Method and system for generating contextual notification of foreign exchange fluctuation

The system addresses the lack of retrospective analysis in foreign exchange by generating context notifications from a context vector, combining internal and external market information, and tailoring content to user expertise, ensuring real-time performance and information density.

KR103046489B1Active Publication Date: 2026-11-10MOIN
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
KR1020260099507
Authority / Receiving Office
KR · KR
Patent Type
Patents
Current Assignee / Owner
Priority Date
2026-05-15
Filing Date
2026-06-01
Publication Date
2026-11-10
Estimated Expiration
2046-06-01

AI Technical Summary

Technical Problem

Existing technologies in the foreign exchange market lack the ability to retrospectively analyze exchange rate fluctuations and reconstruct complex contexts by combining internal and external market information, failing to provide context notifications that cater to the user's expertise level and ensuring real-time performance and information density.

Method used

A system and method that detects exchange rate fluctuations, collects heterogeneous information before the event, generates a context vector, and inputs it into a generative model to create context notifications, incorporating external and internal market information, and adjusts content based on user expertise and notification urgency.

Benefits of technology

Enables real-time generation of context notifications that reflect complex market contexts, cater to user expertise, and ensure information density by providing draft and additional explanatory content when needed, enhancing user understanding and decision-making.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 112026066538325-PAT00002_ABST
    Figure 112026066538325-PAT00002_ABST
Patent Text Reader

Abstract

According to one aspect of the present invention, a method for generating a foreign exchange fluctuation context notification is provided, comprising the steps of: detecting an exchange rate fluctuation event based on fluctuation conditions; collecting heterogeneous information that occurred during a time interval preceding the occurrence of the exchange rate fluctuation event in response to the detection of the exchange rate fluctuation event, and combining the collected heterogeneous information to generate a context vector; and inputting the generated context vector into a generative model learned based on data regarding the foreign exchange field to generate a context notification content regarding the exchange rate fluctuation event.
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The present invention relates to a method and system for generating foreign exchange fluctuation context notifications. Background Technology

[0002] With the advancement of information and communication technology and artificial intelligence technology, various information regarding the foreign exchange market is being collected, processed, and provided in real time, and various users are utilizing information on exchange rate fluctuations for their decision-making.

[0003] In the foreign exchange market, users may receive notifications when the exchange rate of a currency pair of interest reaches a preset threshold or fluctuates above a certain level. For example, notifications containing numerical information about the exchange rate or information regarding whether a threshold has been reached may be provided to the user, such as 'USD / KRW breaks through 1,380 won' or 'EUR / KRW breaks below 1,500 won'.

[0004] Meanwhile, technologies that automatically collect and analyze information regarding the foreign exchange market to predict price changes or generate reports have been presented in the past. However, such prior technologies are limited to predicting future prices or periodically generating standardized reports; they have not presented technologies that retrospectively analyze already occurred exchange rate fluctuations based on the time of occurrence to provide the user with the context of those fluctuations. Furthermore, prior technologies have not presented technologies that reconstruct the context of exchange rate fluctuations from various angles by combining information on events occurring outside the foreign exchange market with information on patterns observed inside the market.

[0005] Accordingly, the inventor(s) propose a technology that enables the context of an exchange rate fluctuation to be immediately provided to a user by detecting an exchange rate fluctuation event based on fluctuation conditions, collecting heterogeneous information generated during a time interval preceding the occurrence of the exchange rate fluctuation event in response to the detection of the exchange rate fluctuation event, combining the collected heterogeneous information to generate a context vector, and inputting the context vector into a generative model learned based on data regarding the foreign exchange field to generate context notification content regarding the exchange rate fluctuation event. Prior art literature

[0006] Korean Published Patent Application No. 10-2017-0041094 (April 14, 2017) The problem to be solved

[0007] The present invention aims to solve at least some of the problems of the aforementioned prior art.

[0008] In addition, the present invention has another objective of detecting an exchange rate fluctuation event based on fluctuation conditions, collecting heterogeneous information generated during a time interval preceding the occurrence of the exchange rate fluctuation event in response to the detection of the exchange rate fluctuation event, combining the collected heterogeneous information to generate a context vector, and inputting the context vector into a generative model learned based on data regarding the foreign exchange field to generate context notification content regarding the exchange rate fluctuation event.

[0009] In addition, the present invention has another objective of enabling context notification content to be generated from a context vector that combines information about events occurring outside the foreign exchange market and information about aspects observed inside the foreign exchange market, thereby allowing complex contexts that are difficult to reconstruct from a single source of information to be reflected in the context notification content.

[0010] In addition, the present invention has another objective of enabling experts and non-experts to accept the context according to their respective levels of understanding, even for the same context, by generating context notification content at an expression level corresponding to the user's level of expertise.

[0011] In addition, the present invention has another objective of ensuring real-time performance and information density simultaneously by sending draft content generated based on a template first and additional explanatory content generated based on a generation model subsequently when conditions regarding the urgency of the notification are satisfied. means of solving the problem

[0012] A representative configuration of the present invention for achieving the above objective is as follows.

[0013] According to one aspect of the present invention, a method for generating a foreign exchange fluctuation context notification is provided, comprising the steps of: detecting an exchange rate fluctuation event based on fluctuation conditions; collecting heterogeneous information that occurred during a time interval preceding the occurrence of the exchange rate fluctuation event in response to the detection of the exchange rate fluctuation event, and combining the collected heterogeneous information to generate a context vector; and inputting the generated context vector into a generative model learned based on data regarding the foreign exchange field to generate a context notification content regarding the exchange rate fluctuation event.

