and progressively reduces reliance on the human validator over time.

Disclosed are an apparatus and method for electing a committee node in a blockchain. The apparatus for electing a committee node in a blockchain is configured to transmit a staking transaction including validator information to a blockchain system to register a validator node, generate a draw pool for electing a committee node based on the validator information, change entries of the draw pool using a predefined shuffling algorithm, select a preset number of top entries as votes of a committee member from among the entries of the shuffled draw pool, and elect, as the committee node, the validator node that has acquired votes of a committee.

This application provides a method and apparatus for training a natural language to SQL model based on reinforcement learning. The method includes: calling an external tool based on natural language instructions and the structural information of a target database to generate a response containing the tool call trajectory and SQL execution results; the external tool includes at least one of a search tool, a calculation tool, and an SQL execution tool; based on the response containing the tool call trajectory and SQL execution results, performing inference verification on the SQL execution results to obtain verification results, the inference verification including comparing the difference between the result table of the SQL generated by the model and the real table through an SQL validator; and adjusting the generation strategy of the large language model according to the verification results to generate executable SQL statements. An end-to-end natural language to SQL training framework is constructed through reinforcement learning algorithms, combining tool fusion inference and interleaved thinking mechanisms to dynamically optimize the model's decision-making capabilities throughout the entire process of data table retrieval, logical calculation, and SQL generation.

This invention discloses a knowledge-driven automated formal verification method and system for smart contracts, belonging to the fields of blockchain and artificial intelligence technology. First, it constructs a formal verification language knowledge graph and an expert experience knowledge graph to extract grammatical rules, vulnerability patterns, and verification logic. Then, it performs project preprocessing and static analysis on the target smart contract to extract an abstract syntax tree, call relationships, and system model. Combining the knowledge graph and large language model, it automatically generates formal verification constraints and executes the verification task through a formal verifier. For errors and counterexamples generated during the verification process, the system uses a feedback mechanism to repair rules and optimize parameters. Finally, it generates a formal verification report containing verification results, risk analysis, and remediation suggestions. This invention achieves full-process automation from verification requirement generation to verification execution and optimization, effectively improving the efficiency and accuracy of smart contract security verification.

The invention provides an incentive and secure federal learning method and system based on a block chain, and the method comprises the steps: a plurality of participants obtain training tasks from the block chain, submit corresponding bidding prices to a smart contract, the smart contract determines a winner set in current training according to the bidding prices, and randomly divides the winner set into a training set and a verification set, obtaining a corresponding trainer and a verifier; the method comprises the following steps: downloading a currently trained local model from a block chain based on winner, adding artificial Gaussian noise, calculating a noise gradient of a target function of the local model by each winner, transmitting the noise gradient to each verifier for model verification, obtaining a list of the selected local model, and performing global aggregation; all verifiers are divided into small fragments, after a verification task of a local model of the trainer is received, each verifier in the small fragments evaluates the quality of the model, and the verifiers exchange votes mutually. The fragmentation consensus algorithm based on circulation can ensure the same security as a non-fragmentation consensus protocol.

Computer-implemented systems and method train a generator and a discriminator, through machine learning, where the generator and discriminator are trained in an adversarial relationship using a simulated, multi-player game. The model parameters for the generator and the discriminator can be updated non-simultaneously. Also, the simulated, multi-player game may comprise a two-person, zero-sum game.

Disclosed is a fail-closed conformance control layer for machine-assisted decision artifacts. The system generates replay-verifiable receipts that commit to a canonical artifact digest, evidence and snapshot identifiers, assumptions, toolchain identifiers, and a policy pack identifier and policy version digest. A deterministic validator evaluates a receipt under the versioned policy pack to output PASS, FAIL, or HOLD with reason codes from a reason-code registry, including time-stamped modeled-variability codes. Downstream side effects are gated in a control path and permitted only on PASS; otherwise they are withheld or routed to remediation or human review. The system may emit portable conformance packs and revalidate receipts over time to detect drift and manage a certification lifecycle. Conformance suites with deterministic test vectors and expected validator outputs may certify validator implementations by issuing signed conformance certificates and publishing a certificate or certificate digest to a verifiable certificate registry for independent verification and procurement acceptance testing.

A system for incorporating user feedback in a large language model (LLM) based retrieval augmentation (RAG) tools includes an application for incorporating user feedback (UFA). The UFA receives an input to the LLM from a host device user, generates an LLM output as a response to the input, causes a system user to verify the LLM output and provide user feedback via a human-machine interface (HMI) of the host device, and in response to the user feedback, engages a performance tuning optimizer that modifies the LLM output by: revising domain knowledge, revising system prompts, and performing constrained regeneration of the LLM output. The performance tuning optimizer progressively reduces computational resource utilization, progressively increases computational efficiency, and progressively reduces reliance on human validators and user feedback over time, and wherein the LLM output is a command to one or more systems of the host device.

The invention provides a cross-chain atomic interaction method of a multi-chain cooperation platform for power data transaction, and relates to the technical field of power data transaction. A payment signature is bound with a data decryption condition through an adapter signature mechanism, so that payment is confirmed on a public chain and can be decrypted into an inseparable event with data on an alliance chain, cross-chain atomic delivery can be completed without a hash time lock, interaction turns on the chain are reduced, and the transaction success rate is remarkably improved. And secondly, a large amount of calculation-intensive data analysis operation is migrated into the TEE to be executed, and the alliance chain and the public chain only need to store lightweight data such as result hash and signature, so that the calculation load on the chains is greatly reduced, and the overall throughput capacity and response efficiency are improved. Finally, the method does not depend on block header verification, multi-signature bridging or external verifiers, can flexibly adapt to block chains with different consensus mechanisms and different performance characteristics, and realizes good portability and deployment convenience.

