AstroNote 2026-88

The Neural Engine for Discovering Luminous Events (NEEDLE) classifier (Sheng et al. 2024) is a machine-learning tool for identifying rare astronomical transients from real-time alerts. NEEDLE is trained to identify Superluminous Supernova (SLSN) and Tidal Disruption Events (TDE) candidates in data from the Zwicky Transient Facility (ZTF; Bellm et al. 2019; Graham et al. 2019). We are running NEEDLE daily on public ZTF alerts, streamed via the Lasair alert broker (Williams et al. 2024), and publicly providing classification predictions as annotations on Lasair.

We provide three SQL filters on Lasair with our candidates (also output as Kafka streams): NEEDLE SLSN candidates, NEEDLE TDE candidates, and all NEEDLE annotations. Lasair users can also incorporate NEEDLE outputs into their own custom SQL filters.

We manually vet all alerts with a SLSN or TDE score > 0.5, and will highlight the most promising candidates in these AstroNotes. A complete set of candidates can be found in the Lasair streams. This report includes recent SLSN and TDE candidates, shown below. We encourage spectroscopic classification and follow-up.