WO2023139130A1 - Computer-implementierte datenstruktur, verfahren und system zum betrieb eines technischen geräts mit einem modell auf basis föderierten lernens - Google Patents
Computer-implementierte datenstruktur, verfahren und system zum betrieb eines technischen geräts mit einem modell auf basis föderierten lernens Download PDFInfo
- Publication number
- WO2023139130A1 WO2023139130A1 PCT/EP2023/051145 EP2023051145W WO2023139130A1 WO 2023139130 A1 WO2023139130 A1 WO 2023139130A1 EP 2023051145 W EP2023051145 W EP 2023051145W WO 2023139130 A1 WO2023139130 A1 WO 2023139130A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- client
- factor
- server
- model
- computer
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
Definitions
- the invention relates to a computer-implemented data structure and a use of the data structure by a client at the edge of a client-server system for operating a technical device connected to a client at the edge with a model based on federated learning.
- the invention also relates to a computer-implemented method and a system for operating a technical device with a model based on federated learning.
- the invention also relates to a computer program, an electronically readable data carrier and a data carrier signal.
- FL Fusion Learning
- Federated learning enables multiple actors to build a common, robust machine learning model without exchanging data, thereby addressing critical issues such as privacy, data security, data access rights, and access to heterogeneous data.
- contributions from individual clients i.e. weights of locally trained models, are not always treated equally, but are weighted in an aggregation process.
- client-server interactions data communications between a client and a server are referred to as client-server interactions, or interactions for short.
- the data can be user data, such as sensor data, or control data for controlling a client or a device connected to a client.
- the operating properties should not only take into account the initial operation, but also current operation and should therefore also include dynamic adjustments of the model for operating the device.
- the object of the invention is achieved by a data structure of the type mentioned, comprising a trust factor for the device, which has a reliability factor, a response time factor and an information quality factor to take into account a data quality of the model for the data of the model, and the trust factor (T) is designed as a model in the form of a Bayesian network, the trust factor being intended to weight the model of the technical device by the client on the fly operation of the system.
- trust can be defined as an agent's belief in attributes such as response time, reliability, and competence in terms of the quality of information provided by a trusted agent.
- a na ⁇ ve Bayesian classifier approach is envisaged to model trust in a loT edge-based network.
- a Bayesian network is a relational network that uses statistical methods to represent probabilistic relationships between different items, specifically conditional probabilities.
- a Bayesian network or Bayesian network is a directed acyclic graph in which the nodes describe random variables and the edges describe conditional dependencies between the variables.
- Each node of the network is assigned a conditional probability distribution of the random variable it represents, given the random variables at the parent nodes. They are described by probability tables. This distribution can be arbitrary, but discrete or normal distributions are often used.
- Parents of a vertex v are those vertices from which an edge leads to v.
- a Bayesian network serves to represent the common probability distribution of all variables involved as compactly as possible using known conditional independences. In doing so, the conditional (in)dependence of subsets of the variables is combined with a priori knowledge.
- a data structure based on a Bayesian network is therefore particularly favorable in order to store probabilities as compactly as possible. This is particularly favorable when the data is transmitted or stored in a data network, such as between client and server.
- a naive Bayesian network is a simple Bayesian network, i.e. a directed graph representing conditional dependencies. It consists of a root node and several leaf nodes.
- a naive Bayesian network is used to favorably represent the trust between two nodes, such as between a server and a client.
- trust in loT node Three trust aspects are taken into account in a trust factor (“trust in loT node”), namely reliability, response time and information quality.
- the term "factor” is used in the present context for the sake of simplicity, although the trust factor is a model in the form of a Bayesian network and is therefore not a scalar in the present context. Alternatively, the term “trust factor” can also be used equivalently to the term “trust model”.
- the trust factor describes a server's trust in a client and can change dynamically during the operation of the client-server system.
- the confidence factor describes the confidence for the device in the form of a numerical, weighting factor.