[0014] According to another aspect of the present invention, a system for generating foreign exchange fluctuation context notifications is provided, comprising: a foreign exchange fluctuation detection unit that detects a foreign exchange fluctuation event based on fluctuation conditions; a complex context collection unit that collects heterogeneous information generated during a time interval preceding the occurrence of the foreign exchange fluctuation event in response to the detection of the foreign exchange fluctuation event, and combines the collected heterogeneous information to generate a context vector; and a context notification content generation unit that inputs the generated context vector into a generative model learned based on data regarding the foreign exchange field to generate context notification content regarding the foreign exchange fluctuation event.

[0015] In addition to this, other methods for implementing the present invention, other systems, and non-transient computer-readable recording media for recording a computer program for executing said methods are further provided. Effects of the invention

[0016] According to the present invention, an exchange rate fluctuation event is detected based on fluctuation conditions, heterogeneous information generated during a time interval preceding the occurrence of the exchange rate fluctuation event is collected in response to the detection of the exchange rate fluctuation event, and the collected heterogeneous information is combined to generate a context vector, and the context vector is input into a generative model learned based on data regarding the foreign exchange field to generate context notification content regarding the exchange rate fluctuation event.

[0017] In addition, according to the present invention, context notification content is generated from a context vector combining information regarding events occurring outside the foreign exchange market and information regarding aspects observed inside the foreign exchange market, thereby enabling complex contexts that are difficult to reconstruct using only a single source of information to be reflected in the context notification content.

[0018] In addition, according to the present invention, context notification content is generated at an expression level corresponding to the user's level of expertise, thereby enabling experts and non-expert users to accept the context according to their respective levels of understanding, even for the same context.

[0019] In addition, according to the present invention, when conditions regarding the urgency of the notification are satisfied, draft content generated based on a template is sent first, and additional explanatory content generated based on a generation model is sent subsequently, thereby enabling real-time performance and information density to be secured simultaneously. Brief explanation of the drawing

[0020] FIG. 1 is a diagram showing the schematic configuration of an entire system for generating foreign exchange fluctuation context notifications according to one embodiment of the present invention. FIG. 2 is a drawing illustrating in detail the internal configuration of a notification generation system according to one embodiment of the present invention. Specific details for implementing the invention

[0021] The following detailed description of the invention refers to the accompanying drawings, which illustrate specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that various embodiments of the invention are different but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be modified from one embodiment to another without departing from the spirit and scope of the invention. It should also be understood that the location or arrangement of individual components within each embodiment may be modified without departing from the spirit and scope of the invention. Accordingly, the following detailed description is not meant to be limiting, and the scope of the invention should be understood to encompass the scope claimed by the claims and all equivalents thereof. Similar reference numerals in the drawings indicate identical or similar components across various aspects.

[0022] Hereinafter, in order to enable a person skilled in the art to easily practice the present invention, various preferred embodiments of the present invention will be described in detail with reference to the attached drawings.

[0023] Configuration of the entire system

[0024] FIG. 1 is a diagram showing the schematic configuration of an entire system for generating foreign exchange fluctuation context notifications according to one embodiment of the present invention.

[0025] As illustrated in FIG. 1, the entire system according to one embodiment of the present invention may include a communication network (100), a notification generation system (200), and a device (300).

[0026] First, a communication network (100) according to one embodiment of the present invention can be configured regardless of the mode of communication, such as wired communication or wireless communication, and can be configured as various communication networks such as a Local Area Network (LAN), a Metropolitan Area Network (MAN), or a Wide Area Network (WAN). Preferably, the communication network (100) referred to in this specification may be the known Internet or the World Wide Web (WWW). However, the communication network (100) may include at least a part of a known wired / wireless data communication network, a known telephone network, or a known wired / wireless television communication network, without being limited thereto.

[0027] For example, the communication network (100) may be a wireless data communication network and may implement conventional communication methods such as WiFi communication, WiFi-Direct communication, Long Term Evolution (LTE) communication, 5G communication, Bluetooth communication (including Bluetooth Low Energy (BLE) communication), infrared communication, ultrasonic communication, etc., in at least a part thereof. As another example, the communication network (100) may be an optical communication network and may implement conventional communication methods such as Light Fidelity (LiFi), etc., in at least a part thereof.

[0028] Next, a notification generation system (200) according to one embodiment of the present invention can perform the function of detecting an exchange rate fluctuation event based on fluctuation conditions, collecting heterogeneous information that occurred during a time interval preceding the occurrence of the exchange rate fluctuation event in response to the detection of the exchange rate fluctuation event, generating a context vector by combining the collected heterogeneous information, and generating context notification content regarding the exchange rate fluctuation event by inputting the context vector into a generative model learned based on data regarding the foreign exchange field.

[0029] The configuration and function of the notification generation system (200) according to the present invention will be examined in detail through the following detailed description.

[0030] Next, the device (300) according to one embodiment of the present invention is a digital device that includes a function to communicate after connecting to the notification generation system (200), and any digital device equipped with memory means and equipped with a microprocessor to have computational capabilities, such as a smartphone, tablet, smart watch, smart band, smart glasses, desktop computer, laptop computer, workstation, PDA, web pad, mobile phone, etc., can be adopted as the device (300) according to the present invention.

