没有合适的资源?快使用搜索试试~ 我知道了~
首页深度学习中的图神经网络最新进展
深度学习中的图神经网络最新进展
需积分: 11 7 下载量 166 浏览量
更新于2024-07-15
收藏 24.38MB PDF 举报
"本次报告由Xavier Bresson副教授主讲,主要探讨了图神经网络(GNN)架构的最新进展,包括传统卷积网络与图卷积网络(GCN)的相关理论,以及一系列新兴的GNN模型和研究方向。报告内容详细介绍了GCN的背景、面临的挑战、发展历史,并深入剖析了各种GNN架构的设计理念和应用场景。"
在深度学习领域,图神经网络已经成为了处理非结构化数据,尤其是图数据的关键技术。GCNs是其中的一种重要类型,它借鉴了卷积神经网络(CNN)的思想,但应用于图结构上,以处理节点、边和整体图的特征。传统的卷积网络主要处理欧几里得空间的数据,如图像和声音,而图卷积则扩展了这一概念,使得网络能够处理更复杂的关系数据。
报告中提到了几种不同类型的图卷积网络:
1. 频域图卷积网络:这种网络通过傅立叶变换在频域中进行操作,允许对图的全局信息进行建模,适用于处理大规模图数据。
2. 空域图卷积网络:直接在图的顶点之间进行操作,通过邻居信息聚合来更新节点特征,这种方法更直观且易于实现。
此外,报告还涵盖了多种GNN变体和研究方向:
- Weisfeiler-Lehman GNNs:这是一种用于测试图同构性的图神经网络,通过迭代细化节点标签以提升模型区分不同图结构的能力。
- Graph Isomorphism Networks:进一步发展了Weisfeiler-Lehman测试,构建出能判断图是否同构的网络结构。
- Principal Neighbourhood Aggregation:基于主成分分析的邻域聚合方法,旨在捕捉节点局部结构的关键信息。
- Equivariant GNNs:这些网络保证了输入图的平移、旋转等变换下的不变性,增强了模型的泛化能力。
- 3-WL/Ring GNNs 和 SparseWL-GNNs:这两种网络提高了GNN的表达能力,尤其是对于复杂图结构的识别。
- Low-Rank Attention GNNs:通过低秩矩阵分解减少计算复杂度,同时保持模型性能。
- Graph Substructure Networks:关注图的子结构,以便更好地理解和建模复杂图的特性。
报告还讨论了GNN的表达能力,包括它们作为通用近似器的潜力,以及在Weisfeiler-Lehman测试下的表达性分析。此外,对于图的位置编码(如Index Positional Encodings、Laplacian Positional Encodings)以及利用边信息进行链接预测也是研究的热点。最后,GNNs for Sets探讨了GNN如何应用于无序集合数据的处理。
这场报告深入浅出地解析了图神经网络的最新进展,展示了GNN在数据科学和人工智能领域的广泛应用前景,同时也指出了当前研究的挑战和未来可能的发展趋势。
14
Principal Neighbourhood Aggregation
[1]
Xavier Bresson 14
Graph Isomorphism Networks
[1]
introduced GNNs as expressive as the WL test to distinguish
non-isomorphic graphs. Aggregation function of the form :
where function
g
is injective and sum aggregator is used (mean/max are not injective).
Principal Neighbourhood Aggregation (PNA)
[1]
: Generalization of the injective sum aggregator
function to multiple aggregators with degree-scalers.
Aggregators are mean, max, min and standard deviation.
Overall injective function (that guarantees GNN expressivity).
Increase the expressivity power of standard MP-GNNs while preserving O(
n
) complexity.
