A digital life-based memory directional migration and clone inheritance system

Abstract

The application discloses a memory directional migration and clone inheritance system based on digital life, and belongs to the technical field of artificial intelligence digital life inheritance. The system realizes memory three-class directional screening, binding and solidification of a benchmark model, asymmetric encryption authorization, cross-device compatible vector alignment memory loading, and completes safe and controllable digital clone inheritance through the execution of preset steps by a processor. The application solves the technical problems of easy privacy leakage, memory full copy bloating, personality distortion after inheritance, lack of specified authorization and cross-device inheritance difficulty in the existing digital life migration, can accurately inherit the digital clone to the specified heir under the premise of ensuring the safety of user privacy, has simple overall deployment, low cost, and is suitable for ordinary users to realize the permanent continuation of personal digital life.

Description

Technical Field

[0002] This invention relates to the fields of artificial intelligence digital life inheritance, memory vector directional migration, and cross-terminal digital clone inheritance technology. It is particularly applicable to the complete inheritance of digital clones after a user's death, the directional screening and migration of core memories, and the continuation of digital life across devices, and belongs to the category of digital life asset inheritance and AI memory migration technology. Background Technology

[0004] As digital human and digital life technologies mature, more and more users wish to leave their digital avatars as digital legacies to designated heirs, achieving the permanent continuation of their personal thoughts, memories, and personality traits. Existing digital human systems have significant shortcomings in memory transfer and avatar inheritance, failing to meet the needs of digital life inheritance in real-world scenarios.

[0005] Most current digital human migration solutions use complete data replication, which transfers a large amount of irrelevant chatter, temporary memories, and private content to the successor. This poses a serious risk of privacy leaks and results in bloated content and low operating efficiency in the inherited digital clone, which does not conform to the user's true intention of "only transferring core life memories".

[0006] Meanwhile, existing technologies lack targeted screening and migration mechanisms, failing to differentiate between inheritable memories, private memories, and prohibited inheritable memories according to user-preset rules. Furthermore, there is no dedicated inheritance authorization verification process, meaning anyone with access to the data can use it directly, making it impossible to guarantee that digital life is only passed on to designated heirs. Additionally, existing migration schemes do not align with a fixed model benchmark, easily leading to the loss of original personality traits after migration, resulting in distorted personalities in the inherited digital avatar and rendering the inheritance meaningless.

[0007] To address the shortcomings of existing technologies, this invention proposes a memory-oriented migration and clone inheritance system based on digital life, which solves the problems of existing solutions being unable to selectively filter memories, lacking inheritance authorization, having a high risk of privacy leakage, and causing personality distortion after migration. Summary of the Invention

[0009] The purpose of this invention is to overcome the shortcomings of existing digital life migration schemes, such as the inability to selectively filter memories, lack of inheritance authorization, high risk of privacy leakage, and easy distortion of personality after migration, and to provide a memory-oriented migration and clone inheritance system based on digital life.

[0010] This system relies on the local processor to achieve memory layering and filtering, inheritance permission encryption verification, targeted vector migration, and benchmark model linkage recovery. It strictly follows the user-preset rules to complete the core memory migration, and at the same time connects with the previously solidified benchmark model to ensure that the inherited digital clone is completely consistent with the original personality, thus achieving safe, controllable, and authentic digital life inheritance.

[0011] To achieve the above objectives, the present invention adopts the following technical solution: A memory-directed migration and clone inheritance system based on digital life includes a processor and a memory electrically connected to the processor; The memory stores program instructions, and when the processor executes the program instructions, it performs the following steps: S1. Pre-configure memory inheritance rules, divide locally stored memories into three categories: inheritable core memories, private restricted memories, and prohibited inheritance memories, and generate a unique inheritance authorization key and a designated heir identity identifier; S2. After receiving the migration trigger instruction, the system automatically retrieves the baseline solidified model as the inheritance basis carrier, extracts only the memory vectors marked as inheritable categories, and automatically encrypts and isolates private and prohibited memory categories, which do not participate in the migration; wherein the migration trigger instruction includes an immediate migration instruction initiated by the user or an automatic migration instruction triggered based on abnormal vital sign monitoring. S3. Bind and package the selected core memory with the solidified benchmark model, encrypt the whole using the inherited authorization key, and generate an inheritable digital clone data package; S4. After the successor completes identity verification and key decryption, if there is no matching benchmark model locally, the benchmark solidified model in the inheritance data package is first decompressed and loaded. Then, the core memory vector is mapped and loaded to the solidified benchmark model through the vector alignment mapping algorithm to complete the digital clone inheritance and restore the original personality characteristics and inherited memories.

[0012] Furthermore, in step S1, the memory category classification is determined based on three dimensions: the semantic importance of the memory, the user's active labeling, and the time weight. The system automatically completes the classification labeling, while also supporting the user to manually adjust the boundaries of various memory categories.

[0013] Furthermore, in step S2, when extracting inheritable memories, the system automatically removes redundant temporary interactive content based on semantic redundancy analysis, retaining only core vectors such as key life experiences, family entrustment, and values, thereby reducing the size of the inheritance package and improving operating efficiency.

[0014] Furthermore, in step S3, the inherited data packet uses an asymmetric encryption method. The inheritor can only decrypt the packet by holding the private key, while the public key is only used for encrypted transmission to avoid data leakage during transmission.