[0031] In particular, the device (300) may include an application (not shown) that enables a user to receive a service according to an embodiment of the present invention from the notification generation system (200). Such an application may be downloaded from the notification generation system (200) or an external application distribution server (not shown). Meanwhile, the nature of such an application may generally be similar to the exchange rate fluctuation detection unit (210), complex context collection unit (220), context notification content generation unit (230), communication unit (240), and control unit (250) of the notification generation system (200) as described below. Here, at least a part of the application may be replaced with a hardware device or firmware device capable of performing substantially the same or equivalent functions as needed.

[0032] In addition, according to one embodiment of the present invention, the device (300) may include a function to provide information necessary for the notification generation system (200) or to reflect the results produced by the notification generation system (200) into these systems by communicating with the enterprise resource planning (ERP) system, trade management platform, accounting system, fund management system, etc., in a manner such as API integration.

[0033] Configuration of the notification generation system

[0034] Below, we will examine the internal configuration of the notification generation system (200) that performs important functions for the implementation of the present invention and the functions of each component.

[0035] FIG. 2 is a drawing illustrating in detail the internal configuration of a notification generation system (200) according to one embodiment of the present invention.

[0036] As illustrated in FIG. 2, a notification generation system (200) according to one embodiment of the present invention may be configured to include an exchange rate fluctuation detection unit (210), a complex context collection unit (220), a context notification content generation unit (230), a communication unit (240), and a control unit (250). According to one embodiment of the present invention, the exchange rate fluctuation detection unit (210), the complex context collection unit (220), the context notification content generation unit (230), the communication unit (240), and the control unit (250) may be program modules in which at least some of them communicate with an external system (not shown). Such program modules may be included in the notification generation system (200) in the form of an operating system, an application program module, or other program modules, and may be physically stored in various known storage devices. Additionally, such program modules may be stored in a remote storage device capable of communicating with the notification generation system (200). Meanwhile, such program modules include, but are not limited to, routines, subroutines, programs, objects, components, data structures, etc. that perform specific tasks or execute specific abstract data types as described below according to the present invention.

[0037] Meanwhile, although the notification generation system (200) has been described as above, this description is exemplary, and it is obvious to those skilled in the art that at least some of the components or functions of the notification generation system (200) may be realized within a device (300) or a server (not shown) or included within an external system (not shown) as needed.

[0038] For example, at least some of the components or functions of the notification generation system (200) may be operated in a manner where they are embedded in at least one of (i) an enterprise’s ERP system or trade management platform, (ii) a foreign exchange trading system of a financial institution (bank, securities firm, etc.), and (iii) an external market data system (a price quote provision system, a forward exchange quote provision system, etc.). However, there are no limitations on this method of operation, and it may be modified in various ways within the scope of achieving the purpose of the present invention.

[0039] First, the exchange rate fluctuation detection unit (210) according to one embodiment of the present invention can perform the function of detecting an exchange rate fluctuation event based on fluctuation conditions.

[0040] Specifically, an exchange rate fluctuation detection unit (210) according to one embodiment of the present invention can detect an exchange rate fluctuation event by receiving real-time or quasi-real-time exchange rate information regarding at least one currency pair and determining whether the received exchange rate information satisfies fluctuation conditions.

[0041] An exchange rate fluctuation detection unit (210) according to one embodiment of the present invention may obtain exchange rate information through various routes. For example, the exchange rate fluctuation detection unit (210) may receive exchange rate information in real-time or near-real-time from an external market data system (e.g., a price provision system, a financial information terminal, etc.) or a foreign exchange trading system of a financial institution. As another example, the exchange rate fluctuation detection unit (210) may receive exchange rate information in a streaming manner through an official API of an external system, a message queue, a WebSocket connection, etc. However, the method of obtaining exchange rate information is not limited thereto and may be changed in various ways within the scope of achieving the purpose of the present invention.

[0042] In addition, the fluctuation conditions according to one embodiment of the present invention may include at least one of a condition regarding the absolute fluctuation range of the exchange rate, a condition regarding the fluctuation rate of the exchange rate, and a condition regarding the breakthrough of a technical indicator.

[0043] For example, a condition regarding the absolute fluctuation range of the exchange rate may be a condition in which the absolute fluctuation range of the exchange rate during a unit of time (e.g., 1 minute, 5 minutes, 15 minutes, 1 hour, etc.) exceeds a threshold value (e.g., 5 won, 10 won, 15 won, etc.). As another example, a condition regarding the rate of change of the exchange rate may be a condition in which the rate of change of the exchange rate during a unit of time exceeds a threshold value (e.g., 0.3%, 0.5%, 1%, etc.). As yet another example, a condition regarding the breakout of a technical indicator may be a condition corresponding to a breakout of a moving average line (e.g., an upward or downward breakout of the 20-day moving average line), entry into an overbought or oversold zone of the Relative Strength Index (RSI) (e.g., when the RSI is 70 or higher or 30 or lower), or a breakout of the upper or lower Bollinger Bands. However, the types of variation conditions and their threshold values ​​are not limited thereto and may be varied within the scope of achieving the purpose of the present invention.

[0044] In addition, the fluctuation condition according to one embodiment of the present invention may be composed of a combination of a plurality of conditions. For example, the exchange rate fluctuation detection unit (210) may detect an exchange rate fluctuation event in response to both the condition regarding the absolute fluctuation range of the exchange rate and the condition regarding the breakthrough of the technical indicator being satisfied, and for another example, it may detect an exchange rate fluctuation event in response to at least one of the condition regarding the absolute fluctuation range of the exchange rate, the condition regarding the fluctuation rate of the exchange rate, and the condition regarding the breakthrough of the technical indicator being satisfied.