[1] Corso, Cavalleri, Beaini, Lio, and Velickovic, Principal neighbourhood aggregation for graph nets, 2020
[2] Xu, Hu, Leskovec, Jegelka, How powerful are graph neural networks?, 2019
f
NN
h
`
i
, {h
`
j
}
j2N
i
=(1+")g(h
`
i
)+
X
j2N
i
g(h
`
j
)
<latexit sha1_base64="gtYN93UwgILojgaKW5cWWSraCjc=">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</latexit>
<latexit sha1_base64="gtYN93UwgILojgaKW5cWWSraCjc=">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</latexit>
<latexit sha1_base64="gtYN93UwgILojgaKW5cWWSraCjc=">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</latexit>
<latexit sha1_base64="gtYN93UwgILojgaKW5cWWSraCjc=">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</latexit>
15
Principal Neighbourhood Aggregation
[1]
Xavier Bresson 15
Illustration of aggregators are mean, max, min and standard deviation :
PNA equation :
[1] Corso, Cavalleri, Beaini, Lio, and Velickovic, Principal neighbourhood aggregation for graph nets, 2020
X
`+1
i
= Concat
⇣
X
`
i
,
˜
X
`
i
W
`
1
⌘
W
`
2
2 R
d
`+1
˜
X
`
i
=
j2N
i
(X
`
i
,X
`
j
) 2 R
12⇥d
`
=
2
6
6
4
Mean
j2N
i
(X
j
)
Std Dev
j2N
i
(X
j
)
max
j2N
i
(X
j
)
min
j2N
i
(X
j
)
3
7
7
5
⌦
2
4
Id
S(d, 1)
S(d, 1)
3
5
,S(d, ↵)=
✓
log(d + 1)
◆
↵
<latexit sha1_base64="9Ntn4VGTdO+OZpLbPoYuN9MtPV0=">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</latexit>
<latexit sha1_base64="9Ntn4VGTdO+OZpLbPoYuN9MtPV0=">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</latexit>
<latexit sha1_base64="9Ntn4VGTdO+OZpLbPoYuN9MtPV0=">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</latexit>
<latexit sha1_base64="9Ntn4VGTdO+OZpLbPoYuN9MtPV0=">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</latexit>
Scalers that amplifies or
decreases amplitude of
incoming message.
16
Outline
Xavier Bresson 16
Weisfeiler-Lehman GNNs
Graph Isomorphism Networks
Principal Neighbourhood Aggregation
Equivariant GNNs
3-WL/Ring GNNs
Sparse WL-GNNs
Low-Rank Attention GNNs
Graph Substructure Networks
Expressivity Power and Universal Approximator
Expressivity with WL
Graph Universal Approximator
Limitations of Expressivity
Graph Positional Encodings
Index Positional Encodings
Structural Message Passing Networks
Laplacian Positional Encodings
Link Prediction with Edge Representation