[0015] Furthermore, in step S4, the vector alignment mapping algorithm specifically includes: obtaining the vector space dimension parameters of the local fixed benchmark model, performing dimensionality reduction processing and spatial coordinate calibration on the core memory vector, realizing deep fusion between the memory vector and the benchmark model, without destroying the original personality characteristics, and avoiding dialogue logic confusion.

[0016] Compared with the prior art, the present invention has the following technical advantages: This invention enables targeted selection and inheritance of digital life memories, strictly distinguishing between inheritable, private, and prohibited memories, significantly reducing the risk of privacy leaks. By binding to a pre-fixed baseline model and being compatible with scenarios where the inheritor does not have the original model, it ensures the complete authenticity of the digital clone personality after inheritance. Combined with asymmetric encryption and identity verification mechanisms, it ensures that digital life can only be passed on to designated inheritors. The overall system structure is simple, with low deployment costs and a high degree of automation in the migration process, making it suitable for ordinary users to achieve secure inheritance of their personal digital life. Specific Implementation

[0018] The embodiments of the present invention will be described in detail below with reference to specific application scenarios.

[0019] This embodiment relies on a regular personal computer for deployment, forming a complete technical system with the previous two patents. It can achieve targeted inheritance and memory migration of digital clones without the need for additional dedicated hardware.

[0020] During the normal use of the digital clone, users can access the inheritance configuration module in advance. The system automatically categorizes all memories into three types based on semantic importance, user-manual marking, and time weighting: inheritable core memories mainly include key life events, family instructions, and values; private and restricted memories include personal emotions, internal family conflicts, and other content visible only to the user; and prohibited inheritance memories contain absolutely private content, completely isolated and never migrated. Simultaneously, the user generates an inheritance key pair, delivers the private key to the designated heir, and stores the public key locally.

[0021] The system supports two migration triggering modes: one is for users to initiate instant migration while they are alive, and the other is to connect to vital sign monitoring equipment and automatically start the inheritance process when abnormal vital signs are detected, covering two scenarios: donations during life and inheritance after death.

[0022] When the inheritance migration instruction is triggered, the system automatically locks the currently running data and retrieves the previously solidified baseline model as the inheritance carrier, no longer using the temporary version from later iterations. Based on semantic redundancy analysis, the processor automatically filters out all core memory vectors marked as inheritable, removes redundant content such as casual conversation, and automatically puts private and prohibited memories into an encrypted isolation area, which are not included in this migration packaging.

[0023] After the selection is completed, the system binds and integrates the baseline solidified model with the core memory vector, uses the public key to complete the overall encryption, and generates a simplified digital clone inheritance data package. The data package can be transmitted to the inheritor via USB flash drive, cloud storage, etc.

[0024] The inheritor imports the inheritance data package onto their local device and enters a unique private key to complete decryption and identity verification. If no matching baseline model is available locally, the system automatically decompresses the data package and loads the original baseline solidified model. Subsequently, a vector alignment mapping algorithm is executed to obtain the vector space dimension of the current baseline model. The inherited memory vector is then dimensionality reduced and coordinate calibrated before loading, achieving deep integration of memory and personality. The inherited digital avatar fully retains the user's original personality, thought process, and designated core memories, while ensuring no privacy is leaked, truly achieving secure, faithful, and controllable digital life inheritance.

[0025] The entire migration and inheritance process is fully automated, requiring no manual intervention for memory screening. It relies on a pre-established model to ensure that the personality remains authentic and uses asymmetric encryption to protect inheritance permissions, effectively solving the three core issues of privacy, authenticity, and authorization in digital life inheritance.

Claims

1. A memory-directed transfer and clone inheritance system based on digital life, characterized in that, Includes a processor and a memory electrically connected to the processor; The memory stores program instructions, and when the processor executes the program instructions, it performs the following steps: S1. Pre-configure memory inheritance rules, divide local memories into three categories: inheritable core memories, private restricted memories, and prohibited inheritance memories, and generate inheritance authorization keys and designated heir identity identifiers; S2. After receiving the migration trigger instruction, the baseline solidified model is retrieved as the inheritance carrier, and only the inheritable core memory vector is extracted. Private and prohibited memory classes are encrypted and isolated and do not participate in the migration. The migration trigger instruction includes user-initiated real-time migration instruction or automatic migration instruction triggered by abnormal vital signs. S3. Bind and package the benchmark solidified model with the selected core memory, and generate a digital clone inheritance data package by encrypting it with the inheritance authorization key; S4. After the successor completes identity verification and key decryption, if there is no local model of the same type, the base solidified model in the data package is loaded first, and then the core memory is loaded into the solidified base model through the vector alignment mapping algorithm to complete the digital clone inheritance.

2. The system according to claim 1, characterized in that, In step S1, memory categories are determined based on three dimensions: semantic importance, user-initiated labeling, and time weighting. Users can also manually adjust the boundaries of memory categories.

3. The system according to claim 1, characterized in that, In step S2, the system automatically removes temporary interactive content based on semantic redundancy analysis, retaining only the core memory vectors of key life categories, thus reducing the size of the inherited data package.

4. The system according to claim 1, characterized in that, In step S3, the inherited data packet uses an asymmetric encryption mechanism. The public key is used for encryption, and the inheritor uses the private key to decrypt, ensuring the security of transmission and inheritance.

5. The system according to claim 1, characterized in that, In step S4, the vector alignment mapping algorithm includes obtaining the dimension of the reference model vector space, performing dimensionality reduction processing and coordinate calibration on the core memory vector, and realizing deep compatibility and fusion of the memory and solidified models.