Australian scientists use AI trained on animals to verify black hole image

A CSIRO team developed a bespoke imaging algorithm to confirm the ring structure of the M87* black hole.

An Australian team of scientists has used an AI model trained on images of cats, dogs, and automobiles to verify a global astronomical breakthrough.

The CSIRO team applied their new technique to data from the Event Horizon Telescope to validate the landmark image of the black hole at the centre of galaxy M87, located 55 million light years away.

The team developed a bespoke imaging algorithm called GenDIReCT and a new data product to confirm the ring shape of the black hole, which had been questioned by an independent group.

CSIRO astronomer and project leader Dr Nithyanandan Thyagarajan said the decision to train the model on terrestrial images rather than existing astronomical imagery removed any chance of bias.

The algorithm accounted for atmospheric errors and those produced by individual observatories, which had historically left room for data sets to be misinterpreted.

Dr Thyagarajan said the powerful imaging algorithm could be applied to areas such as seismic imaging and medical imaging with ultrasound or MRI.

Key facts

  • The CSIRO team developed a bespoke imaging algorithm named GenDIReCT.
  • The AI model was trained on terrestrial images of animals and automobiles rather than astronomical imagery.
  • The findings were published on Thursday in the journal Nature.

Three perspectives

Neutral

An Australian research team used a new AI imaging algorithm to validate the ring structure of the M87* black hole.

Positive

The development of GenDIReCT provides an elegant, independent method that strengthens confidence in black hole research and enriches the toolkit for the scientific community.

Negative

One independent group had previously questioned the validity of the black hole's ring structure, suggesting an alternative interpretation was equally valid.

Your week in news, every Wednesday

The biggest stories of the week for your edition, in your inbox every Wednesday morning, your time. Free, and you can unsubscribe at any time.

Choose your news coverage

Please choose and confirm your coverage.

All sections

Edition