- the trust factor can be calculated by a corresponding computing device in a client or in the server, with calculation in the server being advantageous since the same included computing device can be used for a number of clients and the system is therefore simpler overall.
- a client's trust factor can change over time, for example if the data connection deteriorates during system operation and the information quality factor is adversely affected as a result.
- a client can also take active measures, for example to improve the response time after the Client himself has found that trust a
- Reliability is the ability of a system or component to function under specified conditions for a specified period of time. It is important for an accurate ML model that IoT nodes provide services that are available and can be relied on. This includes, for example, the availability of a client to join a federation.
- the reliability factor describes the reliability of a client towards a server, based on the network communication behavior between client and server.
- the reliability factor describes the reliability of the device in the form of a numerical, weighting factor.
- the reliability factor is advantageously calculated in the server, since the same included computing device in the server can be used for a number of clients and the system is therefore simpler overall.
- Response time is the time difference between requesting information and receiving a response to the request. It is important for an accurate ML model that loT nodes respond to requests in a timely manner.
- the reaction time factor describes the temporal behavior of a client towards a server, based on the network communication behavior between client and server.
- the response time factor describes the behavior of the device over time in the form of a numerical, weighting factor.
- the calculation of the reaction time factor is advantageously carried out on the server, since the same calculation Device can be used in the server and the system is therefore simpler overall.
- Information quality is understood to mean, for example, the accuracy, completeness, consistency, topicality, validity and uniqueness of the information. It is important for an accurate ML model that the models are trained using high quality data.
- the information quality factor describes the information quality of a client's data delivered to a server.
- the information quality factor describes the information quality of data supplied by the device in the form of a numerical, weighting factor.
- the information quality factor is advantageously calculated in the server, with the metrics used for this being determined on the basis of raw data in the client.
- These metrics can be transmitted to the server, which then calculates the information quality factor using an included computing device.
- a model based on this advantageous approach in the form of a computer-implemented data structure is very simple and easy to implement, as well as being quick and particularly compact in terms of low memory requirements and favorable access properties. This allows it to be implemented on various embedded, resource-constrained devices in particular.
- a model based on this approach is highly scalable as it scales linearly with the number of predictors and data points, which is an important property to support computations on IoT nodes.
- the weighting of the model of the device is dynamically re-determined over time during operation of the system.
- the data structure according to the invention is computer-implemented in a client of the client-server system.
- conditional probabilities can be stored which characterize interactions of a client-server system.
- Interactions can be recorded, for example, by data loggers.
- the content of a data transmission can be recorded, but transmission properties can also be determined, such as latency, transmission error rate, transmission bandwidth, time stamp information of data packets, data loss rates, number of transmission repetitions, etc.
- the interactions can take place both in the server and in a
- Clients are recorded and evaluated, as previously explained, for example, using a data logger or fe Simple statistical evaluations of data transmissions carried out between one or more clients and the server.
- the client has the opportunity to derive measures for the client himself from the recorded and evaluated interactions, for example to take measures that directly improve the parameters and thus be able to offer "better" interactions to the server.
- this trust factor can be transmitted to the relevant client so that the client is informed about the status of his trust factor and can take appropriate measures to change it.
- the trust factor can be transmitted to a client periodically or, for example, be initiated by a change above a predefined threshold value.
- the reliability factor is based on the availability of the technical device when the system is in operation.
- reaction time factor is based on the time difference between requesting information and receiving an answer to the request for the device while the system is operating.
- the information quality factor is based at least on the accuracy, completeness, consistency, topicality, validity or uniqueness of the information for the technical device in the operation of the system.
- the object according to the invention is achieved by using the data structure for a computer-implemented method for operating a technical device by a client of a client-server system with a model based on federated learning.
- the internal processes of the FL client-server architecture can be improved by the data structure, in that favorable system properties are increasingly used and unfavorable system properties are accordingly suppressed.
