A structured representation format and cross-device sharing method for personal memory data
Patent Information
- Application Number
- CN202610892559.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-20
- Publication Date
- 2026-10-23
AI Technical Summary
当跨设备交换记忆数据时，出现字段表示方式不一致的问题：例如一个设备上的人物字段为数组格式，另一个设备上为逗号分隔字符串
[0020]与现有技术相比，本发明具有以下有益效果：第一，通过预定义语义锚点字段实现个人记忆数据的标准化表示，与通用数据格式形成明确区分。第二，通过存在性指示标记实现紧凑存储与灵活适配的统一。第三，通过版本标识符保障不同版本格式的互操作性。第四，通过跨设备格式转换方法封堵任意自研格式的绕路路径。
Smart Images
Figure FT_1 
Figure FT_2 
Figure FT_3
Abstract
Description
Technical Field
[0001] This invention belongs to the field of data structure and serialization technology, specifically relating to a unified structured representation format for personal memory data and a cross-device sharing method, applicable to memory data exchange between wearable devices, mobile terminals, vehicle systems and smart home devices.
[0002] Terminology Definition
[0003] The terms used in this specification and claims are defined as follows:
[0004] (1) Personal memory data: including but not limited to structured data, unstructured text, raw audio and video data, feature identification data, fingerprint data and any combination thereof, as long as its content or identification is related to the experience, behavior and status of a specific individual.
[0005] (2) Processing and calling: including but not limited to any computational behavior that can acquire, present or utilize memory data content, such as retrieval, matching, addressing, mapping, generation, reasoning, context window association, etc.
[0006] (3) Predefined semantic anchor fields: These refer to the set of predefined fields that constitute the core semantic dimensions of personal memory data, including at least six dimensions: time, location, people, dialogue, events, and emotions. These fields can be explicit structured fields or implicit in unstructured text and extracted by algorithms. Background Technology
[0007] Memory cross-connection systems use different storage formats on different computing devices—JSON might be used on smart glasses, Protobuf on smartwatches, and SQLite composite columns on mobile phones. When exchanging memory data across devices, inconsistencies arise in field representation: for example, a person field on one device might be an array, while on another it might be a comma-separated string. Existing general serialization schemes such as JSON, Protobuf, BSON, and MessagePack are all general data exchange formats and lack predefined semantic anchor dimensions and existence indication mechanisms for personal memory data. As of June 2026, there is still no standardized structured representation format for personal memory data that covers multiple types of computing devices. Summary of the Invention
[0008] The technical problem to be solved by the present invention is to provide a structured representation format for personal memory data and a method for sharing across devices, so that personal memory data generated by any computing device has a unified semantic definition and serialization standard.
[0009] This invention provides a structured representation format for personal memory data, comprising:
[0010] Predefined semantic anchor field units define a set of standardized memory attribute types. Each field corresponds to a preset personal memory attribute (including at least six-dimensional semantic anchors: time, place, person, dialogue, event, and emotion). When storing or transmitting personal memory data on different devices, the data is organized according to the same set of standardized field definitions.
[0011] The existence indicator unit identifies which predefined semantic anchor fields contain valid data in the current data instance by setting a bitmap in the format header, so that the format can maintain a compact structure and flexibly adapt to different acquisition scenarios.
[0012] The version identifier unit enables compatibility and interoperability between different versions of the format by recording the data model version number in the format header, and supports the continuous evolution of the format.
[0013] The present invention also provides a method for cross-device sharing of personal memory data, comprising: identifying at least one segment of personal memory data on a first computing device storing the personal memory data in a first format; converting the personal memory data from the first format into a target format suitable for cross-device sharing, the target format including at least one predefined semantic anchor field; and using the personal memory data in the target format on at least a second computing device.
[0014] Regarding the specific implementation of serialization, this invention supports multiple serialization schemes as equivalent implementations, including but not limited to:
[0015] (1) Binary format of fixed-length header plus variable-length body: The header contains a version identifier (2 bytes), an existence bitmap (1 byte) and a total length (2 bytes). The body arranges the semantic anchor data in the order of the fields. The byte order is little-endian and supports direct reading by offset.
[0016] (2) Schema definition based on Protocol Buffers: nest each semantic anchor field in the message structure of protobuf, and use the optional field of proto3 to achieve the equivalent function of existence marking.
[0017] (3) Schema serialization based on Apache Avro: Using the field structure of Avro's schema registry, the equivalent function of variable length body is achieved through Avro's default value and union type.
[0018] All equivalent substitutions for the above serialization schemes fall within the protection scope of this invention.
[0019] The present invention also provides an encoder and a decoder configured to serialize the structured representation format into a binary data stream and deserialize it from the binary data stream.
[0020] Compared with existing technologies, this invention has the following advantages: First, it achieves standardized representation of personal memory data through predefined semantic anchor fields, clearly distinguishing it from general data formats. Second, it achieves a balance between compact storage and flexible adaptation through existence indicator markers. Third, it ensures interoperability between different version formats through version identifiers. Fourth, it blocks bypasses the loopholes of arbitrary self-developed formats through cross-device format conversion methods. Attached Figure Description
[0021] Figure 1 This is a schematic diagram of the data structure for the structured representation format of personal memory data in this invention.
[0022] Figure 2 This is a schematic diagram of the bitmap encoding of the existence indicator mark of the present invention.
[0023] Figure 3 This is a flowchart of the cross-device format conversion method of the present invention. Detailed Implementation
[0024] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the present invention.