[0045] In addition, the exchange rate fluctuation detection unit (210) according to one embodiment of the present invention may apply different fluctuation conditions for each currency pair. For example, by applying a fluctuation condition having a relatively large threshold value to a currency pair with relatively high volatility and applying a fluctuation condition having a relatively small threshold value to a currency pair with relatively low volatility, fluctuation conditions corresponding to the volatility of each currency pair may be set. As another example, the exchange rate fluctuation detection unit (210) may dynamically adjust the fluctuation conditions by further referring to the market opening time, the average volatility during the preceding period, etc.

[0046] In addition, the exchange rate fluctuation detection unit (210) according to one embodiment of the present invention may apply fluctuation conditions differently depending on the market phase determined based on the level of volatility of the foreign exchange market.

[0047] Specifically, a market phase according to one embodiment of the present invention may include at least one of a stable phase in which the volatility of the foreign exchange market is stable, a volatility expansion phase in which the volatility of the foreign exchange market is expanded, and a crisis phase in which a systemic crisis occurs in the foreign exchange market. According to one embodiment of the present invention, a market phase may be classified based on at least one of an option implied volatility indicator (e.g., USD / KRW 1-month ATM implied volatility), the VIX index, the range of change in the interest rate difference between currency pairs, the range of change in the bid-ask spread, etc.

[0048] For example, the exchange rate fluctuation detection unit (210) may apply a fluctuation condition with a relatively small threshold value when the volatility of the foreign exchange market is classified as a stable period, thereby allowing an exchange rate fluctuation event to be detected even by a small exchange rate fluctuation, so as to provide rich contextual information to the user, and may apply a fluctuation condition with a relatively large threshold value when the volatility of the foreign exchange market is classified as an volatility expansion period, thereby adjusting the frequency of notifications and alleviating the user's notification fatigue. As another example, when the volatility of the foreign exchange market is classified as a crisis phase, the exchange rate fluctuation detection unit (210) may further adjust the threshold value of the fluctuation condition and at the same time set the length of the time interval described later to be shorter, so that external events and internal aspects that have the most direct impact when an exchange rate fluctuation event occurs may be reflected in the context vector first.

[0049] However, the types of market phases, the method of classifying market phases, and the method of applying differential conditions of variation according to market phases are not limited thereto and may be modified in various ways within the scope of achieving the purpose of the present invention.

[0050] An exchange rate fluctuation detection unit (210) according to one embodiment of the present invention can, when detecting an exchange rate fluctuation event, generate event information including information regarding the time of occurrence of the exchange rate fluctuation event, currency pair, direction of fluctuation, fluctuation range, type of satisfied fluctuation condition, etc., and transmit it to a complex context collection unit (220) described later, thereby triggering subsequent processing.

[0051] Next, a complex context collection unit (220) according to one embodiment of the present invention can perform the function of collecting heterogeneous information that occurred during a time interval preceding the time of occurrence of an exchange rate fluctuation event in response to the detection of an exchange rate fluctuation event, and combining the collected heterogeneous information to generate a context vector.

[0052] Specifically, a time interval according to one embodiment of the present invention may be a time interval with the point of occurrence of an exchange rate fluctuation event as the end point and a point in time that is retroactive from the point of occurrence by a preset amount of time (e.g., 5 minutes, 15 minutes, 30 minutes, 1 hour, etc.). According to one embodiment of the present invention, the length of the time interval may be determined differently depending on the currency pair, the user, or the market conditions. For example, a short time interval (e.g., 5 minutes to 15 minutes) may be applied during periods when the volatility of the foreign exchange market is relatively high, and a long time interval (e.g., 30 minutes to 1 hour) may be applied during periods when the volatility is relatively low, and a time interval set by the user may also be applied.

[0053] In addition, heterogeneous information according to one embodiment of the present invention may include a combination of information regarding events occurring outside the foreign exchange market and information regarding aspects observed inside the foreign exchange market.

[0054] Specifically, information regarding events occurring outside the foreign exchange market according to one embodiment of the present invention may refer to exogenous information that can influence the foreign exchange market independently of the trading dynamics of the foreign exchange market itself. For example, information regarding events occurring outside the foreign exchange market may include at least one of: (i) information regarding monetary policy announcements by central banks (e.g., resolutions of the Monetary Policy Committee of the Bank of Korea, statements of the Federal Open Market Committee (FOMC) of the U.S. Federal Reserve System, monetary policy announcements by the European Central Bank (ECB), etc.); (ii) information regarding announcements of economic indicators by government agencies (e.g., announcements of employment indicators by the U.S. Department of Labor, press releases on import and export trends by the Ministry of Trade, Industry and Energy of Korea, announcements of the consumer price index, etc.); (iii) news reports regarding geopolitical events; (iv) professional columns or reports by market participants regarding the foreign exchange market; and (v) information regarding financial statements of major companies or disclosures related to trade transactions.