- the object of the invention is also achieved by a computer-implemented method for operating a technical device, which is connected to a client of a client-server system, the client-server system comprising a server and connected clients and the clients interacting at least temporarily with the server, which is recorded as interactions, with a model based on federated learning, the following steps being carried out: System is based, and which is determined from the interactions recorded in step a), b) detecting a reaction time factor, which is based on the time difference between requesting information and receiving an answer to the request for the Device in the operation of the system is based, and which is determined from the interactions recorded in step a), c) recording an information quality factor, which is based at least on the accuracy, completeness, consistency, timeliness, validity or uniqueness of the information for the technical device in the operation of the system, and which is determined from the recorded interactions in step a), d) determining a respective trust factor, which is a model in the form of a Bayes is formed, for the client,
- the respective trust factor during operation of the System is determined again and chronologically younger recorded data with regard to the reliability factor, the reaction time factor and the information quality factor are considered more weighted than chronologically older.
- the object according to the invention is achieved by a system of the type mentioned at the beginning, comprising a server and connected clients, which are each connected to technical devices, and the system is set up to carry out the method according to the invention.
- the object according to the invention is achieved by a computer program comprising instructions which, when executed by a computer, cause the computer to carry out the method according to the invention on a system according to the invention.
- the object according to the invention is achieved by an electronically readable data carrier with readable control information stored thereon, which comprises at least the computer program according to the invention and is designed in such a way that when the data carrier is used in a computing device, they carry out the method according to the invention.
- the object according to the invention is achieved by a data carrier signal which transmits the computer program according to the invention.
- Fig. 1 a prior art FL loop
- 2 shows an example of a trust factor data structure according to the invention
- FIG. 6 shows an example of federated averaging in the form of a pseudo code.
- Clients ED1-ED3, EDK with connected devices FE1-FE3, FEK (“front end”) have corresponding weights within respective local models MLM1-MLM3, MLMK, which are transmitted to a server S1.
- the server S1 performs an aggregation of those models MLM1-MLM3, MLMK received from the clients ED1-ED3, EDK with a federated averaging algorithm and creates a global model MLMG with appropriate weights.
- the server S1 distributes the resulting averaged weights back to the clients ED1-ED3, EDK.
- the clients can now control or operate the connected technical devices ED1-ED3, EDK using the new ML model, for example to carry out predictive maintenance on the device.
- This loop is repeated for N rounds of communication.
- Contributions of individual clients i.e. weights of locally trained models, are not treated equally, but are weighted in an aggregation process.
- This aspect ⁇ Ixt is shown in the figure, where n is the number of data samples and n k is the number of data samples of client k.
- Fig. 2 graphically represents an example of a trust factor data structure according to the invention, which is applied in a FL model.
- a trust factor T is determined by a reliability factor R, a response time factor RT and an information quality factor Qol.
- Fig. 3 shows an example of a device according to the invention with a server S and clients C1-C3, which are set up to interact with the respective technical devices D1-D3.
- Each client C1-C3 also referred to as an edge device, has a computer-implemented data structure in the form of a confidence factor T1-T3 according to the previous figure, and of course a respective local trained ML model for operating the respective connected device D1-D3.
- the server S has an aggregated global ML model GM, which is provided to the clients C1-C3.
- the trust factors T1-T3 are determined by the server from interactions, such as data transactions between the server S and clients C1-C3, and are stored in the server.
- the trust factors T1-T3 can also be stored in the clients C1-C3, with the trust factors T1-T3 each being transmitted with the trained local models M1-M3 in order to then be aggregated in the server S to form the global model GM.
- the trust factor TI and the local model M1 and the technical device D1 are therefore assigned to the client CI.
- the trust factor T2 and the local model M2 and the technical device D2 are assigned to the client C2.
- the trust factor T3 and the local model M3 and the technical device D3 are assigned to the client C3.