[0025] ### Example 1: Structured Storage of Data Memory on Smart Glasses
[0026] The smart glasses collected a piece of personal memory data (Time: June 17, 2026, 14:30; Location: Nanshan District, Shenzhen; People: [Zhang San, Li Si]; Dialogue: "Let's have dinner together next time"). The glasses' encoder organized the data into a structured format according to predefined semantic anchor fields—time field containing timestamps (Unix milliseconds), location field containing geographic coordinates in JSON format, people field containing a list of identifiers, and dialogue field containing UTF-8 text. The existence indicator is a bitmap, with each bit corresponding to the existence status of each field in the instance; in this example, there is no emotion field, so the corresponding bit is 0. The version identifier records the current format version number (v1.0).
[0027] ### Example 2: Bitmap Encoding of Existence Indicator Markers
[0028] There are a total of 6 predefined semantic anchor fields (time, location, person, dialogue, event, and emotion). The encoder generates a one-byte bitmap marker during output, with the lower 6 bits corresponding to the 6 predefined fields and the higher 2 bits reserved. Optional dimensions such as scene and tag are carried through extended field slots. When only the time and person fields have valid data, the bitmap value is 00010001 (bit 0 corresponds to time, and bit 2 corresponds to person). After reading the bitmap, the decoder only decodes the fields marked as 1, ignoring fields marked as 0.
[0029] ### Example 3: Cross-device format conversion (Protobuf → Standard format)
[0030] The first computing device (such as a smartwatch) stores the memory data in Protobuf format. After the watch encoder identifies a segment of memory data that needs to be synchronized to the second computing device (such as smart glasses), it maps the Protobuf fields to predefined semantic anchor fields—timestamp to the time dimension, gps_coordinates to the location dimension, and participants list to the people dimension. After mapping, it is serialized in binary format (little-endian fixed-length header + variable-length body) and transmitted to the glasses via Bluetooth. The glasses decoder parses and uses the memory data according to the predefined fields.
[0031] ### Example 4: Cross-Device Format Conversion (SQLite → Standard Format)
[0032] The primary computing device (such as a mobile phone) uses SQLite structured column storage. The mobile encoder extracts the memory columns (create_time, location, people, content, etc.) from the database and maps them to predefined semantic anchor fields. After mapping, it can choose to serialize the data in protobuf or binary format and transmit it to the target device.
[0033] ### Example 5: Semantic Anchor Extraction from Unstructured Text
[0034] The user recorded a freely expressed natural language text message using the voice input function of smart glasses: "Discussed patent layout with Zhang San at Shenzhen Science Park on the afternoon of May 27, 2026". The system sent this text into the NLP entity extraction and semantic parsing pipeline, automatically extracting the following fields: time ("May 27, 2026 afternoon" → Unix millisecond timestamp 2037387600000), location ("Shenzhen Science Park" → geographic coordinates 22.5431, 113.9535), person ("Zhang San" → user's contact identifier contact-zhangsan), and event ("discussed patent layout" → event tag event-discussion). The extracted results were filled into the corresponding slots of the predefined semantic anchor fields. The bits corresponding to the time, location, person, and event fields in the existence indicator were set to 1, while the dialogue and sentiment fields were set to 0. Although this memory data was input as unstructured text, it was still stored and transmitted in the structured format defined by this patent after system processing. This embodiment illustrates that by using unstructured text as user input and then having the system automatically extract semantic anchors, the underlying storage still falls within the protection scope of the predefined semantic anchor field of this patent.
Claims
1. A structured representation format for personal memory data, characterized in that, include: One or more predefined semantic anchor fields, each field corresponding to a semantic dimension of a personal memory attribute, used to store personal memory attribute data of a preset type; An existence indicator flag is used to identify which fields among the predefined semantic anchor fields contain valid data; A version identifier is used to indicate the version of the data model in the structured representation format.
2. A method for cross-device sharing of personal memory data, characterized in that, include: Identify at least one segment of personal memory data on a first computing device that stores personal memory data in a first format; The personal memory data is converted from the first format to a target format suitable for cross-device sharing, the target format containing at least one predefined semantic anchor field; The personal memory data is used in the target format on at least a second computing device.
3. The method as described in claim 1 or 2, characterized in that, The computing device includes at least one of wearable devices, mobile terminals, in-vehicle systems, smart home devices, and any computing device that includes a processor and memory.
4. The format as described in claim 1, characterized in that, The predefined semantic anchor fields include at least a time field, a location field, a person field, a dialogue field, an event field, and an emotion field.
5. The format as described in claim 1, characterized in that, The existence indicator is a bitmap marker, with each bit corresponding to the existence status of a predefined semantic anchor field.
6. The format as described in claim 1, characterized in that, The structured representation format further includes extended field slots, which support adding custom fields without breaking compatibility with older versions of the format.
7. The format as described in claim 1, characterized in that, The structured representation format is serialized into a binary format with a fixed-length header and a variable-length body. The fixed-length header includes at least the version identifier, the existence indicator, and the total length information.
8. The format as described in claim 7, characterized in that, The serialized byte order of the binary format is little-endian, which supports reading each field data directly by offset without deserialization.
9. The format as described in claim 1, characterized in that, The predefined semantic anchor field can be implicit in unstructured natural language text and extracted from the unstructured text by a natural language processing algorithm. In this case, the existence indicator is the truth value identifier of the algorithm extraction result.