[0055] In addition, information regarding aspects observed within the foreign exchange market according to one embodiment of the present invention may refer to endogenous information directly observed from trading dynamics or price dynamics within the foreign exchange market. For example, information regarding aspects observed within the foreign exchange market may include at least one of: (i) information regarding changes in technical indicators (e.g., breakout of a moving average line, entry into an overbought or oversold zone of the RSI, deviation from the upper or lower Bollinger Band, crossing of the signal line of a momentum indicator (MACD), etc.); (ii) information regarding changes in interest rate differences between currency pairs; (iii) information regarding changes in the Non-Deliverable Forward (NDF) spread; (iv) information regarding changes in swap points; (v) information regarding changes in bid-ask spreads and trading volume; and (vi) information regarding changes in foreign exchange reserves.

[0056] In addition, the complex context collection unit (220) according to one embodiment of the present invention can collect heterogeneous information through various routes. For example, the complex context collection unit (220) can collect information regarding events occurring outside the foreign exchange market through methods such as RSS (Rich Site Summary) feeds, official APIs, web scraping, and messaging channels from external news provision systems, official websites of central banks or government agencies, financial information terminals, etc. As another example, the complex context collection unit (220) can receive time-series data of exchange rates, quote data, interest rate data, NDF data, swap point data, etc. from external market data systems, foreign exchange trading systems of financial institutions, etc., and derive information regarding aspects observed within the foreign exchange market from the received data.

[0057] In addition, the complex context collection unit (220) according to one embodiment of the present invention may perform appropriate preprocessing on the collected heterogeneous information according to the form of the information. For example, the complex context collection unit (220) may perform processing such as morphological analysis, tokenization, lemmatization, term recognition based on a foreign exchange domain term dictionary (e.g., a dictionary containing foreign exchange domain terms such as NDF, swap point, offshore exchange rate, foreign exchange reserves, currency swap (CRS), etc.), and embedding for information in text form, and may perform processing such as normalization, missing value interpolation, time series matching, and difference transformation for information in numeric form. For information in voice or video form, processing to convert it into text form may be performed by performing speech-to-text processing, subtitle extraction processing, and visual information extraction processing using a vision-language model.

[0058] A complex context collection unit (220) according to one embodiment of the present invention can generate a context vector by combining preprocessed heterogeneous information. A context vector according to one embodiment of the present invention may be a structured vector in which heterogeneous information is time-matched and arranged in slots according to information sources.

[0059] For example, the context vector may be a multidimensional vector in which at least one component representing information regarding events occurring outside the foreign exchange market and at least one component representing information regarding aspects observed inside the foreign exchange market are aligned on the same time axis and combined. As another example, the context vector may be a sequence vector in which heterogeneous information is aligned in a time-series form, a vector in which partial vectors embedded by information source are concatenated, or a vector encoded with an attention-based representation that reflects the interaction between information sources. As yet another example, the context vector may be a vector in which weights based on the reliability or influence of the information source are reflected. However, the structure and representation method of the context vector are not limited thereto and may be modified in various ways within the scope of achieving the purpose of the present invention.

[0060] In addition, a context vector according to one embodiment of the present invention may be a vector structured to reflect the temporal causal relationship between an event occurring outside the foreign exchange market and an aspect observed inside the foreign exchange market.

[0061] Specifically, a complex context collection unit (220) according to one embodiment of the present invention recognizes the temporal sequence between the time of occurrence of an event that occurred outside the foreign exchange market, the time of change of an aspect observed inside the foreign exchange market, and the time of occurrence of an exchange rate fluctuation event, and can configure a context vector so that the recognized temporal sequence is explicitly expressed in the context vector.

[0062] For example, the complex context collection unit (220) can construct a context vector such that a temporal causal sequence, such as “an increase in the interest rate difference between Korea and the US (at time T0+5) is observed following the FOMC’s announcement of a freeze on the base rate (at time T0), an increase in the downward pressure of the NDF 1-month spread (at time T0+10), and an increase in USD / KRW (at time T0+15) is observed following”, is explicitly expressed as a component of the context vector.

[0063] As another example, the complex context collection unit (220) may construct a context vector in the form of a graph in which an event occurring outside the foreign exchange market is used as a trigger node, a change in the pattern observed inside the foreign exchange market is used as a follower node, and the time interval and degree of causal influence between the two nodes are used as edges. As yet another example, the complex context collection unit (220) may classify heterogeneous information within a time interval into multi-stage reactions such as first-order reactions, second-order reactions, and third-order reactions according to chronological order, and ensure that the classification results are reflected in the context vector. However, the method of reflecting temporal causal relationships in the context vector is not limited to this and can be varied within the scope of achieving the purpose of the present invention.

[0064] According to the structure of the context vector according to one embodiment of the present invention, beyond simply combining heterogeneous information, a causal path through which an external event propagates influence within the foreign exchange market can be reflected together in the context vector, so that the context notification content generated by the context notification content generation unit (230) described later can be generated in a form that includes a causal explanation rather than a simple list of facts.

[0065] In addition, the complex context collection unit (220) according to one embodiment of the present invention may, when generating a context vector, perform missing value processing or default value processing for information sources where information is absent or insufficient within a time interval, and may construct the context vector by lowering the weight or excluding information sources determined to have low reliability.

[0066] According to the complex context collection unit (220) according to one embodiment of the present invention, a context vector is generated in which information regarding an event occurring outside the foreign exchange market and information regarding an aspect observed inside the foreign exchange market are aligned and combined on the same time axis, thereby allowing a complex context that is difficult to reconstruct with only a single information source to be reflected in the context notification content generation process by the context notification content generation unit (230) described later.

[0067] Next, the context notification content generation unit (230) according to one embodiment of the present invention can perform the function of generating context notification content regarding exchange rate fluctuation events by inputting a context vector generated by the complex context collection unit (220) into a generation model learned based on data regarding foreign exchange fields.