- Fig. 4 shows an exemplary embodiment of the method according to the invention for the operation of a technical device D1-D3, which is connected to a client C1-C3 of a client-server system, the client-server system comprising a server S and connected clients C1-C3 and the clients C1-C3 interacting at least temporarily with the server S, which is recorded as interactions, with a model based on federated learning, in which the following steps are carried out : a) detection of a reliability factor R, which is based on the availability of the technical device D1-D3 in the operation of the system, and which is determined from the detected interactions, by the client C1-C3, b) detection of a reaction time factor RT, which is based on the time difference between requesting information and receiving an answer to the request for the device in the operation of the system, and which is determined from the detected interactions, by the client C1 -C3, c) detection of an information quality factor Qol, which is based at least on the accuracy, completeness, consistency, timeline
- the information quality factor Qol can be determined using further metrics CRT (“conditional probability table”), which includes conditional probabilities in the form of a table, with each leaf node being linked via the metric elements in the form of a table.
- CRT conditional probability table
- steps a) to d) can also be carried out independently of steps e) and f) in terms of time, ie also in parallel, as shown.
- the interaction of a client with the server is understood as communication or data transport, for example, to communicate data from a terminal connected to a client in the form of a sensor to the server.
- a client can also be viewed as a root node.
- Each trust factor T1-T3 can have two values as the root node, namely “1” for “satisfactory” and “0” for “unsatisfactory”.
- a value P(T) represents the server's total trust in a client, i.e. an IoT node, in the ability to provide actionable knowledge in the form of trained weights.
- the value is determined by the ratio of "satisfactory" interactions to the total number of interactions.
- the node of the information quality factor Qol represents a number of different data quality matrices DQM (“Data Quality Metrics”).
- the information quality factor Qol includes three metric values:
- CPL completeness refers to the completeness of data with regard to all values as a part and as a whole. Accordingly, there should be no data gaps and none
- Timeliness TLN Data should be up-to-date at the intended retrieval time and be available through the server, and corresponding access should also be possible.
- Validity refers to how accurately a method measures what it is supposed to measure. This means that it produces results that correspond to real-world properties, properties and variations in the physical world.
- the data quality matrices DQM are determined in the client from raw data from an IoT device connected to the client, transmitted to the server and can, for example, form the information quality factor Qol by summation, which is calculated in the server.
- the information quality factor Qol thus itself comprises a Bayesian network through the data quality matrices DQM.
- Table 1 shows examples of data quality matrices DQM.
- 5 shows an example for calculating the confidence factor in the form of a pseudo code.
- the result of this calculation provides a trust value for an IoT node between 0 and 1 .
- Each node has a threshold value s_t for a "satisfactory" interaction.
- Fig. 6 shows an example of federated averaging in the form of a pseudo-code "FederatedAveraging".
- This exemplary embodiment shows an advantageous extension of the approach, with a possibility for "aging" of the trust factor being provided.
- This property is particularly relevant and advantageous for a dynamic loT environment.
- the sensor measuring device connected to an IoT node can be renewed and improved and thus provide more accurate estimates, which leads to a favorable influence on the information quality factor Qol.
- the ability of individual IoT nodes to provide valuable information may well change, depending on environmental factors.
- This aging can be applied to the metrics for the information quality factor Qol, the response time factor RT and the reliability factor R either individually or all together, for example by applying a higher weighting to more recent data and a lower weighting to older data.
- the set of favorable interactions between the server S and a client C1-C3 in the same time window t is characterized by the ratio num_satisf ying/num_interactions.
- An aging factor is between 0 and 1, with younger interactions or their degree of "satisfaction" being weighted higher than older interactions.
- an adaptive weighting scheme AWS is used in the federated averaging algorithm “FederatedAveraging”.
- the weight term weights w ⁇ +1 is marked.
- the set of weights w ⁇ +1 of a model of a client k is multiplied by the trust weight w T , where the trust weight w T can be determined by the computeTrust ( ) function shown in the previous figure.
