Database Paper Browser

Back to papers

PreLog: A Pre-trained Model for Log Analytics

Summary: PreLog: foundation model for log analytics, pre-trained on unlabeled heterogeneous logs with log-specific entry/sequence objectives to capture semantics/structure. Prompt tuning casts downstream parsing and anomaly detection into the pretraining format, yielding unified, cost-effective transfer across tasks. (summarized by gpt-5.4-mini on May 24 2026)

Paper ID
6927
Venue
SIGMOD
Year
2024
Pagerank
4.8925595e-05
Overall Rank
6,897 | 52.02%
DOI
10.1145/3654966

Incoming Non-self Citations Over Time

Authors

Incoming Citations (Sorted by Pagerank)

Showing 1 of 1 citing papers.

Rank Citing Paper Year Venue Pagerank
10,218 Unseen Anomaly Detection from System Logs 2026 SIGMOD 4.1945683e-05
Previous Page 1 / 1 Next

Outgoing Citations (Sorted by Pagerank)

Showing 5 of 5 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Previous Page 1 / 1 Next

Semantically Similar Papers