This invention relates to a web automation testing method based on a two-layer AI architecture of perception and orchestration, comprising: an orchestration layer for receiving test tasks, generating and sending operation instructions through LLM dynamic programming, and invoking the perception layer, and integrating a state monitoring engine and a schema validator; a perception layer, communicatively connected to the orchestration module, for performing multimodal analysis using a large visual language model (VLM) based on the received image data to be analyzed and structured analysis instructions, and returning predefined formatted structured data; wherein, the orchestration module is further configured to generate a new sequence of operation instructions based on the structured data returned by the perception module, through a schema validator, to drive the browser execution agent of the execution layer to complete the test task. This invention can achieve a high degree of intelligence while ensuring the absolute reliability of the test execution process. The architecture further introduces a state transition engine and an output verification mechanism to enhance the robustness and generalization ability of the system.

A computer-implemented method, the method comprising, at a validator node in a blockchain network: acquiring a first transaction, the first transaction including executable code; receiving a second transaction, the second transaction containing information including at least a commit instance of the r portion and the s portion of a first ECDSA signature, and the information further including a random number; and executing the code of the first transaction. The code is configured to verify H. PoW If (f(r, d)) satisfies the predetermined conditions defined in the code, return TRUE if the predetermined conditions defined in the code are met, where r is the submission instance of the r part, d is the random number, and H PoW q is a hash function, and f is a function that combines q and d.

Various aspects of the subject technology relate to systems, methods, and machine-readable media for cross-chain communication in a blockchain platform. Various aspects may include accepting, at a first blockchain, a first transaction including a message and a message payload. Aspects may also include validating, at the first blockchain, the message by signing the message using signature keys of one or more validators in a first set of validators of the first blockchain. Aspects may also include generating an aggregate signature based on the signature keys of the one or more validators in a first set of validators. Aspects may also include submitting a second transaction on to a second blockchain, the second transaction including the message and the aggregate signature. Aspects may include validating, at the second blockchain, the second transaction based on a shared registry.

The present disclosure relates to a system and a method for processing distributed data files. The processor executes instructions to receive a set of instructions from a primary device, wherein the set of instructions comprises verification rules, validators, primary transformers and structure query transformers; generate processed data files by processing the distributed data files. The distributed data files are processed by performing at least one of: executing one of the verification rules, the validators and the primary transformers on the distributed data files; and transforming the distributed data files by executing the structure query transformers. The execution of the structured query transformers comprises steps of generating a dependency graph based upon dependencies between the structure query transformers; and determining a sequence of execution of the structured query transformers based upon the dependency graph; and transfer the processed data files to a data warehouse.

A method of performing local device attestation is provided. The method includes receiving, by a trustable device, an instruction from a user to perform attestation of a user device, creating, by the trustable device, an attestation request comprising a random number whereby attestation evidence is generated and includes the random number, the attestation evidence is then signed, and the attestation evidence is communicated to a verifier to generate an attestation report of the attestation evidence, receiving, by the trustable device, the attestation result of the verifier, the attestation result including a signature of the verifier, and validating, by the trustable device, content of the user device by verifying the signature of the verifier in the attestation report.

This invention discloses a hallucination detection method and system based on atomic statement decomposition and multi-source evidence fusion. The hallucination detection method includes: responding to a user query input by the user, acquiring the answer text generated by the retrieval enhancement generation system and multiple retrieval evidence documents; decomposing the answer text into multiple atomic statements; classifying each atomic statement to obtain the statement type corresponding to the atomic statement; using a statement validator to verify the atomic statements based on the retrieval evidence documents to obtain a verification score; determining the fusion weight according to the statement type of the atomic statement, and calculating the comprehensive score of the atomic statement based on the fusion weight and the verification score; marking the verification status of the atomic statements according to the comprehensive score, and marking atomic statements with abnormal verification status in the answer text. This invention solves the problems of excessively coarse granularity of hallucination detection and inaccurate error localization in the RAG system through a fine-grained atomic statement and verification mechanism, thereby improving the interpretability of the answer content.

Various aspects of the technologies described herein relate to systems, methods, and machine-readable media for cross-chain communication on a blockchain platform. Various aspects may include accepting a first transaction, comprising a message and a message payload, on a first blockchain. The aspects may also include verifying a message on the first blockchain by signing the message using the signing keys of one or more validators in a first set of validators on the first blockchain. The aspects may also include generating an aggregate signature based on the signing keys of one or more validators in a first set of validators. The aspects may also include submitting a second transaction, comprising a message and an aggregate signature, on a second blockchain. The aspects may also include verifying the second transaction on the second blockchain based on a shared registry.

Systems, computer program products, and methods are described herein for authoring, validating, and indexing inbound and outbound documents. The invention utilizes a fine-tuned Large Language Model (LLM) for authoring, validation, and indexing of both outbound and inbound correspondence within regulated entities. This system is designed to enhance efficiency and compliance by addressing three key areas: document authoring, content validation, and document indexing. For outbound document distribution and final artifact generation, the system incorporates a content authoring co-pilot that employs a prompt-based content generation, tagging, and content translation, while a content validator checks for static content, fonts, styles, logos, and images. For inbound document receiving, the system automates classification, data extraction, and document comparison. This optimization supports a unified approach to handling both outgoing and incoming documents, reducing manual effort and ensuring consistency in document processing across diverse channels.

Systems and methods are validating data in a data set. A data set including data to validate and a validator to use in validating the data is selected based on user input generated based on interactions of a user with a graphical user interface. The validator is applied to the data to determine whether one or more statistics generated through application of the validator to the data is valid or invalid based on a validation routine associated with the validator. A data quality report indicating whether the data set is valid or invalid, based on a determination of whether the one or more statistics is valid or invalid, is generated and selectively presented to the user through the graphical user interface.