[0068] Specifically, a generative model according to one embodiment of the present invention may be a model trained based on data regarding the foreign exchange field, capable of generating content in the form of natural language regarding exchange rate fluctuation events from context vectors.

[0069] For example, a generative model according to one embodiment of the present invention may be a language model that has been pre-trained on a general domain and adaptedly trained based on data regarding the foreign exchange field (e.g., a foreign exchange domain term dictionary, data on past foreign exchange market fluctuation cases, a corpus of columns and reports regarding the foreign exchange market, monetary policy statements from central banks and economic indicator releases from government agencies, commentary texts from market participants regarding exchange rate fluctuations, etc.). As another example, the generative model may be a model trained from the beginning with data regarding the foreign exchange field, a model to which a Retrieval-Augmented Generation (RAG) method is applied that searches for relevant information from a knowledge base composed of data regarding the foreign exchange field and utilizes it for content generation, or a model additionally trained based on instruction data regarding the foreign exchange field. However, the type and training method of the generative model are not limited thereto and may be varied within the scope of achieving the purpose of the present invention.

[0070] A context notification content generation unit (230) according to one embodiment of the present invention can generate context notification content regarding an exchange rate fluctuation event by inputting a context vector into a generation model.

[0071] In a context notification content according to one embodiment of the present invention, at least one of information regarding the direction of fluctuation of an exchange rate fluctuation event, information regarding the magnitude of fluctuation, and information regarding the persistence of fluctuation may be included. For example, the context notification content may include directional information regarding the rise or fall of the exchange rate, quantitative information regarding the range of fluctuation, and qualitative information regarding the persistence of fluctuation, such as the nature of a short-term shock or a trend. Specifically, a context notification content may be generated in which information regarding the direction of fluctuation (rise), the magnitude of fluctuation (15 won), and the persistence of fluctuation (nature of a short-term shock) is included, such as, "USD / KRW rose by 15 won in the previous 30 minutes and exceeded 1,380 won, and the nature of the fluctuation is evaluated as a short-term shock."

[0072] In addition, context notification content according to one embodiment of the present invention can be generated at an expression level corresponding to the user's level of expertise.

[0073] Information regarding the level of expertise of a user according to one embodiment of the present invention may be determined based on at least one of the following: the user's job group, whether the user possesses foreign exchange-related qualifications, foreign exchange transaction history, and information regarding a level of expression pre-set by the user. For example, if the user's job group corresponds to a professional job group such as a foreign exchange dealer, a fund manager, or a treasurer, the user's level of expertise may be determined to be relatively high, and if the user's job group corresponds to a general employee of a small or medium-sized enterprise, a trade manager, or an individual user, the user's level of expertise may be determined to be relatively low.

[0074] A context notification content generation unit (230) according to one embodiment of the present invention can generate context notification content having different expression levels depending on the user's level of expertise, even for the same context vector.

[0075] For example, when the user's level of expertise is relatively high, contextual notification content at the level of expression using specialized foreign exchange terminology may be generated, such as, "Due to the FOMC freeze, the interest rate differential between Korea and the U.S. has widened, causing the 1-month NDF spread to shift by +0.5 won in the direction of increasing bearish pressure, acting as short-term upward pressure on USD / KRW." As another example, when the user's level of expertise is relatively low, contextual notification content at the level of expression using plain everyday language may be generated for the same context vector, such as, "As the U.S. maintains its base interest rate, demand for the U.S. dollar is increasing, causing the value of the dollar against the won to rise in the short term."

[0076] By doing so, the context notification content generation unit (230) according to one embodiment of the present invention can generate context notification content so that even for the same context, experts and non-expert users can accept the context according to their respective levels of understanding.

[0077] In addition, the context notification content generation unit (230) according to one embodiment of the present invention can generate draft content based on a template and generate additional explanatory content based on a generation model in response to the satisfaction of a condition regarding the urgency of the notification.

[0078] A condition regarding the urgency of a notification according to one embodiment of the present invention may be satisfied when the fluctuation range of an exchange rate fluctuation event exceeds a threshold value, when the exchange rate fluctuation event occurs in temporal proximity to an event with a significant market impact (e.g., announcement of policy decisions by central banks of major countries, large-scale geopolitical events, etc.), or when an exchange rate fluctuation event occurs in a currency pair for which the user has previously set an urgent notification. However, the types of conditions regarding the urgency of a notification are not limited thereto and may be varied within the scope of achieving the purpose of the present invention.

[0079] Draft content generated based on a template according to one embodiment of the present invention may be short-form content generated quickly by filling in key information such as the direction of fluctuation, the magnitude of fluctuation, and currency pairs of an exchange rate fluctuation event into predefined phrases. For example, the draft content may be generated in a form that quickly conveys only the facts and key figures of an exchange rate fluctuation, such as "[USD / KRW] exceeded 1,380 won, increased by 15 won in the previous 30 minutes. Detailed explanation to be sent."

[0080] Additional explanatory content generated based on a generative model according to one embodiment of the present invention may be content in a more detailed form generated by inputting a context vector into the generative model. For example, the additional explanatory content may be generated in a form that comprehensively describes the external causes of the fluctuation, changes in internal aspects, and an assessment regarding the persistence of the fluctuation, such as, "The recent rise in USD / KRW is primarily due to the US FOMC's decision to freeze the base interest rate announced during the previous 30 minutes and the subsequent widening of the interest rate differential between Korea and the US, and it is assessed that the spread of the 1-month NDF widened by +0.5 won during the same time period, reflecting increased bearish pressure in the market. The nature of the fluctuation is assessed as a short-term shock, and whether there is a trend reversal needs to be determined through additional observation over the next 24 hours."