GNNs for Sets
17
Graph Representations
Xavier Bresson 17
Graphs are rank-
2
tensors. They can be represented as
List of edges (short lists for sparse graphs):
E
=
[ {1,2}, {1,3}, {2,3}, {3,4} ]
Matrix (dense tensor):
Other tensors :
Vectors (ordered and unordered sets) are rank-
1
tensors :
General tensors (s.a. 3D-MRIs or hyper-graphs) of rank-
m
:
index 1
index 4
index 3
index 2
G =(V,E)
<latexit sha1_base64="239FBV4Lx15Q0Pwecwy+zRLKQeE=">AAAB/XicbVDLSgMxFL1TX7W+xsfOTbAIFaTMFEE3QkFElxXsA9qhZNK0Dc1khiQj1KH4K25cKOLW/3Dn35hpZ6GtBwKHc+7lnhw/4kxpx/m2ckvLK6tr+fXCxubW9o69u9dQYSwJrZOQh7LlY0U5E7Sumea0FUmKA5/Tpj+6Sv3mA5WKheJejyPqBXggWJ8RrI3UtQ86AdZDgnlyM0GXqNQ4vT4pdO2iU3amQIvEzUgRMtS69lenF5I4oEITjpVqu06kvQRLzQink0InVjTCZIQHtG2owAFVXjJNP0HHRumhfijNExpN1d8bCQ6UGge+mUyzqnkvFf/z2rHuX3gJE1GsqSCzQ/2YIx2itArUY5ISzceGYCKZyYrIEEtMtCksLcGd//IiaVTKrlN2786K1UpWRx4O4QhK4MI5VOEWalAHAo/wDK/wZj1ZL9a79TEbzVnZzj78gfX5A94Dk3c=</latexit>
<latexit sha1_base64="239FBV4Lx15Q0Pwecwy+zRLKQeE=">AAAB/XicbVDLSgMxFL1TX7W+xsfOTbAIFaTMFEE3QkFElxXsA9qhZNK0Dc1khiQj1KH4K25cKOLW/3Dn35hpZ6GtBwKHc+7lnhw/4kxpx/m2ckvLK6tr+fXCxubW9o69u9dQYSwJrZOQh7LlY0U5E7Sumea0FUmKA5/Tpj+6Sv3mA5WKheJejyPqBXggWJ8RrI3UtQ86AdZDgnlyM0GXqNQ4vT4pdO2iU3amQIvEzUgRMtS69lenF5I4oEITjpVqu06kvQRLzQink0InVjTCZIQHtG2owAFVXjJNP0HHRumhfijNExpN1d8bCQ6UGge+mUyzqnkvFf/z2rHuX3gJE1GsqSCzQ/2YIx2itArUY5ISzceGYCKZyYrIEEtMtCksLcGd//IiaVTKrlN2786K1UpWRx4O4QhK4MI5VOEWalAHAo/wDK/wZj1ZL9a79TEbzVnZzj78gfX5A94Dk3c=</latexit>
<latexit sha1_base64="239FBV4Lx15Q0Pwecwy+zRLKQeE=">AAAB/XicbVDLSgMxFL1TX7W+xsfOTbAIFaTMFEE3QkFElxXsA9qhZNK0Dc1khiQj1KH4K25cKOLW/3Dn35hpZ6GtBwKHc+7lnhw/4kxpx/m2ckvLK6tr+fXCxubW9o69u9dQYSwJrZOQh7LlY0U5E7Sumea0FUmKA5/Tpj+6Sv3mA5WKheJejyPqBXggWJ8RrI3UtQ86AdZDgnlyM0GXqNQ4vT4pdO2iU3amQIvEzUgRMtS69lenF5I4oEITjpVqu06kvQRLzQink0InVjTCZIQHtG2owAFVXjJNP0HHRumhfijNExpN1d8bCQ6UGge+mUyzqnkvFf/z2rHuX3gJE1GsqSCzQ/2YIx2itArUY5ISzceGYCKZyYrIEEtMtCksLcGd//IiaVTKrlN2786K1UpWRx4O4QhK4MI5VOEWalAHAo/wDK/wZj1ZL9a79TEbzVnZzj78gfX5A94Dk3c=</latexit>
