US-20260127616-A1 - Neuro-Symbolic Intelligence Configured for Transaction Dispute Resolution

US20260127616A1US 20260127616 A1US20260127616 A1US 20260127616A1US-20260127616-A1

Abstract

Transactions may be disputed. Accordingly, streamlining and anticipating possible disputes with transactions may help achieve improved transaction processing efficiency and enhance customer satisfaction. In some arrangements, a transaction processing and dispute resolution system may determine a dispute propensity score for a transaction in order to anticipate and possibly avoid possible disputes for a transaction. This determination may be performed by configuring and applying transaction parameters to a neuro-symbolic artificial intelligence (AI) model. Additionally, the transaction processing and dispute resolution system may also automatically identify resolutions to transaction disputes when they arise using smart contracts. In such a system, the transaction-in-dispute may be analyzed using a neuro-symbolic AI model to identify possible updates to the smart contract rules. This allows the model to further adapt and learn based on new contexts and information.

Inventors

Jemlin Lucas
Suryanarayana Adivi
Pushkar Taneja

Assignees

BANK OF AMERICA CORPORATION

Dates

Publication Date: 20260507
Application Date: 20241104

Claims (20)

1 . A method comprising: detecting, by a dispute resolution processing computing system, a dispute associated with a transaction of a first entity; determining one or more transaction parameters for the transaction subject to the dispute, the transaction parameters including a device type or model through which a transaction request, corresponding to the transaction of the first entity, was made, a geographic location from which the transaction request was received, and a unique identifier of the first entity; blocking transactions for one or more accounts of the first entity; determining, using a first neuro-symbolic artificial intelligence model, that the transaction is complex, wherein determining that the transaction is complex includes analyzing, using the first neuro-symbolic artificial intelligence model, the transaction and historical transactions of the first entity to determine that the transaction establishes a pattern of transaction behavior that is new to the first entity; in response to determining that the transaction is complex: generating one or more modifications to one or more existing smart contracts in a blockchain network using a second neuro-symbolic artificial intelligence model different from the first neuro-symbolic artificial intelligence model, wherein the one or more smart contracts are configured to generate dispute resolutions, the second neuro-symbolic artificial intelligence model being specific to a transaction type of the transaction of the first entity; updating the one or more smart contracts based on the generated one or more modifications; and after updating the one or more existing smart contracts, determining a resolution to the dispute using the updated one or more smart contracts.
2 . The method of claim 1 , wherein determining, using the first neuro-symbolic artificial intelligence model, that the transaction is complex, includes: obtaining transaction data for the historical transactions; analyzing the transaction data along with the one or more transaction parameters for the transaction subject to the dispute to identify one or more transaction patterns; and determining that at least one of the one or more transaction patterns are new.
3 . The method of claim 1 , wherein updating the one or more smart contracts includes: generating a new node in the blockchain with a new smart contract having one or more updated rules.
4 . The method of claim 3 , further comprising: associating the new node with an existing node in the blockchain, the existing node in the blockchain storing a prior version of the new smart contract.
5 . (canceled)
6 . The method of claim 1 , further comprising generating the first neuro-symbolic artificial intelligence model, including: obtaining transaction data for one or more historical transactions; extracting transaction features from the transaction data; defining a plurality of symbolic rules using at least one of the plurality of transaction features; and generating a plurality of machine learning nodes and connections, wherein each of the nodes represents one of the transaction features and each of the connections represents a relationship between at least two of the nodes.
7 . The method of claim 1 , further comprising selecting the second neuro-symbolic artificial intelligence model based on the transaction type.
8 . An apparatus comprising: a processor; and memory storing computer-readable instructions that, when executed, cause the apparatus to: detect a dispute associated with a transaction of a first entity; determine one or more transaction parameters for the transaction subject to the dispute, the transaction parameters including a device type or model through which a transaction request, corresponding to the transaction associated with the dispute, was made, a geographic location from which the transaction request was received, and a unique identifier of the first entity; block transactions for one or more accounts of the first entity; determine, using a first neuro-symbolic artificial intelligence model, whether the transaction is complex, wherein determining whether the transaction is complex includes analyzing, using the first neuro-symbolic artificial intelligence model and historical transactions of the first entity, whether the transaction establishes a pattern of transaction behavior that is new to the first entity; in response to determining that the transaction is complex ;: generate one or more modifications to one or more existing smart contracts in a blockchain network using a second neuro-symbolic artificial intelligence model different from the first neuro-symbolic artificial intelligence model, wherein the one or more smart contracts are configured to generate dispute resolutions, the second neuro-symbolic artificial intelligence model being specific to a transaction type of the transaction of the first entity; update the one or more smart contracts based on the generated one or more modifications; and after updating the one or more existing smart contracts, determine a resolution to the dispute using the updated one or more smart contracts.
