Info: Zenodo’s user support line is staffed on regular business days between Dec 23 and Jan 5. Response times may be slightly longer than normal.

There is a newer version of the record available.

Published June 3, 2024 | Version 1.0.0
Dataset Open

Towards Neural Scaling Laws for Foundation Models on Temporal Graphs

  • 1. ROR icon McGill University
  • 2. ROR icon Mila - Quebec Artificial Intelligence Institute
  • 3. ROR icon University of Manitoba
  • 4. ROR icon The University of Texas at Dallas
  • 5. ROR icon Université de Montréal
  • 6. ROR icon Canadian Institute for Advanced Research
  • 7. ROR icon University of Central Florida

Description

Datasets provided in this storage are introduced in the paper: Towards Neural Scaling Laws for Foundation Models on Temporal Graphs

  • Each .csv file represents all transactions of the token network that has the same name as the file name (<tokenname.csv>)
  • Each transaction corresponds to a row in each file.
  • Each transaction has:
    •  blockNumber :  is the block ID of Ethereum that includes this transaction
    • timestamp: time that the transaction is made in UNIX timestamp format
    • tokenAddress : the address that specifies a unique ERC20 token
    • from: address of sender
    • to: address of receiver
    • value: the amount the transaction
    • fileBlock: we split the whole number of blocks count to 35 buckets and assigned the bucket ID to the transaction to trace the blocks 
  • Raw .csv will be used to divide into generate edgeslist and label, which indicates all node interactions and labels for each snapshot respectively, with help of functions from TGS_Handler defined in TGS.py inside the TGS package (see the code in the Github repository provided along with this data storage)



Abstract

The field of temporal graph learning aims to learn from evolving network data to forecast future interactions. Given a collection of observed temporal graphs, is it possible to predict the evolution of an unseen network from the same domain?
To answer this question, we first present the Temporal Graph Scaling(TGS) dataset, a large collection of temporal graphs, consisting of eighty-four ERC20 token transaction networks collected from 2017 to 2023. Next, we evaluate the transferability of Temporal Graph Neural Networks(TGNNs) for the temporal graph property prediction task by pre-training on a collection of up to sixty-four token transaction networks and then evaluate their downstream performance on twenty unseen token types. We observe that the neural scaling law observed in NLP and Computer Vision also applies in temporal graph learning where pre-training on more networks and more parameters leads to better downstream performance. To the best of our knowledge, this is the first time that the transferability of temporal graphs has been shown empirically. Notably, on the downstream token networks, the largest pre-trained model outperforms fine-tuned TGNNs on thirteen unseen test networks. Therefore, we believe this is a promising first step towards building foundation models for temporal graphs. The code and datasets are publicly available at https://github.com/benjaminnNgo/ScalingTGNs under MIT licence.

 

Files

0x0.csv

Files (3.4 GB)

