Bloom filter index
An index based on a Bloom filter — a probabilistic data structure that can quickly say "the row definitely does not match" or "maybe matches". In PostgreSQL an extension (bloom) good for tables with many columns queried in arbitrary combinations.
The problem it solves: if you have 10 columns and query different subsets, B-tree would require many indexes. A bloom index covers all columns compactly and filters out most non-matching rows; the few "maybe" hits are verified against the table. Trade-off: false positives (never false negatives) → good for reducing the search set, not for exact lookups. Also used internally: ClickHouse, Parquet, and LSM trees (RocksDB, Cassandra) have bloom filters to skip data blocks/SSTables without disk I/O. A tunable size controls the false-positive rate.