<latexit sha1_base64="239FBV4Lx15Q0Pwecwy+zRLKQeE=">AAAB/XicbVDLSgMxFL1TX7W+xsfOTbAIFaTMFEE3QkFElxXsA9qhZNK0Dc1khiQj1KH4K25cKOLW/3Dn35hpZ6GtBwKHc+7lnhw/4kxpx/m2ckvLK6tr+fXCxubW9o69u9dQYSwJrZOQh7LlY0U5E7Sumea0FUmKA5/Tpj+6Sv3mA5WKheJejyPqBXggWJ8RrI3UtQ86AdZDgnlyM0GXqNQ4vT4pdO2iU3amQIvEzUgRMtS69lenF5I4oEITjpVqu06kvQRLzQink0InVjTCZIQHtG2owAFVXjJNP0HHRumhfijNExpN1d8bCQ6UGge+mUyzqnkvFf/z2rHuX3gJE1GsqSCzQ/2YIx2itArUY5ISzceGYCKZyYrIEEtMtCksLcGd//IiaVTKrlN2786K1UpWRx4O4QhK4MI5VOEWalAHAo/wDK/wZj1ZL9a79TEbzVnZzj78gfX5A94Dk3c=</latexit>
2 R
n
m
<latexit sha1_base64="z7CqSd7qrIzjKmZq9xLPmUXSFD8=">AAAB/XicbVDLSsNAFL3xWesrPnZugkVwVZIi6LLgxmUV+4AmLZPppB06MwkzE6GG4q+4caGIW//DnX/jpM1CWw8MHM65l3vmhAmjSrvut7Wyura+sVnaKm/v7O7t2weHLRWnEpMmjlksOyFShFFBmppqRjqJJIiHjLTD8XXutx+IVDQW93qSkICjoaARxUgbqW8f+1T4HOlRGGZ3014menxa7tsVt+rO4CwTryAVKNDo21/+IMYpJ0JjhpTqem6igwxJTTEj07KfKpIgPEZD0jVUIE5UkM3ST50zowycKJbmCe3M1N8bGeJKTXhoJvOgatHLxf+8bqqjqyCjIkk1EXh+KEqZo2Mnr8IZUEmwZhNDEJbUZHXwCEmEtSksL8Fb/PIyadWqnlv1bi8q9VpRRwlO4BTOwYNLqMMNNKAJGB7hGV7hzXqyXqx362M+umIVO0fwB9bnD59RlUM=</latexit>
<latexit sha1_base64="z7CqSd7qrIzjKmZq9xLPmUXSFD8=">AAAB/XicbVDLSsNAFL3xWesrPnZugkVwVZIi6LLgxmUV+4AmLZPppB06MwkzE6GG4q+4caGIW//DnX/jpM1CWw8MHM65l3vmhAmjSrvut7Wyura+sVnaKm/v7O7t2weHLRWnEpMmjlksOyFShFFBmppqRjqJJIiHjLTD8XXutx+IVDQW93qSkICjoaARxUgbqW8f+1T4HOlRGGZ3014menxa7tsVt+rO4CwTryAVKNDo21/+IMYpJ0JjhpTqem6igwxJTTEj07KfKpIgPEZD0jVUIE5UkM3ST50zowycKJbmCe3M1N8bGeJKTXhoJvOgatHLxf+8bqqjqyCjIkk1EXh+KEqZo2Mnr8IZUEmwZhNDEJbUZHXwCEmEtSksL8Fb/PIyadWqnlv1bi8q9VpRRwlO4BTOwYNLqMMNNKAJGB7hGV7hzXqyXqx362M+umIVO0fwB9bnD59RlUM=</latexit>
<latexit sha1_base64="z7CqSd7qrIzjKmZq9xLPmUXSFD8=">AAAB/XicbVDLSsNAFL3xWesrPnZugkVwVZIi6LLgxmUV+4AmLZPppB06MwkzE6GG4q+4caGIW//DnX/jpM1CWw8MHM65l3vmhAmjSrvut7Wyura+sVnaKm/v7O7t2weHLRWnEpMmjlksOyFShFFBmppqRjqJJIiHjLTD8XXutx+IVDQW93qSkICjoaARxUgbqW8f+1T4HOlRGGZ3014menxa7tsVt+rO4CwTryAVKNDo21/+IMYpJ0JjhpTqem6igwxJTTEj07KfKpIgPEZD0jVUIE5UkM3ST50zowycKJbmCe3M1N8bGeJKTXhoJvOgatHLxf+8bqqjqyCjIkk1EXh+KEqZo2Mnr8IZUEmwZhNDEJbUZHXwCEmEtSksL8Fb/PIyadWqnlv1bi8q9VpRRwlO4BTOwYNLqMMNNKAJGB7hGV7hzXqyXqx362M+umIVO0fwB9bnD59RlUM=</latexit>
<latexit sha1_base64="z7CqSd7qrIzjKmZq9xLPmUXSFD8=">AAAB/XicbVDLSsNAFL3xWesrPnZugkVwVZIi6LLgxmUV+4AmLZPppB06MwkzE6GG4q+4caGIW//DnX/jpM1CWw8MHM65l3vmhAmjSrvut7Wyura+sVnaKm/v7O7t2weHLRWnEpMmjlksOyFShFFBmppqRjqJJIiHjLTD8XXutx+IVDQW93qSkICjoaARxUgbqW8f+1T4HOlRGGZ3014menxa7tsVt+rO4CwTryAVKNDo21/+IMYpJ0JjhpTqem6igwxJTTEj07KfKpIgPEZD0jVUIE5UkM3ST50zowycKJbmCe3M1N8bGeJKTXhoJvOgatHLxf+8bqqjqyCjIkk1EXh+KEqZo2Mnr8IZUEmwZhNDEJbUZHXwCEmEtSksL8Fb/PIyadWqnlv1bi8q9VpRRwlO4BTOwYNLqMMNNKAJGB7hGV7hzXqyXqx362M+umIVO0fwB9bnD59RlUM=</latexit>