[0081] In addition, the context notification content generation unit (230) according to one embodiment of the present invention can generate additional explanation content with an expression level corresponding to the user's level of expertise when generating additional explanation content.

[0082] In addition, the context notification content generation unit (230) according to one embodiment of the present invention may further perform the function of verifying whether the generated context notification content has consistency with the information included in the context vector.

[0083] Specifically, a context notification content generation unit (230) according to one embodiment of the present invention can extract factual claims included in the generated context notification content (e.g., numerical value of an exchange rate, fluctuation range, timing of occurrence of an external event, announced value of an economic indicator, numerical value of interest rate difference between currency pairs, etc.) and verify whether the extracted factual claims are consistent with source information constituting a context vector.

[0084] For example, the context notification content generation unit (230) may verify consistency by using natural language processing techniques to extract parts corresponding to numerical facts or temporal facts within the context notification content and checking whether the extracted parts match the source numerical values ​​or source temporal points included in the context vector within a preset tolerance range. As another example, the context notification content generation unit (230) may verify consistency by using a separately trained consistency verification model to evaluate semantic consistency between the context notification content and the context vector.

[0085] A context notification content generation unit (230) according to one embodiment of the present invention may perform follow-up measures, such as (i) regenerating the context notification content, (ii) correcting the part with consistency issues and regenerating it, or (iii) withholding the sending of the context notification content to the device (300) and notifying the administrator, when it is determined that the consistency verification result does not satisfy a preset standard. However, the method of consistency verification and the follow-up measures based on the verification result are not limited thereto and may be varied within the scope of achieving the purpose of the present invention.

[0086] According to the consistency verification function according to one embodiment of the present invention, cases where information different from the facts is included in context notification content due to hallucination phenomena that may occur in the generation model can be prevented, thereby enabling reliable context notification content to be provided to the user.

[0087] Meanwhile, context notification content generated by the context notification content generation unit (230) according to one embodiment of the present invention may be utilized as background explanation for information regarding currency hedging recommendations by a currency hedging decision support system (not shown). For example, when the currency hedging decision support system recommends a currency hedging ratio for a specific scheduled payment date to a user, the context notification content generated by the context notification content generation unit (230) according to one embodiment of the present invention may be provided together as background explanation for information regarding the recommendation of the currency hedging ratio. By doing so, even a user with relatively little expertise in currency hedging can participate in the decision-making regarding the currency hedging recommendation while contextually understanding the market conditions that serve as the background for the currency hedging recommendation.

[0088] Additionally, the context notification content generating unit (230) according to one embodiment of the present invention may also perform the function of sending the generated context notification content to the device (300). For example, the context notification content generating unit (230) may send the context notification content to the device (300) through at least one of various methods, such as a push notification, a messaging service, email, an instant messenger, a webhook, or an API callback.

[0089] Additionally, the context notification content generation unit (230) according to one embodiment of the present invention may, in response to the satisfaction of a condition regarding the urgency of the notification, first send draft content generated based on a template to the device (300), and subsequently send additional explanatory content generated based on a generation model to the device (300).

[0090] For example, the context notification content generation unit (230) may send draft content to the device (300) within seconds after an exchange rate fluctuation event is detected in response to the condition regarding the urgency of the notification being satisfied, and subsequently send additional explanatory content generated by the generation model to the device (300) within tens of seconds or minutes. As another example, the context notification content generation unit (230) may send the draft content and the additional explanatory content together under the same notification identifier so that the draft content can be naturally updated and expanded into the additional explanatory content and displayed on the device (300).

[0091] In addition, the context notification content generating unit (230) according to one embodiment of the present invention may send context notification content to a device (300) by combining numerical information regarding an exchange rate fluctuation event with the context notification content. For example, the context notification content generating unit (230) may send to the device (300) a combined form of numerical information of the exchange rate (currency pair 1,382.5 won, +15 won) and context notification content (an explanation that the widening of the interest rate difference between Korea and the US due to the US FOMC's freezing of the base rate) such as "[USD / KRW 1,382.5 won, +15 won] The main cause of this rise in the exchange rate is the widening of the interest rate gap between Korea and the US following the US FOMC's freezing of the base rate." By doing so, the user can simultaneously receive numerical information of the exchange rate fluctuation and the context information within the same notification.

[0092] In addition, the context notification content generation unit (230) according to one embodiment of the present invention may transmit the context notification content to external systems, such as an enterprise resource planning (ERP) system, a trade management platform, an accounting system, a fund management system, and a foreign exchange hedging decision support system, in addition to the device (300), in an API integration manner, thereby enabling the external systems to utilize the context notification content according to one embodiment of the present invention in a manner that is suitable for them.

[0093] Next, a communication unit (240) according to one embodiment of the present invention can perform the function of enabling data transmission and reception from / to an exchange rate fluctuation detection unit (210), a complex context collection unit (220), and a context notification content generation unit (230).