9 . The apparatus of claim 8 , wherein determining, using the first neuro-symbolic artificial intelligence model, whether the transaction is complex, includes: obtaining transaction data for the historical transactions; analyzing the transaction data along with the one or more transaction parameters for the transaction subject to the dispute to identify one or more transaction patterns; and determining whether any of the identified one or more transaction patterns are new.
10 . The apparatus of claim 8 , wherein updating the one or more smart contracts includes: generating a new node in the blockchain with a new smart contract having one or more updated rules.
11 . The apparatus of claim 10 , wherein the instructions, when executed, further cause the apparatus to: associate the new node with an existing node in the blockchain, the existing node in the blockchain storing a prior version of the new smart contract.
12 . The apparatus of claim 8 , wherein the instructions, when executed, further cause the apparatus to: in response to determining that the transaction is not complex, determining the resolution to the dispute using the one or more existing smart contracts.
13 . The apparatus of claim 8 , wherein the instructions, when executed, further cause the apparatus to generate the first neuro-symbolic artificial intelligence model, including: obtaining transaction data for the historical transactions; extracting transaction features from the transaction data; defining a plurality of symbolic rules using at least one of the plurality of transaction features; and generating a plurality of machine learning nodes and connections, wherein each of the nodes represents one of the transaction features and each of the connections represents a relationship between at least two of the nodes.
14 . The apparatus of claim 8 , wherein the instructions, when executed, further cause the apparatus to select the second neuro-symbolic artificial intelligence model based on the transaction type.
15 . A non-transitory computer-readable medium storing computer readable instructions that, when executed, cause an apparatus to: detect a dispute associated with a transaction of a first entity; determine one or more transaction parameters for the transaction subject to the dispute, the transaction parameters including a device type or model through which a transaction request, corresponding to the transaction associated with the dispute, was made, a geographic location from which the transaction request was received, and a unique identifier of the first entity; block transactions for one or more accounts of the first entity; determine, using a first neuro-symbolic artificial intelligence model, whether the transaction is complex, wherein determining whether the transaction is complex includes analyzing, using the first neuro-symbolic artificial intelligence model and historical transactions of the first entity, whether the transaction establishes a pattern of transaction behavior that is new to the first entity; in response to determining that the transaction is complex: generate one or more modifications to one or more existing smart contracts in a blockchain network using a second neuro-symbolic artificial intelligence model different from the first neuro-symbolic artificial intelligence model, wherein the one or more smart contracts are configured to generate dispute resolutions, the second neuro-symbolic artificial intelligence model being specific to a transaction type of the transaction of the first entity; update the one or more smart contracts based on the generated one or more modifications; and after updating the one or more existing smart contracts, determine a resolution to the dispute using the updated one or more smart contracts.
16 . The non-transitory computer-readable medium of claim 15 , wherein determining, using the first neuro-symbolic artificial intelligence model, whether the transaction is complex, includes: obtaining transaction data for the historical transactions; analyzing the transaction data along with the one or more transaction parameters for the transaction subject to the dispute to identify one or more transaction patterns; and determining whether any of the identified one or more transaction patterns are new.
17 . The non-transitory computer-readable medium of claim 15 , wherein updating the one or more smart contracts includes: generating a new node in the blockchain with a new smart contract having one or more updated rules.
18 . The non-transitory computer-readable medium of claim 17 , wherein the instructions, when executed, further cause the apparatus to: associate the new node with an existing node in the blockchain, the existing node in the blockchain storing a prior version of the new smart contract.
19 . The non-transitory computer-readable medium of claim 15 , wherein the instructions, when executed, further cause the apparatus to: in response to determining that the transaction is not complex, determining the resolution to the dispute using the one or more existing smart contracts.
20 . The non-transitory computer-readable medium of claim 15 , wherein the instructions, when executed, further cause the apparatus to generate the first neuro-symbolic artificial intelligence model, including: obtaining transaction data for the historical transactions; extracting transaction features from the transaction data; defining a plurality of symbolic rules using at least one of the plurality of transaction features; and generating a plurality of machine learning nodes and connections, wherein each of the nodes represents one of the transaction features and each of the connections represents a relationship between at least two of the nodes.