2
4
3
5
2 R
n
<latexit sha1_base64="H2ojkUv6xoIXh/dZUhqXHtKfwhI=">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</latexit>
<latexit sha1_base64="H2ojkUv6xoIXh/dZUhqXHtKfwhI=">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</latexit>
<latexit sha1_base64="H2ojkUv6xoIXh/dZUhqXHtKfwhI=">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</latexit>
<latexit sha1_base64="hP+6LrUf2d3tZaldqaQQvEKMXyw=">AAAB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i77+gsTLXTzQrMMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GSS7KDDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RRtRxzzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D77odBu3wMYA6nMMFXEEIN3AHD9CBLghI4BXevYn35n2suqp569LO4I+8zx84xIo4</latexit>
<latexit sha1_base64="hnJ+4vLmYL7ONXRI/h4zSG5Blzo=">AAACJ3icbVBNT9tAFHyG0qYmlMCVywqExCmyubRHJC70FiqSIMUmWm+ek1XWa2v3GSmy/Jt66S+p1AMcihB/hHWSSuVjtCuNZnY1701SKGkpCO68jc0PWx8/tT772+2dL7udvfbA5qUR2Be5ys11wi0qqbFPkhReFwZ5ligcJvPzxh/eorEy11e0KDDO+FTLVApOThp3vkcKUxr5UYJTqStuDF/Ulaj9KPp3UE/Wuh8ZOZ1R7EdSRxmnWZJUP+qbSteM+ePOUdANlmBvSbgmR7BGb9z5HU1yUWaoSShu7SgMCopdFEmh0IWVFgsu5nyKI0c1z9DG1XLlmh07ZcLS3LiriS3V/39UPLN2kSXuZTOofe014nveqKT0W1xJXZSEWqyC0lIxylnTH5tIg4LUwhEujHSzMjHjhgtyLTclhK9XfksGp90w6IaXAbTgAA7hBEL4CmdwAT3og4Cf8Af+woP3y7v3Hld1bXjr3vbhBbynZ8SwqGE=</latexit>
<latexit sha1_base64="hnJ+4vLmYL7ONXRI/h4zSG5Blzo=">AAACJ3icbVBNT9tAFHyG0qYmlMCVywqExCmyubRHJC70FiqSIMUmWm+ek1XWa2v3GSmy/Jt66S+p1AMcihB/hHWSSuVjtCuNZnY1701SKGkpCO68jc0PWx8/tT772+2dL7udvfbA5qUR2Be5ys11wi0qqbFPkhReFwZ5ligcJvPzxh/eorEy11e0KDDO+FTLVApOThp3vkcKUxr5UYJTqStuDF/Ulaj9KPp3UE/Wuh8ZOZ1R7EdSRxmnWZJUP+qbSteM+ePOUdANlmBvSbgmR7BGb9z5HU1yUWaoSShu7SgMCopdFEmh0IWVFgsu5nyKI0c1z9DG1XLlmh07ZcLS3LiriS3V/39UPLN2kSXuZTOofe014nveqKT0W1xJXZSEWqyC0lIxylnTH5tIg4LUwhEujHSzMjHjhgtyLTclhK9XfksGp90w6IaXAbTgAA7hBEL4CmdwAT3og4Cf8Af+woP3y7v3Hld1bXjr3vbhBbynZ8SwqGE=</latexit>
