# Ninety-nine million Hubble cutouts, and eight hundred overlooked oddities.

AnomalyMatch processed 99.6 million Hubble Legacy Archive cutouts and the study reports 1,339 anomalies, more than 800 of them previously undocumented, with experts reviewing the candidates.

Canonical: https://brightaifuture.com/discoveries/hubble-anomalymatch
Format: discovery
Source publication: 2025-12-16
Bright publication: 2026-09-19
Substantive update: None recorded
Evidence and review: Demonstrated; confidence: high; approved; ai-assisted. AI-assisted editorial comparison with the cited primary sources, explicit evidence limits, and held alternatives. Publication authorized by the site owner on 2026-09-19; no human source review or independent replication is claimed.

## The human problem

Decades of telescope imagery contain unusual objects that researchers cannot inspect one cutout at a time.

## The prior constraint

Fully supervised searches find what has already been labeled, while unsupervised systems can overwhelm scientists with ambiguous outliers.

## AI’s actual role

Semi-supervised and active learning ranked unusual images and incorporated expert feedback.

## The documented result

The peer-reviewed paper reports 99.6 million cutouts processed, 1,339 anomalies, and more than 800 previously undocumented candidates.

## Why it may matter

The system expands what astronomers can notice while preserving the distinction between a flagged candidate and a scientific explanation.

## Limitations

An anomaly is a selection result, not a confirmed new phenomenon.

Candidates require catalog checks and follow-up observations.

Archive artifacts and rare imaging conditions can look unusual.

## Unresolved questions

Which candidates receive follow-up observations?

How many are artifacts or known classes after deeper review?

Can the active-learning process transfer to other archives?

## Provenance and history

{
  "dates": {
    "eventDate": "2026-01-27",
    "publicationDate": "2025-12-16",
    "captureDate": "2026-09-19",
    "lastReviewedDate": "2026-09-19"
  },
  "provenance": {
    "origin": "editorial",
    "externalId": "https://www.aanda.org/articles/aa/full_html/2025/12/aa55512-25/aa55512-25.html"
  },
  "revisions": [
    {
      "id": "revision:sept26-anomaly-01",
      "recordedAt": "2026-09-19",
      "summary": "Initial archive-discovery draft with candidate-status boundary.",
      "sourceIds": [
        "source-anomaly-esa",
        "source-anomaly-paper",
        "source-anomaly-preprint"
      ]
    }
  ],
  "corrections": []
}

## Original sources

- [Researchers discover hundreds of cosmic anomalies with help from AI](https://esahubble.org/news/heic2603/)
- [Identifying astrophysical anomalies in 99.6 million cutouts from the Hubble Legacy Archive using AnomalyMatch](https://www.aanda.org/articles/aa/full_html/2025/12/aa55512-25/aa55512-25.html)
- [Identifying Astrophysical Anomalies in 99.6 Million Cutouts from the Hubble Legacy Archive Using AnomalyMatch](https://arxiv.org/abs/2505.03508)

## Continue exploring

- [Frontier](https://brightaifuture.com/worlds/frontier)
- [What changes when computing leaves Earth?](https://brightaifuture.com/threads/space)
- [Where can human judgment go with a new instrument?](https://brightaifuture.com/threads/discovery)