- MLM1-MLM3 MLMK, M1-M3 ML model
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Algebra (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computer And Data Communications (AREA)
Abstract
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202380017781.2A CN118575182A (zh) | 2022-01-18 | 2023-01-18 | 用于利用基于联邦学习的模型来运行技术设备的计算机实现的数据结构、方法和系统 |
US18/729,354 US20250094843A1 (en) | 2022-01-18 | 2023-01-18 | Computer-Implemented Data Structure, Method and System for Operating a Technical Device with a Model Based on Federated Learning |
EP23703702.3A EP4430537A1 (de) | 2022-01-18 | 2023-01-18 | Computer-implementierte datenstruktur, verfahren und system zum betrieb eines technischen geräts mit einem modell auf basis föderierten lernens |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EPPCT/EP2022/051043 | 2022-01-18 | ||
PCT/EP2022/051043 WO2023138756A1 (de) | 2022-01-18 | 2022-01-18 | Computer-implementierte datenstruktur, verfahren und system zum betrieb eines technischen geräts mit einem modell auf basis föderierten lernens |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2023139130A1 true WO2023139130A1 (de) | 2023-07-27 |
Family
ID=80682710
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/EP2022/051043 WO2023138756A1 (de) | 2022-01-18 | 2022-01-18 | Computer-implementierte datenstruktur, verfahren und system zum betrieb eines technischen geräts mit einem modell auf basis föderierten lernens |
PCT/EP2023/051145 WO2023139130A1 (de) | 2022-01-18 | 2023-01-18 | Computer-implementierte datenstruktur, verfahren und system zum betrieb eines technischen geräts mit einem modell auf basis föderierten lernens |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/EP2022/051043 WO2023138756A1 (de) | 2022-01-18 | 2022-01-18 | Computer-implementierte datenstruktur, verfahren und system zum betrieb eines technischen geräts mit einem modell auf basis föderierten lernens |
Country Status (4)
Country | Link |
---|---|
US (1) | US20250094843A1 (de) |
EP (1) | EP4430537A1 (de) |
CN (1) | CN118575182A (de) |
WO (2) | WO2023138756A1 (de) |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021118452A1 (en) * | 2019-12-10 | 2021-06-17 | Agency For Science, Technology And Research | Method and server for federated machine learning |
-
2022
- 2022-01-18 WO PCT/EP2022/051043 patent/WO2023138756A1/de active Application Filing
-
2023
- 2023-01-18 WO PCT/EP2023/051145 patent/WO2023139130A1/de not_active Application Discontinuation
- 2023-01-18 CN CN202380017781.2A patent/CN118575182A/zh active Pending
- 2023-01-18 US US18/729,354 patent/US20250094843A1/en active Pending
- 2023-01-18 EP EP23703702.3A patent/EP4430537A1/de not_active Withdrawn
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021118452A1 (en) * | 2019-12-10 | 2021-06-17 | Agency For Science, Technology And Research | Method and server for federated machine learning |
Non-Patent Citations (5)
Title |
---|
AMIT PORTNOY ET AL: "Towards Federated Learning With Byzantine-Robust Client Weighting", ARXIV.ORG, CORNELL UNIVERSITY LIBRARY, 201 OLIN LIBRARY CORNELL UNIVERSITY ITHACA, NY 14853, 18 May 2021 (2021-05-18), XP081949237 * |
KANG JIAWEN ET AL: "Reliable Federated Learning for Mobile Networks", IEEE WIRELESS COMMUNICATIONS, 14 October 2019 (2019-10-14), US, pages 1 - 8, XP055835834, ISSN: 1536-1284, DOI: 10.1109/MWC.001.1900119 * |
SHENGHUI LI ET AL: "Auto-weighted Robust Federated Learning with Corrupted Data Sources", ARXIV.ORG, CORNELL UNIVERSITY LIBRARY, 201 OLIN LIBRARY CORNELL UNIVERSITY ITHACA, NY 14853, 31 March 2021 (2021-03-31), XP081901753 * |
THOMAS HIESSL ET AL.: "Industrial Federated Learning - Requirements and System Design", 14 May 2020, CORNELL UNIVERSITY LIBRARY |
