# US 2026/0087680 A1 · Reinforcement learning-based rate control for end-to-end neural network based video compression

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

- Number: US 2026/0087680 A1
- Document ID: US20260087680A1
- Status: Published application
- Kind: A1
- Application number: 19/114,220
- Filed: 2023-09-22
- Published: 2026-03-26
- Earliest claimed: 2022-09-23
- Assignees: InterDigital VC Holdings, Inc.
- Applicants: Interdigital Vc Holdings, Inc.
- Inventors: Fabien RACAPE; Ujwal DINESHA; Hyomin CHOI; Syed Mateen UL HAQ
- Web page: https://patentgrep.com/patent/US20260087680A1

## CPC

- G06T 9/002: {using neural networks}
- H04N 19/147: according to rate distortion criteria (rate-distortion as a criterion for motion estimation H04N19/567)
- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/047: Probabilistic or stochastic networks
- G06N 3/0475: Generative networks
- G06N 3/092: Reinforcement learning
- H04N 19/124: Quantisation
- H04N 19/177: the unit being a group of pictures [GOP]
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An end-to-end neural network-based rate control method based on reinforcement learning implements video codec embodiments. In one embodiment, the codec environment is based on an Asymmetric Gained Variational Auto-Encoder (AG-VAE) architecture. A Reinforcement Learning (RL) agent is implemented through a deep convolutional neural network. In an embodiment, the RL agent conveys a choice of gain vector to the AG-VAE codec and receives reward data from the AG-VAE environment. Rate control is optimized over a period of frames, such as a Group of Pictures (GOP).

## Claims

### Claim 1

A method, comprising:
encoding a portion of video using a determined number of bits; and,
determining the number of bits to allocate for the encoded portion of video based on a number of frames, wherein said determining comprises using a gain vector from a reinforcement learning agent that uses a latent determined from the encoding.

### Claim 2

An apparatus configured to perform:
encoding a portion of video using a determined number of bits; and,
determining the number of bits to allocate for the encoded portion of video based on a number of frames, wherein said determining comprises using a gain vector from a reinforcement learning agent that uses a latent determined from the encoding.

### Claim 3

A method, comprising:
parsing video data for a lambda value;
determining an index of a vector with corresponding lambda that is closest to a target lambda value, wherein lambda defines a rate-distortion operating point;
interpolating values of vectors between the determined index vector and one having a consecutive index, using the target lambda value and lambda values of the vectors between the determined index vector and one having a consecutive index; and,
decoding the video data using the interpolated vectors.

### Claim 4

An apparatus configured to perform:
parsing video data for a lambda value;
determining an index of a vector with corresponding lambda that is closest to a target lambda value, wherein lambda defines a rate-distortion operating point;
interpolating values of vectors between the determined index vector and one having a consecutive index, using the target lambda value and lambda values of the vectors between the determined index vector and one having a consecutive index; and,
decoding the video data using the interpolated vectors.

### Claim 5 (depends on 1)

The method of claim 1, wherein an Asymmetric Gained Variational Auto-Encoder (AG-VAE) architecture is used.

### Claim 6 (depends on 1)

The method of claim 1, wherein the reinforcement learning is implemented by a Deep Neural Network (DNN).

### Claim 7 (depends on 1)

The method of claim 1, wherein said number of frames is a Group of Pictures (GOP).

### Claim 8 (depends on 1)

The method of claim 1, wherein said reinforcement learning agent determines a range of bitrate points.

### Claim 9 (depends on 2)

The apparatus of claim 2, wherein the video processed is all intra mode prediction.

### Claim 10 (depends on 1)

The method of claim 1, wherein the reinforcement learning agent receives frame type as input.

### Claim 11 (depends on 4)

A device comprising:
an apparatus according to claim 4; and
at least one of (i) an antenna configured to receive a signal, the signal including the video block, (ii) a band limiter configured to limit the received signal to a band of frequencies that includes the video block, and (iii) a display configured to display an output representative of a video block.

### Claim 12 (depends on 1)

A non-transitory computer readable medium containing data content generated according to the method of claim 1, for playback using a processor.

### Claim 13

(canceled)

### Claim 14 (depends on 3)

A computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method of claim 3.

### Claim 15 (depends on 1)

A non-transitory computer readable medium containing data content comprising instructions to perform the method of claim 1.

### Claim 16 (depends on 3)

The method of claim 3, wherein an Asymmetric Gained Variational Auto-Encoder (AG-VAE) architecture is used.

### Claim 17 (depends on 2)

The apparatus of claim 2, wherein the reinforcement learning is implemented by a Deep Neural Network (DNN).

### Claim 18 (depends on 2)

The apparatus of claim 2, wherein said number of frames is a Group of Pictures (GOP).

### Claim 19 (depends on 2)

The apparatus of claim 2, wherein said reinforcement learning agent determines a range of bitrate points.

### Claim 20 (depends on 3)

The method of claim 3, wherein the video processed is all intra mode prediction.

### Claim 21 (depends on 2)

The apparatus of claim 2, wherein the reinforcement learning agent receives frame type as input.
