Neural Network Representation Formats

Patent metadata

Field	Value
Publication number	US-2025384298-A1
Application number	US-202519303931-A
Country	US
Kind code	A1
Filing date	Aug 19, 2025
Priority date	Oct 1, 2019
Publication date	Dec 18, 2025
Grant date	—

Abstract

Data stream having a representation of a neural network encoded thereinto, the data stream including serialization parameter indicating a coding order at which neural network parameters, which define neuron interconnections of the neural network, are encoded into the data stream.

First claim

I/We claim:

1 . Data stream having a representation of a neural network encoded thereinto, the data stream comprising serialization parameter indicating a coding order at which neural network parameters, which define neuron interconnections of the neural network, are encoded into the data stream, wherein the neural network parameters are coded into the data stream using context-adaptive arithmetic coding.

2 . Apparatus for encoding a representation of a neural network into a data stream, wherein the apparatus is configured to provide the data stream with a serialization parameter indicating a coding order at which neural network parameters, which define neuron interconnections of the neural network, are encoded into the data stream, wherein the apparatus is configured to encode, into the data stream, the neural network parameters using context-adaptive arithmetic encoding.

3 . Apparatus for decoding a representation of a neural network from a data stream, wherein the apparatus is configured to decode from the data stream a serialization parameter indicating a coding order at which neural network parameters, which define neuron interconnections of the neural network, are encoded into the data stream, wherein the apparatus is configured to decode, from the data stream, the neural network parameters using context-adaptive arithmetic decoding.

4 . Apparatus of claim 3 , wherein the data stream is structured into one or more individually accessible portions, each individually accessible portion representing a corresponding neural network layer of the neural network, and
wherein the apparatus is configured to decode serially, from the data stream, neural network parameters, which define neuron interconnections of the neural network within a predetermined neural network layer, and use the coding order to assign neural network parameters serially decoded from the data stream to the neuron interconnections.

5 . Apparatus of claim 3 , wherein the serialization parameter is indicative of a permutation using which the coding order permutes neurons of a neural network layer relative to a default order.

6 . Apparatus of claim 5 , wherein the permutation orders the neurons of the neural network layer in a manner so that the neural network parameters monotonically increase along the coding order or monotonically decrease along the coding order.

7 . Apparatus of claim 5 , wherein the permutation orders the neurons of the neural network layer in a manner so that, among predetermined coding orders signalable by the serialization parameter, a bitrate for coding the neural network parameters into the data stream is lowest for the permutation indicated by the serialization parameter.

8 . Apparatus of claim 3 , wherein the neural network parameters comprise weights and biases.

9 . Apparatus of claim 3 , wherein the apparatus is configured to decode, from the data stream, individually accessible sub-portions, into which individually accessible portions the data stream is structured, each sub-portion representing a corresponding neural network portion of the neural network, so that each sub-portion is completely traversed by the coding order before a subsequent sub-portion is traversed by the coding order.

10 . Apparatus of claim 4 , wherein the neural network parameters are decoded from the data stream using context-adaptive arithmetic decoding and using context initialization at a start of any individually accessible portion or sub-portion.

11 . Apparatus of claim 4 , wherein the apparatus is configured to decode, from the data stream, start codes at which each individually accessible portion or sub-portion begins, and/or pointers pointing to beginnings of each individually accessible portion or sub-portion, and/or pointers data stream lengths of each individually accessible portion or sub-portion for skipping the respective individually accessible portion or sub-portion in parsing the data stream.

12 . Apparatus of claim 3 , wherein the apparatus is configured to decode, from the data stream, a numerical computation representation parameter indicating a numerical representation and bit size at which the neural network parameters are to be represented when using the neural network for inference.

13 . Apparatus of claim 3 , wherein the data stream, is structured into individually accessible sub-portions, each individually accessible sub-portion representing a corresponding neural network portion of the neural network, so that each individually accessible sub-portion is completely traversed by the coding order before a subsequent individually accessible sub-portion is traversed by the coding order, wherein the apparatus is configured to decode, from the data stream, for a predetermined individually accessible sub-portion the neural network parameter and a type parameter indicting a parameter type of the neural network parameter decoded from the predetermined individually accessible sub-portion.

14 . Apparatus of claim 13 , wherein the type parameter discriminates, at least, between neural network weights and neural network biases.

15 . Apparatus of claim 3 , wherein the data stream, is structured into one or more individually accessible portions, each one or more individually accessible portion representing a corresponding neural network layer of the neural network, and
wherein the apparatus is configured to decode, from the data stream, for a predetermined neural network layer, a neural network layer type parameter indicating a neural network layer type of the predetermined neural network layer of the neural network.

16 . Apparatus of claim 15 , wherein the neural network layer type parameter discriminates, at least, between a fully-connected and a convolutional layer type.

17 . Apparatus of claim 3 , wherein the apparatus is configured to decode a representation of a neural network from the data stream, wherein the data stream is structured into one or more individually accessible portions, each individually accessible portion representing a corresponding neural network layer of the neural network, and wherein the data stream is, within a predetermined portion, further structured into individually accessible sub-portions, each sub-portion representing a corresponding neural network portion of the respective neural network layer of the neural network, wherein the apparatus is configured to decode from the data stream, for each of one or more predetermined individually accessible sub-portions
a start code at which the respective predetermined individually accessible sub-portion begins, and/or a pointer pointing to a beginning of the respective predetermined individually accessible sub-portion, and/or a data stream length parameter indicating a data stream length of the respective predetermined individually accessible sub-portion for skipping the respective predetermined individually accessible sub-portion in parsing the data stream.

18 . Apparatus of claim 17 , wherein the apparatus is configured to decode, from the data stream, the representation of the neural network using context-adaptive arithmetic decoding and using context initialization at a start of each individually accessible portion and each individually accessible sub-portion.

19 . Apparatus of claim 3 , wherein the apparatus is configured to decode a representation of a neural network from a data stream, wherein the data stream is structured into individually accessible portions, each portion representing a corresponding neural network portion of the neural network, wherein t

Assignees

• Fraunhofer Ges Forschung

Inventors

• Matlage Stefan
• Haase Paul
• Kirchhoffer Heiner
• Mueller Karsten
• Samek Wojciech
• Wiedemann Simon
• Marpe Detlev
• Schierl Thomas
• Sánchez De La Fuente Yago
• Skupin Robert
• Wiegand Thomas

Classifications

• H03M7/70

  Type of the data to be coded, other than image and sound · CPC title

• H03M7/4018

  Context adapative binary arithmetic codes [CABAC] · CPC title

• G06N3/08

  Learning methods · CPC title

• G06N3/0495

  Quantised networks; Sparse networks; Compressed networks · CPC title

• H03M7/6023

  Parallelization · CPC title