<latexit sha1_base64="qzp6qQPmWnjk4ReJk9tuCOUBTOA=">AAACMnicbVBNS8NAEN34WeNX1aOXxSJ4KokXPQpe9KZiq9DEstlO2sXNJuxOhBLym7z4SwQPelDEqz/CTY3g17ALj/dmeDMvyqQw6HmPztT0zOzcfGPBXVxaXlltrq13TZprDh2eylRfRsyAFAo6KFDCZaaBJZGEi+j6sNIvbkAbkapzHGcQJmyoRCw4Q0v1m8eBhBh7bhDBUKiCac3GZcFLNwi+HqhBzbuBFsMRhm4gVJAwHEVRcVZeFaqk1O03W17bmxT9C/watEhdJ/3mfTBIeZ6AQi6ZMT3fyzC0Vii4BGuWG8gYv2ZD6FmoWAImLCYnl3TbMgMap9p+hXTCfp8oWGLMOIlsZ7Wo+a1V5H9aL8d4PyyEynIExT+N4lxSTGmVHx0IDRzl2ALGtbC7Uj5imnG0KVch+L9P/gu6u23fa/unXutgt46jQTbJFtkhPtkjB+SInJAO4eSWPJBn8uLcOU/Oq/P22Trl1DMb5Ec57x8FLqnq</latexit>
<latexit sha1_base64="H2ojkUv6xoIXh/dZUhqXHtKfwhI=">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</latexit>
<latexit sha1_base64="H2ojkUv6xoIXh/dZUhqXHtKfwhI=">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</latexit>
<latexit sha1_base64="H2ojkUv6xoIXh/dZUhqXHtKfwhI=">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</latexit>
<latexit sha1_base64="H2ojkUv6xoIXh/dZUhqXHtKfwhI=">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</latexit>
<latexit sha1_base64="H2ojkUv6xoIXh/dZUhqXHtKfwhI=">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</latexit>
<latexit sha1_base64="H2ojkUv6xoIXh/dZUhqXHtKfwhI=">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</latexit>
A =
2
6
6
4
0110
1010
1101
0010
3
7
7
5
2 R
n⇥n
<latexit sha1_base64="IY45M3e7jKl2yOFMl66x3TQQ12Y=">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</latexit>
<latexit sha1_base64="IY45M3e7jKl2yOFMl66x3TQQ12Y=">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</latexit>
<latexit sha1_base64="IY45M3e7jKl2yOFMl66x3TQQ12Y=">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</latexit>
<latexit sha1_base64="IY45M3e7jKl2yOFMl66x3TQQ12Y=">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</latexit>
V 1 2 3 4
1
2
3
4
剩余80页未读,继续阅读
2019-11-12 上传
2019-11-27 上传
2023-07-02 上传
2024-07-01 上传
2023-07-25 上传
2023-08-21 上传
2023-09-14 上传
2023-09-29 上传
syp_net
- 粉丝: 158
- 资源: 1187
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
最新资源
- 高清艺术文字图标资源,PNG和ICO格式免费下载
- mui框架HTML5应用界面组件使用示例教程
- Vue.js开发利器:chrome-vue-devtools插件解析
- 掌握ElectronBrowserJS:打造跨平台电子应用
- 前端导师教程:构建与部署社交证明页面
- Java多线程与线程安全在断点续传中的实现
- 免Root一键卸载安卓预装应用教程
- 易语言实现高级表格滚动条完美控制技巧
- 超声波测距尺的源码实现
- 数据可视化与交互:构建易用的数据界面
- 实现Discourse外聘回复自动标记的简易插件
- 链表的头插法与尾插法实现及长度计算
- Playwright与Typescript及Mocha集成:自动化UI测试实践指南
- 128x128像素线性工具图标下载集合
- 易语言安装包程序增强版:智能导入与重复库过滤
- 利用AJAX与Spotify API在Google地图中探索世界音乐排行榜
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功