[0094] Finally, a control unit (250) according to one embodiment of the present invention can perform the function of controlling the flow of data between an exchange rate fluctuation detection unit (210), a complex context collection unit (220), a context notification content generation unit (230), and a communication unit (240). That is, by controlling the flow of data from / to / from the outside of the system (200) according to the present invention or the flow of data between each component of the system (200) according to the present invention, the control unit (250) can control the exchange rate fluctuation detection unit (210), the complex context collection unit (220), the context notification content generation unit (230), and the communication unit (240) to perform their respective unique functions.

[0095] The embodiments according to the present invention described above may be implemented in the form of program instructions that can be executed through various computer components and recorded on a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc., either individually or in combination. The program instructions recorded on the computer-readable recording medium may be those specifically designed and configured for the present invention or those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. Hardware devices may be modified into one or more software modules to perform processing according to the present invention, and vice versa.

[0096] Although the present invention has been described above with reference to specific details such as specific components, limited embodiments, and drawings, this is provided only to aid in a more comprehensive understanding of the invention, and the invention is not limited to the above embodiments, and a person skilled in the art to which the invention belongs can make various modifications and changes from this description.

[0097] Accordingly, the scope of the present invention should not be limited to the embodiments described above, and all scopes equivalent to or equivalently modified from the claims set forth below, as well as the claims set forth below, shall be considered to fall within the scope of the concept of the present invention. Explanation of the symbols

[0098] 100: Communication network 200: Notification generation system 210: Exchange rate fluctuation detection unit 220: Complex Context Collection Unit 230: Context Notification Content Generation Section 240: Communications Department 250: Control unit 300: Device

Claims

Claim 1 A method implemented in an alert generation system for generating foreign exchange fluctuation context alerts, wherein the system comprises an exchange rate fluctuation detection unit, a complex context collection unit, and a context alert content generation unit, and wherein the method comprises the steps of: the exchange rate fluctuation detection unit detecting an exchange rate fluctuation event based on fluctuation conditions; the complex context collection unit collecting heterogeneous information that occurred during a time interval preceding the occurrence of the exchange rate fluctuation event in response to the detection of the exchange rate fluctuation event, and combining the collected heterogeneous information to generate a context vector; and the context alert content generation unit inputting the generated context vector into a generation model learned based on data regarding the foreign exchange field to generate context alert content regarding the exchange rate fluctuation event, wherein the heterogeneous information includes a combination of information regarding an event that occurred outside the foreign exchange market and information regarding aspects observed inside the foreign exchange market, and wherein the context vector is a vector structured to reflect the temporal causal relationship between an event that occurred outside the foreign exchange market and aspects observed inside the foreign exchange market, and wherein, in the context alert content generation step, the complex context collection unit, information regarding an event that occurred outside the foreign exchange market and inside the foreign exchange market A method for generating a context vector in which information regarding observed aspects is aligned and combined on the same time axis, and in the context notification content generation step, the context notification content generation unit generates the context notification content in a form including a causal explanation associated with the temporal causal relationship. Claim 2 delete Claim 3 In claim 1, a method in which, in the context notification content generation step, the context notification content generation unit generates the context notification content at an expression level corresponding to the user's level of expertise. Claim 4 A method according to claim 1, wherein the context notification content generating unit further includes the step of sending the generated context notification content to a device, wherein in the sending step, the context notification content generating unit first sends a draft content generated based on a template in response to the satisfaction of a condition regarding the urgency of the notification, and subsequently sends additional explanatory content generated based on the generation model. Claim 5 A non-transient computer-readable recording medium for recording a computer program for executing the method according to paragraph 1. Claim 6 A system for generating foreign exchange fluctuation context notifications, comprising: a foreign exchange fluctuation detection unit that detects an exchange rate fluctuation event based on fluctuation conditions; a complex context collection unit that, in response to the detection of the foreign exchange fluctuation event, collects heterogeneous information that occurred during a time interval preceding the time of occurrence of the foreign exchange fluctuation event and combines the collected heterogeneous information to generate a context vector; and a context notification content generation unit that inputs the generated context vector into a generative model learned based on data regarding the foreign exchange field to generate context notification content regarding the foreign exchange fluctuation event, wherein the heterogeneous information includes a combination of information regarding an event that occurred outside the foreign exchange market and information regarding aspects observed inside the foreign exchange market, and the context vector is a vector structured to reflect the temporal causal relationship between the event that occurred outside the foreign exchange market and the aspects observed inside the foreign exchange market, wherein the complex context collection unit generates the context vector in which information regarding the event that occurred outside the foreign exchange market and information regarding aspects observed inside the foreign exchange market are aligned on the same time axis and combined, and the context notification content generation unit generates the context notification content in a form that includes a causal explanation associated with the temporal causal relationship. Claim 7 delete Claim 8 In paragraph 6, the context notification content generation unit is a system that generates the context notification content at an expression level corresponding to the user's level of expertise. Claim 9 In claim 6, the context notification content generation unit sends the generated context notification content to a device, wherein, in response to the satisfaction of a condition regarding the urgency of the notification, it first sends a draft content generated based on a template and subsequently sends additional explanatory content generated based on the generation model.

Citation Information

Patent Citations

  • System and method for trading foreign currency asset

    KR1020190013095A

  • Method for provide enterprise stock forecasting algorithm based on situation cognition

    KR1020200053005A

  • Information Platform System on Foreign Exchange Rate and its Information Processing Method

    KR1020160053624A

  • Method for providing exchage rate prediction system based on multi aritificial intelligence models

    KR1020250032378A

  • Transformer-based electronic disclosure document service platform apparatus and method

    KR1020260063079A