DARKER: Efficient Transformer with Data-driven Attention Mechanism for Time Series
Summary: DARKER approximates softmax attention with multiple data-driven, trainable kernel projections rather than a fixed random feature map. Its pIndex selects projections and an input index accelerates training, yielding linear sequence complexity and 3–4× speedups with competitive TSC accuracy. (summarized by gpt-5.6-luna on Jul 24 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Rundong Zuo (Hong Kong Baptist University)
- 2. Guozhong Li (King Abdullah University of Science and Technology)
- 3. Rui Cao (Hong Kong Baptist University)
- 4. Byron Choi (Hong Kong Baptist University)
- 5. Jianliang Xu (Hong Kong Baptist University)
- 6. Sourav S Bhowmick (Nanyang Technological University)
BibTeX Citation
@article{zuo_vldb24,
title = {{DARKER: Efficient Transformer with Data-driven Attention Mechanism for Time Series}},
author = {Zuo, Rundong and Li, Guozhong and Cao, Rui and Choi, Byron and Xu, Jianliang and Bhowmick, Sourav S},
journal = {PVLDB},
series = {{VLDB} '24},
volume = {17},
number = {11},
pages = {3229--3242},
doi = {10.14778/3681954.3681996},
url = {https://doi.org/10.14778/3681954.3681996},
year = {2024}
}
Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 14 of 14 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next