# US 2025/0386221 A1 · Identify infrastructure with ml embeddings

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0386221 A1
- Document ID: US20250386221A1
- Status: Published application
- Kind: A1
- Application number: 18/741,687
- Filed: 2024-06-12
- Published: 2025-12-18
- Earliest claimed: 2024-06-12
- Assignees: QUALCOMM Incorporated
- Applicants: QUALCOMM Incorporated
- Inventors: Himanshu JOSHI; June NAMGOONG; Taesang YOO; Jay Kumar SUNDARARAJAN; Eren BALEVI
- Web page: https://patentgrep.com/patent/US20250386221A1

## CPC

- H04W 24/10: Scheduling measurement reports {; Arrangements for measurement reports}
- H04W 72/02: Selection of wireless resources by user or terminal
- H04L 1/0003: {by switching between different modulation schemes}
- H04L 1/1812: Hybrid protocols; Hybrid automatic repeat request [HARQ]
- H04W 24/02: Arrangements for optimising operational condition
- H04W 72/1273: of downlink data flows
- H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72)
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

Apparatus, methods, and computer program products for wireless communication are provided. An example method may include collecting a set of observation data associated with at least one network node. The example method may further include feeding the set of observation data to an embedding engine. The example method may further include receiving, from the embedding engine, a set of embeddings, the set of embeddings being a set of representations of one or more characteristics of a set of network nodes including the at least one network node. The example method may further include communicating with a second network node based on the set of embeddings.

## Claims

### Claim 1

An apparatus for wireless communication at a user equipment (UE), comprising:
at least one memory; and
at least one processor coupled to the at least one memory, and based at least in part on information stored in the at least one memory, the at least one processor, individually or in any combination, is configured to cause the UE to:
collect a set of observation data associated with at least one network node;
feed the set of observation data to an embedding engine;
receive, from the embedding engine, a set of embeddings, the set of embeddings being a set of representations of one or more characteristics of a set of network nodes including the at least one network node; and
communicate with a second network node based on the set of embeddings.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein the set of observation data comprises a set of link level protocol interaction data between the UE and the at least one network node, wherein the set of link level protocol interaction data comprises data associated with at least one of: a physical (PHY) layer signaling, a medium access control (MAC) level signaling, a hybrid automatic repeat request (HARQ) result, a channel state feedback (CSF) report, a channel quality indicator (CQI), a rank indicator, a physical downlink shared channel (PDSCH) grant rank, or a modulation and coding scheme (MCS).

### Claim 3 (depends on 1)

The apparatus of claim 1, wherein the set of embeddings are categorized based on a set of respective manufacturers associated with the set of network nodes, and wherein the set of representations are associated with a set of schedulers associated with the set of network nodes.

### Claim 4 (depends on 3)

The apparatus of claim 3, wherein the set of embeddings are based on: (1) a set of linear projections of the set of observation data, each linear projection of the set of linear projections corresponding to one respective slot associated with the set of observation data, (2) a set of transform encoder outputs based on the set of linear projections, one or more classify token associated with the set of linear projections being categorized based on the set of respective manufacturers, and (3) at least one context and at least one rank associated with the set of transform encoder outputs.

### Claim 5 (depends on 1)

The apparatus of claim 1, wherein the embedding engine is located at the UE or a server separate from the UE.

### Claim 6 (depends on 1)

The apparatus of claim 1, wherein to communicate with the second network node, the at least one processor, individually or in any combination, is configured to cause the UE to:
identify the second network node as having similar characteristics as one particular network node of the set of network nodes based on the set of embeddings;
perform at least one grant prediction of a next grant from the second network node based on an identification of the second network node; and
communicate with the second network node based on the at least one grant prediction.

### Claim 7 (depends on 6)

The apparatus of claim 6, wherein to perform the at least one grant prediction based on the identification, the at least one processor, individually or in any combination, is configured to cause the UE to:
perform each of the at least one grant prediction in parallel using a respective embedding of the set of embeddings; and
select a best performing grant prediction of the at least one grant prediction.

### Claim 8 (depends on 1)

The apparatus of claim 1, wherein the at least one processor, individually or in any combination, is further configured to cause the UE to:
transmit a capability message indicating a capability associated with utilization of the set of embeddings.

