# US 2025/0165802 A1 · Collaborative caching framework for multi-edge systems with robust federated deep learning

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

- Number: US 2025/0165802 A1
- Document ID: US20250165802A1
- Status: Published application
- Kind: A1
- Application number: 18/408,610
- Filed: 2024-01-10
- Published: 2025-05-22
- Earliest claimed: 2023-11-20
- Assignees: Fuzhou University
- Applicants: Fuzhou University
- Inventors: Zheyi CHEN; Jie LIANG; Luying ZHONG; Jiayu ZHENG
- Web page: https://patentgrep.com/patent/US20250165802A1

## CPC

- G06N 3/098: Distributed learning, e.g. federated learning
- G06N 3/084: Backpropagation, e.g. using gradient descent
- H04L 67/1004: Server selection for load balancing
- H04L 67/2885: Hierarchically arranged intermediate devices, e.g. for hierarchical caching
- H04L 67/289: Intermediate processing functionally located close to the data consumer application, e.g. in same machine, in same home or in same sub-network
- H04W 4/50: Service provisioning or reconfiguring
- H04W 28/086: among access entities
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72)

## Abstract

A Collaborative Caching Framework for Multi-edge Systems with Robust Federated Deep Learning is provided. First, we design a new partitioning mechanism for multi-dimensional cache space, enabling precise content recommendations in user classification intervals. Next, we develop a VQ-VAE-based accurate prediction for content popularity by overcoming posterior collapse. Finally, we create a new training mode and proactive cache replacement strategy based on robust federated deep learning. Specifically, residual-based detection for adversarial model updates and similarity-based federated aggregation are integrated to avoid the model destruction caused by adversarial updates, which enables the proactive cache replacement adapting to optimized cache resources and thus enhances cache performance. Using real-world testbed and MovieLens datasets, extensive experiments verify that RoCoCache achieves higher cache hit rates and efficiency than benchmark methods while ensuring better robustness. Moreover, we demonstrate the effectiveness of the components designed in RoCoCache via ablation experiments.

## Claims

### Claim 1

A Collaborative Caching Framework for Multi-edge Systems with Robust Federated Deep Learning, wherein the Collaborative Caching Framework for Multi-edge Systems which consists of M MEC nodes, each contains a MEC server and a base station, donated by the set E={e1, e2, . . . , em, . . . , eM}, and N users, donated by the set U={u1, u2, . . . , un, . . . , UN}; the caching space of MEC nodes is donated as the set C={C1, C2, . . . , Cm, . . . , CM}; each user is connected to a MEC node, and they communicate with each other via the wireless link provided by the associated base station; furthermore, the communications among MEC nodes and between MEC nodes and the cloud data center are conducted via the backhaul link; the caching space status of each MEC node is periodically broadcast to the other MEC nodes within the proposed system; moreover, the content library of the cloud data center, denoted by F={f1, f2, . . . , fi, . . . , fI}, where I indicates the number of accessible contents; it is noted that users are discretely distributed in the service zone of each edge node.

### Claim 2 (depends on 1)

The Collaborative Caching Framework for Multi-edge Systems with Robust Federated Deep Learning according to claim 1, wherein when the user un sends a request for the content fi to its connected MEC node, the workflow is given as follows:
Step 1: the current MEC node checks whether it has cached fi; if fi is cached, the MEC node will send it to un directly; otherwise, it goes to Step 2;
Step 2: the current MEC node searches for whether there exists a collaborative MEC node that caches fi; if there exists, the collaborative MEC node will forward fi to the current MEC node via the backhaul link, and then fi will be sent to un; otherwise, it goes to Step 3;
Step 3: if no collaborative MEC node caches fi, the content library in the cloud data center will provide fi and forward it to the current MEC node through the backhaul link, and then fi will be sent to un.

### Claim 3 (depends on 1)

The Collaborative Caching Framework for Multi-edge Systems with Robust Federated Deep Learning according to claim 1, wherein the popularity of fi on the MEC node em is defined as
P
i
,
m
=
req
i
,
m
req
m
(
1
)
where redi,m is the number of requests for fi received by em, and reqm is the total number of requests received by em;
the proposed RoCoCache enables precise prediction of content popularity; to evaluate the prediction accuracy, the global loss function is defined as
J
⁡
(
w
(
r
)
)
=
∑
m
=
1
M
r
⁢
e
⁢
q
m
r
⁢
e
⁢
q
⁢
M
⁢
S
⁢
E
m
(
w
m
(
r
)
)
(
2
)
where r indicates the FL communication round, w(r) is the parameter of the global prediction model, req is the total number of requests received by all MEC nodes, wm(r) is the parameter of a local prediction model, and Mean-Square Error (MSE) is defined as
M
⁢
S
⁢
E
m
⁢
(
w
m
(
r
)
)
=
1
I
⁢
∑
j
=
1
I
(
y
i
,
m
⁢
(
w
m
(
r
)
)
-
P
i
,
m
⁢
(
r
)
)
2
(
3
)
where i, m(wm(r) is the predicted popularity value of fi on em and Pi,m(r) is the actual value;
moreover, the cache hit rate is defined as
H
=
∑
m
=
1
M
⁢
∑
i
=
1
M
⁢
θ
m
⁢
(
f
i
)
∑
m
=
1
M
⁢
req
m
×
100
⁢
%
(
4
)
where θm(fi) indicates whether em caches the content requested by users or not, and it is defined as
θ
m
⁢
(
f
i
)
=
{
1
,
if
⁢
f
i
⁢
in
⁢
C
m
0
,
otherwise
.
(
5
)