Name Size Download all
md5:a9582a58b96a52cafd57ebee54d2f40c
30.0 MB Preview Download
md5:4771652a3960da2b36927ed474052bad
32.7 MB Preview Download
md5:015d78f3355566b460c8593e3ed7d6a3
21.8 MB Preview Download
md5:adcffe9f4da97c596e9248769a21103e
21.1 MB Preview Download
md5:23d03fea4bbafd57e9d202b9118f45dc
76.5 MB Preview Download
md5:e24f5cd5876b323a307fbbc5571e7344
35.1 MB Preview Download
md5:ce796db83bc28382e9e3ac8f6198aa29
12.5 MB Preview Download
md5:a2e597c55da4b093708c4e86ce46de74
63.7 MB Preview Download
md5:5d3c5f8bb0cf8ef395945fd271f18cad
77.5 MB Preview Download
md5:c8ac5c3ca9fbfc67a32e9f7e835b81c2
21.6 MB Preview Download
md5:170948181c9ae186a3d7e7c59aa441f1
16.6 MB Preview Download
md5:7095bdaf8fed09f5c08d21b7fe65b2a1
21.4 MB Preview Download
md5:ff7e1362a18b9c0aa0a44253abc16259
57.9 MB Preview Download
md5:595f8191db9b87cfabdc818afc20184a
38.1 MB Preview Download
md5:ebc549baa014362c792bb38980fbc1ea
18.7 MB Preview Download
md5:db9165aacaa839bd185c2507f6196b45
59.4 MB Preview Download
md5:52ebe3792d2692541288bda2eb5610b3
41.6 MB Preview Download
md5:578410bd83675bd974dda22969d5d644
35.8 MB Preview Download
md5:4eabdcdfe036c6c9903b3931780936dd
20.6 MB Preview Download
md5:1e90d8baff0144f74f9386b1722b3fef
15.6 MB Preview Download
md5:be577f6abf3e866194d8c90dbafed5d8
19.5 MB Preview Download
md5:8013903049cac187ac723baab5bee3ea
16.6 MB Preview Download
md5:b8ee83f9e1586f1318f017aa482bd5ca
68.3 MB Preview Download
md5:e0426f55bcc86fec94bcb60e825efa76
13.7 MB Preview Download
md5:a025a5d7aa295632b2e9078f4435a9af
40.6 MB Preview Download
md5:60b4bef9d3f2610a44ea82841e9f9147
59.5 MB Preview Download
md5:ad1408e40ec4cc60872935bf4d581dfc
34.7 MB Preview Download
md5:6886731e182f3f89adce18d62f65d1ac
14.1 MB Preview Download
md5:6f604d43cd8fb72fdf7c0ea0b5628f3d
64.1 MB Preview Download
md5:4318235fa64b455546520fb7707afdea
31.8 MB Preview Download
md5:3e04988723a1999cce64e616c2dc3603
41.4 MB Preview Download
md5:f9bc7be410dda73aa159a86ae01613ff
13.9 MB Preview Download
md5:3868d7944e1e5215493255bc4375a79c
18.0 MB Preview Download
md5:6e8d0a6b0657cd8cde3e32a96b557f0f
54.7 MB Preview Download
md5:3c607af90da2e126ef229d1d86e2a3cc
12.1 MB Preview Download
md5:5d1db14a740d7b5d512f16b1654edcb7
50.2 MB Preview Download
md5:60890a3b1b4102352c1012a078fb47b4
85.4 MB Preview Download
md5:790ba6543771309e5b2128c69069a01f
35.1 MB Preview Download
md5:5a9fd96cf706f0c2fb1fc8608681fa02
40.2 MB Preview Download
md5:1f06668d9e0b7a3b44b678ad84b39912
65.1 MB Preview Download
md5:f6bb8db1367b4fc361d6bd988ef4c699
75.3 MB Preview Download
md5:52c0354d42fdb1c94d8c1992d664a4b1
16.8 MB Preview Download
md5:c70438a531a3beab45699230d784f3fb
59.7 MB Preview Download
md5:dbaaeb1c4af53a98d3153a027a10e051
47.5 MB Preview Download
md5:b18dcd11e67d4575a151950eb36b6f63
78.1 MB Preview Download
md5:52c43356437d0eb4a7fc3f172c553073
40.7 MB Preview Download
md5:70e4e76a51c66598bfef9e305c867ca7
14.4 MB Preview Download
md5:fec951072c7ccc0678f9867b01261d7e
32.7 MB Preview Download
md5:c50b472a969d5e84fcf6fb819ebbab50
62.3 MB Preview Download
md5:d85a00932f6bf8c9b6b275e03cf3da9d
39.5 MB Preview Download
md5:e0b0c7ec9b220cdd15dd7d32212f4c6b
74.5 MB Preview Download
md5:56b32eebf73031bf57fbae317fd4bfec
97.3 MB Preview Download
md5:5a412bdbe7bd7e9cd32b049342f73b9c
20.4 MB Preview Download
md5:9433ddadf24cd5e2742a4b8460f9fe2b
20.9 MB Preview Download
md5:4dd111e82e8e46e422d91ab44489d0f5
30.0 MB Preview Download
md5:6bd136262477a87dbc3cd4d9b9fa5c93
16.3 MB Preview Download
md5:7a84733d2dfffa03b4a0520ea7dc3882
21.6 MB Preview Download
md5:d16595bb0d2229fe771e8f251ea70dfb
52.5 MB Preview Download
md5:499aaf2cf42dd45dfbda19c38c02c543
87.7 MB Preview Download
md5:dadc8323400353bfeedd8900eca7642f
38.5 MB Preview Download
md5:068ea6a6051e29cf4c68d96c1d44fb7a
33.4 MB Preview Download
md5:10f89e730531034e4ecc4bba4f39ffed
31.3 MB Preview Download
md5:681a6db766c33c7f0902bbc3c2931196
37.1 MB Preview Download
md5:6786fe5f9009d0e60f3792b586ece04f
42.4 MB Preview Download
md5:bac6e32f2f43ac086630e447527a1916
36.1 MB Preview Download
md5:290410fc80a48847df1690408b195489
18.6 MB Preview Download
md5:c04014d5d0e420e94e41b5b2db0f4b43
24.4 MB Preview Download
md5:9abb8be69be45400772e9ce0457aa99c
16.1 MB Preview Download
md5:539ac59895b97341042f89bb0d7adfa1
15.9 MB Preview Download
md5:55925e53ee31b629d87abbae56278a18
16.1 MB Preview Download
md5:5b131d7601e2b71bab5a678c75f3dccd
66.8 MB Preview Download
md5:e4ef91f720c0df0cc56bcb5b3fc1d2e7
36.5 MB Preview Download
md5:453265ac61ae417602c835e60928a33d
19.2 MB Preview Download
md5:f731d885c4b0d27576f9626a99defce3
88.0 MB Preview Download
md5:1dfa9eb115765ae607f70d192f585b39
89.3 MB Preview Download
md5:660560ffb8103896184b124075390625
52.0 MB Preview Download
md5:bde8071199ecee33fed73a95a867ac04
52.5 MB Preview Download
md5:06ad706e2fa7acdc0eb891d6714d93f4
33.8 MB Preview Download
md5:86b4938697ff5ac3d48aac7bb3fb9c30
36.7 MB Preview Download
md5:e838d8d9d8d5ea37c3b7161d6c49f960
35.6 MB Preview Download
md5:ec3b36fd754e0f73b6133a2b27217ff4
61.9 MB Preview Download
md5:00f7434fc1e84173694eeb827ee86b9b
18.5 MB Preview Download
md5:ff490bd325ce194ffe1824388cb5f815
39.0 MB Preview Download
md5:0962bbad6d851754ca3a5def597b9d41
67.5 MB Preview Download

Additional details

Dates

Submitted
2024-06-05
NeurIPS 2024 Datasets and Benchmarks Track

Software

Repository URL
https://github.com/benjaminnNgo/ScalingTGNs
Programming language
Python
Development Status
Active