Description

BACKGROUND Aspects described herein relate to electrical computers, systems, and devices for processing transactions and possible transaction disputes in an efficient, secure, and reliable manner. Online and computer transactions have rapidly evolved, and real-time transactions have become the norm. Although such transactions provide convenience and efficiency, disputes arising from such transactions are increasing and oftentimes pose significant challenges. Existing dispute resolution processes are largely manual and can be slow and labor-intensive. Accordingly, such processes may lead to delays, additional costs, and negative impacts on the customer experience and satisfaction. In one example, users have the capability to make payments through a mobile app to transfer funds to another user's account. However, in some cases, a transaction may be disputed by the sending bank due to various reasons, including the sender's account not having less than the amount of needed funds. Some current dispute resolution processes would require a customer to visit a physical bank facility or raise the complaint in a customer center, provide identification documents, and wait for the bank to investigate and resolve the issue. This process can take several days or even weeks, adding to computing processing load, consuming personnel resources, and causing frustration and inconvenience to the customer or user. BRIEF SUMMARY The following presents a simplified summary in order to provide a basic understanding of some aspects of the disclosure. The summary is not an extensive overview of the disclosure. It is neither intended to identify key or critical elements of the disclosure nor to delineate the scope of the disclosure. The following summary merely presents some concepts of the disclosure in a simplified form as a prelude to the description below. Aspects described herein relate to systems, methods, apparatuses and processes that combine artificial intelligence (AI) models, blockchain technology, adaptive smart contracts, and AI-enhanced oracles to streamline claims settlement and dispute resolution. In some arrangements, the systems, methods, apparatuses and processes implement a neuro-symbolic AI model to facilitate the settlement and resolution. The systems, methods, apparatuses and processes are configured to analyze complex rules and conditions, dynamically adjust security settings, and make informed decisions in real-time. In some examples, the neuro-symbolic AI model may provide reasoning capabilities to interpret complex rules and conditions, enabling the system to understand the context of each transaction and make informed decisions that consider the specific circumstances of each case. According to one or more aspects, systems, methods, apparatuses and processes may learn from transaction and/or dispute resolution data and adapt its behavior to improve dispute resolution, ensuring that the system becomes more effective over time. According to one or more aspects, the systems, methods, apparatuses and processes include smart contract orchestration features which are configured to provide reasoning capabilities to interpret complex rules and conditions, enabling dynamic adjustment of smart contract rules and behavior based on AI predictions and analysis. According to one or more further aspects, systems, methods, apparatuses and processes provide dynamic rule updates, such as rules applied in a smart contract. Such a feature may analyze data and context to suggest updates or modifications to the rules governing smart contracts, ensuring that the system adapts to changing circumstances and remains up-to-date. According to one or more aspects, the systems, methods, apparatuses and processes described herein provide fraud detection and prevention features. For example, the system's fraud detection and prevention capabilities may identify high-risk transactions and dynamically adjust security settings to prevent fraud, ensuring the security and integrity of transactions. According to further aspects, the systems, methods, apparatuses and processes are configured to provide more detailed and clearer explanations for dispute resolution decisions and improves transparency throughout the transaction completion and dispute resolution process, enabling stakeholders to understand the reasoning behind the system's decisions. These features, along with many others, are discussed in greater detail below. BRIEF DESCRIPTION OF THE DRAWINGS The present disclosure is illustrated by way of example and not limited in the accompanying figures in which like reference numerals indicate similar elements and in which: FIGS. 1A and 1B depict an illustrative computing environment for a transaction and dispute processing system and service in accordance with one or more aspects described herein; FIGS. 2A and 2B illustrate an example process flow for transaction processing and dispute resolution accordance with one or more aspects described