### Claim 4 (depends on 1)

The Collaborative Caching Framework for Multi-edge Systems with Robust Federated Deep Learning according to claim 1, wherein proposing a user partitioning method based on multi-dimensional user features including gender, age, and occupation, which are mapped to coordinate axes, denoted by Θ={l1, l2, . . . , lt, . . . , lT}; at the initial stage (Grade=0), all users with different features are placed within the same user interval (h0); if the number of users in h0 exceeds the threshold ζ(Grade), it will be equally divided into 2T user intervals along each dimension, where the length of each divided dimension will be halved (lt=lt/2); ζ(Grade) determines the number of users in a user interval; when ζ(Grade) is larger, there are more users in each user interval, which cannot well reflect the unique preferences of different users; when ζ(Grade) is smaller, there are fewer users in each user interval, which may lead to inaccurate cache prediction; to achieve adaptive partitioning of user intervals and capture the potential relationships between interval users and their preferred contents, we set ζ(Grade)=α2Grade, where α is a hyper-parameter; the partitioned user intervals are denoted as H={h1, h2, . . . , hs, . . . , hS}, where Sis the number of user intervals; the partitioning may continue and go to the following stages (e.g., Grade=1, 2, . . . ) according to performance requirements.

### Claim 5 (depends on 1)

The Collaborative Caching Framework for Multi-edge Systems with Robust Federated Deep Learning according to claim 1, wherein the user activity and memory access interval are defined as
active
⁢
(
h
s
)
=
req
s
req
m
(
6
)
diverge
⁢
(
h
s
)
=
Γ
⁢
(
h
s
)
I
(
7
)
where reqs is the number of user requests and Γ(hs) is the memory access interval of hs;
considering the above factors, the size of cache space allocated to hs is defined as
cache
s
=
φ
⁢
(
num
⁢
(
h
s
)
,
active
⁢
(
h
s
)
,
diverge
⁢
(
h
s
)
)
×
cache
m
∑
s
=
1
S
⁢
φ
⁢
(
num
⁢
(
h
s
)
,
active
⁢
(
h
s
)
,
diverge
⁢
(
h
s
)
)
(
8
)
where cachem indicates the size of cache space on the MEC node connected to hs.

### Claim 6 (depends on 1)

The Collaborative Caching Framework for Multi-edge Systems with Robust Federated Deep Learning according to claim 1, wherein the user request matrix X contains historical information of user-requested contents on MEC nodes, which is defined as
X
=
[
x
i
,
…
,
x
n
,
…
,
x
N
]
T
∈
ℝ
N
×
I
(
9
)
where 1≤n≤N and with n presenting the number of users connected to a MEC node; xn=[xn1, . . . , xni, . . . , xnI]T indicates the content request record of the user n, where 1≤i≤I and t is the index of the content library; xni=1 indicates the successful content request; xni=0 indicates either a failed content request or the content that is not of interest, and these two cases are hard to be distinguished, leading to inaccurate prediction; to solve this issue, we supplement and calibrate the matrix X;
in VQ-VAE, the implicit embedded space is defined as v∈K×D, where K is the space size and D is the dimension of the embedded vector; thus, there are K embedded vectors vk∈D (k∈1, 2, 3, . . . . K); the VQ-VAE inputs xn and outputs tv(xn) via the encoder network; next, discrete hidden variable t is calculated by the nearest neighbor algorithm, and the posterior probability distribution q(t|xn) is one-hot encoding, which is defined as
q
⁢
(
t
=
k
❘
x
n
)
=
{
1
,
for
⁢
k
=
argmin
j
⁢

t
v
⁢
(
x
n
)
-
v
j

2
0
,
otherwise
(
10
)
the input of the decoder is defined as
t
q
⁢
(
x
n
)
=
v
k
(
11
)
where k is the index of the decoder input, and it is defined as
k
=
argmin
j
⁢

t
v
⁢
(
x
n
)
-
v
j

2
(
12
)
to address the problem of gradient collapse caused by introducing the implicit embedded space, we replicate the gradient ∇zL from the decoder network to the encoder network during the back-propagation;
when training the VQ-VAE, the loss function is defined as
L
=
log
⁢
p
⁢
(
x
n
❘
t
q
⁢
(
x
n
)
)
+

sg
[
t
v
⁢
(
x
n
)
]
-
v

2
2
+
λ
⁢

t
v
⁢
(
x
n
)
-
sg
[
v
]

