{
  "id": 19246713,
  "title": "Machine learning reveals how often fish schooling evolved—and repeatedly disappeared",
  "url": "https://urgent.news/2026/11/09/machine-learning-reveals-how-often-fish-schooling-evolved-and",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-09T23:30:05.000Z",
  "source": {
    "name": "Phys.org",
    "slug": "phys-org",
    "url": "https://phys.org/news/2026-11-machine-reveals-fish-schooling-evolved.html"
  },
  "original_language": "en",
  "account": "Scientists have employed new machine learning techniques to analyze the evolution of fish schooling across thousands of species. This study, published in the Proceedings of the National Academy of Sciences on November 6, is described as one of the most extensive examinations of any behavior in any group of species, according to the authors. The researchers found that fish species have gained the ability to move collectively some 60 times and lost it 320 times over the past 350 million years.\n\nAndrew Hein, an associate professor of computational biology, explained that species gain a trait through genetic mutations that favor fitness and survival. Once established, this trait can be inherited by subsequent generations, while it can be lost if it becomes disadvantageous. Despite initial expectations that losses of collective behavior would be rare, the study revealed that hundreds of lineages experienced a shift from collective movement to solitary individuals.\n\nThe team used advanced machine learning methods, including text transformers and image analysis techniques, to classify species behaviors and behaviors from citizen science images. This allowed them to quantify the phenotypes of fish schooling behaviors for 2,839 species, representing nearly 10% of the total 39,000 living fish species.\n\nThe study supports two main hypotheses for the evolution of collective movement. Firstly, it suggests that schooling may provide predator avoidance through collective defense, as species low in the food chain with many predators are more likely to engage in collective behavior. Secondly, the researchers found that schooling may have evolved to help individuals locate scarce resources, as predators that move in groups typically consume prey that also move in groups.\n\nDespite the benefits of collective movement, the researchers propose several potential drawbacks. Fish in groups may be more susceptible to misinformation, competition for limited resources, and disease risk, which could contribute to the loss of schooling behavior over time. These findings offer valuable insights into the complex dynamics of evolutionary behavior and the factors that drive the emergence and disappearance of collective movement in fish species.",
  "summary": "Scientists have used new machine learning techniques to investigate a topic no one has ever found a way to study: the evolution of fish moving together in schools across thousands of species. The new paper, published Nov. 6 in the Proceedings of the National Academy of Sciences, is among the largest studies of the evolution of any behavior ever conducted in any group of species, according to the…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 2,
    "also_reported_by": [
      {
        "outlet": "bioRxiv",
        "title": "Leave-One-Out-Cross-Validation Based Machine Learning Algorithms for Efficient, Scalable, and Robust Prediction of Artemisinin Resistance in Plasmodium falciparum Kelch13 mutants.",
        "url": "https://urgent.news/2026/11/09/leave-one-out-cross-validation-based-machine-learning-algorithms-for",
        "published": "2026-11-09T00:00:00.000Z"
      }
    ]
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}