THOMAS HIESSL ET AL: "Industrial Federated Learning -- Requirements and System Design", ARXIV.ORG, CORNELL UNIVERSITY LIBRARY, 201 OLIN LIBRARY CORNELL UNIVERSITY ITHACA, NY 14853, 14 May 2020 (2020-05-14), XP081673589 * |
Also Published As
Publication number | Publication date |
---|---|
EP4430537A1 (de) | 2024-09-18 |
CN118575182A (zh) | 2024-08-30 |
WO2023138756A1 (de) | 2023-07-27 |
US20250094843A1 (en) | 2025-03-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
DE202017007517U1 (de) | Aggregatmerkmale für maschinelles Lernen | |
DE69324052T2 (de) | Neuronalnetzwerk-Lernsystem | |
DE112021004908T5 (de) | Computerbasierte systeme, rechenkomponenten und rechenobjekte, die eingerichtet sind, dynamische ausreisser-verzerrungs-verringerung bei maschinenlernmodellen zu implementieren | |
DE102021004591A1 (de) | Für Graphen vorgesehene neuronale Netzwerke für Datensätze mit Heterophilie | |
DE102017125256A1 (de) | Suche nach einer neuronalen Architektur | |
DE112021003908T5 (de) | Föderales maschinenlernen durch verwenden von ortsabhängigem hashing | |
DE112019006156T5 (de) | Erkennung und behandlung von unsachgemässen eingaben durch neuronale netzwerke | |
DE112021004174T5 (de) | Föderiertes lernen zur anomalieerkennung | |
DE102019007340A1 (de) | Technik zum Einrichten und Betreiben eines neuronalen Netzwerks | |
EP3901713B1 (de) | Verfahren und system zum betrieb einer technischen anlage mit einem optimalen modell | |
DE102022132540A1 (de) | Adaptive modellbereinigung zum verbessern der leistung von föderalem lernen | |
DE112016007411T5 (de) | Fuzzy-eingabe für autoencoder | |
DE112015005501T5 (de) | Alterungsprofilbildungsmaschine für physikalische Systeme | |
DE102021000450A1 (de) | Empfehlungssystem auf Grundlage einer Trendsetter-Inferenz | |
WO2020126168A1 (de) | Verfahren zur zusammenarbeit von mehreren geräten eines lokalen netzwerks | |
EP4430537A1 (de) | Computer-implementierte datenstruktur, verfahren und system zum betrieb eines technischen geräts mit einem modell auf basis föderierten lernens | |
DE112021000619T5 (de) | Adaptive zustandsverwaltung für statusunabhängige services | |
DE10252445A1 (de) | Verfahren und Computer-Anordnung zum Bereitstellen von Datenbankinformation einer ersten Datenbank und Verfahren zum rechnergestützten Bilden eines statistischen Abbildes einer Datenbank | |
WO2001065421A1 (de) | Verfahren und anordnung zur modellierung eines systems | |
DE102022131760A1 (de) | Modellgenerierungsverfahren, modellgenerierungsprogramm, modellgenerierungsvorrichtung und datenverarbeitungsvorrichtung | |
EP4399646A1 (de) | Computer-implementiertes verfahren und system zur anomalie-erkennung in sensordaten | |
EP4405869A1 (de) | Computer-implementiertes verfahren zum betrieb eines technischen geräts mit einem modell | |
DE102024212013A1 (de) | Verfahren und vorrichtung zur modellaggregation im föderierten lernen | |
EP4513391A1 (de) | Computer-implementiertes verfahren und system zum betrieb eines technischen geräts | |
EP3507943A1 (de) | Verfahren zur kommunikation in einem kommunikationsnetzwerk |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23703702 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2023703702 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 2023703702 Country of ref document: EP Effective date: 20240611 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 18729354 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202380017781.2 Country of ref document: CN |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: 2023703702 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 18729354 Country of ref document: US |