2
2
,
(
13
)
where log p(xn|tq(xn)) is the reconstruction loss, aiming to optimize the encoder and decoder networks; since the back-propagation gradient is directly replicated to the encoder network, the loss log p(xn|tq(xn)) is not considered; in ∥sg[tv(xn)]−v∥22, L2 error is used to drive vk towards tv(xn), aiming to optimize the implicit embedded space; λ∥tn(xn)−sg[v]∥22 is to prevent the encoder output from exceeding the scope of the implicit embedding space, where λ depends on the reconstruction loss, and sg is the stop-gradient operator that is constant with the partial derivative of 0 during the forward propagation;
next, the log-likelihood function is defined as
log
⁢
p
⁢
(
x
n
)
≈
log
⁢
p
⁢
(
x
n
❘
t
q
⁢
(
x
n
)
)
⁢
p
⁢
(
t
q
⁢
(
x
n
)
)
(
14
)
according to Jensen's Inequality, Eq. (13) is rewritten as
log
⁢
p
⁢
(
x
n
)
≥
log
⁢
p
⁢
(
x
n
❘
t
q
⁢
(
x
n
)
)
⁢
p
⁢
(
t
q
⁢
(
x
n
)
)
.
(
15
)

### Claim 7 (depends on 1)

The Collaborative Caching Framework for Multi-edge Systems with Robust Federated Deep Learning according to claim 1, wherein first combine the model parameters from all MEC nodes into a matrix R∈M×θ, which is defined as
R
m
,
:
=

w
m
(*
)
-
w
m
(
0
)

M
(
16
)
w
m
(*
)
-
w
m
(
0
)
-
1
M
⁢
∑
m
=
1
M
∂
L
⁢
(
w
m
;
d
m
)
∂
w
(
17
)
w
m
(
0
)
-
w
m
(*
)
-
1
M
⁢
∑
m
=
1
M
∂
L
⁢
(
w
m
;
d
m
)
∂
w
(
18
)
where the local training model is parameterized by wm, wm(*) is the updated local model, and dm is the training data;
next, we arrange the elements of each column in R in descending order, retain their sorted positions, and transform them into {tilde over (R)}; for example, R(5.3, 6.7, 0.7, 0.4)→{tilde over (R)}(2, 1, 3, 4); specifically, the mean and standard deviation (STD) of {tilde over (R)} are defined as
mean
m
=
1
θ
⁢
∑
ϑ
=
1
θ
R
~
m
,
ϑ
(
19
)
std
m
=
1
θ
⁢
∑
ϑ
=
1
θ
⁢
(
R
~
m
,
ϑ
-
mean
m
)
2
(
20
)
following the mean and STD, we can divide the normal and adversarial model updates into two clusters through the K-means, where the adversarial model updates can be easily identified by the proposed residual-based detection; thus, the MEC nodes that offer normal model updates can be filtered, denoted by E′={e1, e2, . . . , eM′}; to avoid the model destruction caused by adversarial model updates, we design a similarity-based federated aggregation method; specifically, we adopt the canonical correlation analysis (CCA) to measure the similarity between the model updates of each MEC node and the average one, which determines the weights of different model updates when performing federated aggregation; this process is described as
w
r
+
1
=
(
1
-
β
)
⁢
w
r
+
β
⁢
∑
m
=
1
M
′
w
m
r
❘
"\[LeftBracketingBar]"
d
❘
"\[RightBracketingBar]"
*
τ
(
21
)
where τ indicates the similarity score.

### Claim 8 (depends on 1)

The Collaborative Caching Framework for Multi-edge Systems with Robust Federated Deep Learning according to claim 1, wherein based on the proposed RFDL, design a proactive cache replacement strategy with multi-edge collaboration; for each MEC node, initialize the cache space cachetemp and set of user intervals H through the multi-dimensional cache space partitioning; while cachetemp≥0, and the user-interest contents will be placed into the temporary cache library Ctemp; to avoid the cache redundancy caused by overlapping userinterest contents in different intervals, replace Ctemp by Cs that selects cacheh most popular contents in the current user interval hs from Ctemp; next, remove the duplicates in the cache library Cm on each MEC node and update the available cache space; the above steps will be iterated until the cache space is fully occupied.