### Claim 9 (depends on 1)

The apparatus of claim 1, wherein the at least one processor, individually or in any combination, is further configured to cause the UE to:
store and maintain the set of embeddings in the at least one memory.

### Claim 10 (depends on 1)

The apparatus of claim 1, wherein the set of representations in the set of embeddings is based on a location associated with the UE.

### Claim 11

An apparatus for wireless communication at a user equipment (UE), comprising:
at least one memory; and
at least one processor coupled to the at least one memory, and based at least in part on information stored in the at least one memory, the at least one processor, individually or in any combination, is configured to cause the UE to:
transmit, to an embedding engine upon a cell change, a set of link level protocol interaction data between the UE and a network node;
receive, from the embedding engine, an embedding associated with the network node, the embedding being a representation of one or more characteristics of the network node; and
communicate with the network node based on the embedding.

### Claim 12 (depends on 11)

The apparatus of claim 11, wherein the set of link level protocol interaction data comprises data associated with at least one of: a physical (PHY) layer signaling, a medium access control (MAC) level signaling, a hybrid automatic repeat request (HARQ) result, a channel state feedback (CSF) report, a channel quality indicator (CQI), a rank indicator, a physical downlink shared channel (PDSCH) grant rank, or a modulation and coding scheme (MCS).

### Claim 13 (depends on 11)

The apparatus of claim 11, wherein the embedding is associated with a scheduler associated with the network node.

### Claim 14 (depends on 13)

The apparatus of claim 13, wherein the embedding is based on: (1) a set of linear projections of a set of observation data, each linear projection of the set of linear projections corresponding to one respective slot associated with the set of observation data, (2) a set of transform encoder outputs based on the set of linear projections, one or more classify token associated with the set of linear projections being categorized based on a set of manufacturers, and (3) at least one context and at least one rank associated with the set of transform encoder outputs.

### Claim 15 (depends on 11)

The apparatus of claim 11, wherein to communicate with the network node, the at least one processor, individually or in any combination, is configured to cause the UE to:
perform at least one grant prediction of a next grant from the network node based on the embedding; and
communicate with the network node based on the at least one grant prediction.

### Claim 16 (depends on 15)

The apparatus of claim 15, wherein the at least one processor, individually or in any combination, is further configured to cause the UE to:
transmit, to the embedding engine, a set of additional link level protocol interaction data after communication with the network node based on the at least one grant prediction; and
receive an updated embedding from the embedding engine based on the set of additional link level protocol interaction data.

### Claim 17

A method for wireless communication performed by a user equipment (UE), comprising:
collecting a set of observation data associated with at least one network node;
feeding the set of observation data to an embedding engine;
receiving, from the embedding engine, a set of embeddings, the set of embeddings being a set of representations of one or more characteristics of a set of network nodes including the at least one network node; and
communicating with a second network node based on the set of embeddings.

### Claim 18 (depends on 17)

The method of claim 17, wherein the set of observation data comprises a set of link level protocol interaction data between the UE and the at least one network node, wherein the set of link level protocol interaction data comprises data associated with at least one of: a physical (PHY) layer signaling, a medium access control (MAC) level signaling, a hybrid automatic repeat request (HARQ) result, a channel state feedback (CSF) report, a channel quality indicator (CQI), a rank indicator, a physical downlink shared channel (PDSCH) grant rank, or a modulation and coding scheme (MCS).

### Claim 19 (depends on 17)

The method of claim 17, wherein the set of embeddings are categorized based on a set of respective manufacturers associated with the set of network nodes, and wherein the set of representations are associated with a set of schedulers associated with the set of network nodes.

### Claim 20 (depends on 19)

The method of claim 19, wherein the set of embeddings are based on: (1) a set of linear projections of the set of observation data, each linear projection of the set of linear projections corresponding to one respective slot associated with the set of observation data, (2) a set of transform encoder outputs based on the set of linear projections, one or more classify token associated with the set of linear projections being categorized based on the set of respective manufacturers, and (3) at least one context and at least one rank associated with the set of transform encoder outputs.
