AI tool that 'learns like a human' could be a game-changer for disease diagnosis

Imagine a world where tricky diseases could be spotted faster and more accurately. That's the promise of a clever new human-like AI tool. Researchers at The University of Hong Kong (HKU) have unveiled 'MorphoGenie', an innovative system designed to give doctors a helping hand.

Learning like us

This isn't just any AI. MorphoGenie analyses microscope images of individual cells to find tiny, meaningful patterns linked to their health, state, and how they behave. Crucially, it's designed to be interpretable, meaning researchers can see how it reaches its conclusions, not just what it predicts.

The tool learns a small set of reusable visual 'building blocks' from cell images. Think of it like a child learning to recognise shapes and then combining them to understand more complex objects. These blocks include things like cell size, shape, internal texture, and fine local details.

This groundbreaking concept is inspired by how humans learn – a fundamental principle in AI called compositionality. We don't learn everything from scratch; we build understanding by combining simpler ideas.

Professor Kevin Tsia, from HKU’s Faculty of Engineering, explained: "One of the long-term goals of AI is to build systems that learn from reusable concepts, rather than simply memorising patterns."

"MorphoGenie applies a similar principle to cell morphology. This helps make AI more transparent, adaptable and potentially more useful for future disease diagnostics," he added.

Unlocking hidden clues

Cell morphology, or the study of cell shape and structure, holds a wealth of biological information. However, much of it is tough to quantify consistently or even see with the naked eye. Traditional methods can be slow and biased.

MorphoGenie steps in by learning directly from cell images without needing manual labelling. It organises complex image information into a clear, understandable format. This could help classify cell states more objectively and uncover biological patterns we currently miss.

What can it do?

The HKU team has already shown MorphoGenie's impressive capabilities. It successfully distinguished major lung cancer cell subtypes. It also detected changes in cell morphology caused by drugs.

Beyond that, it tracked dynamic biological processes. These include cell-cycle progression and epithelial-to-mesenchymal transition. Both are closely linked to disease progression and how cancer spreads.

Dr Rashmi Sreeramachandra Murthy, the study's first author, said: "Cell images contain much richer information than what we can easily describe using conventional measurements alone."

"By learning interpretable visual primitives, MorphoGenie helps reveal meaningful biological patterns that might otherwise remain hidden. It still allows researchers to understand what the AI is using to interpret the data," she noted.

Why transparency matters

A huge advantage of MorphoGenie is its versatility. It works across different microscopy techniques, from label-free quantitative phase imaging to fluorescence microscopy. It can even transfer what it learns from one dataset to a completely new one.

This suggests massive potential for biomedicine, including disease research and drug discovery. The team believes MorphoGenie could be vital for the next generation of AI tools in this field.

As AI tackles more complex scientific challenges, the need for systems that scientists can understand and verify becomes crucial. MorphoGenie’s transparent approach to analysing biological images is a significant step forward.

Professor Tsia reiterated: "Interpretability is important not only for trust, but also for scientific usefulness."

"If AI is to help researchers find meaningful changes in cells, its findings need to be presented in a way people can understand. And crucially, verify," he concluded.

Related Content

MORE: New computer chip that 'thinks' like a human could slash AI energy use by 70% — https://trendwiremedia.com/2026/03/23/new-computer-chip-that-thinks-like-a-human-could-slash-ai-energy-use-by-70/

MORE: Robots could get human-like muscles in major breakthrough — https://trendwiremedia.com/2026/04/12/robots-could-get-human-like-muscles-in-major-breakthrough/

MORE: Payments boost for four million households on Universal Credit — https://trendwiremedia.com/2026/02/09/payments-boost-for-four-million-households-on-universal-credit/

OFFICIAL SOURCE VERIFICATION: This report is based on official data from University Newsroom. Document: HKU Engineering Researchers Unveil Human-Like AI Tool to Boost Disease Diagnosis Source Link: [Read the official report from University Newsroom](http://www.hku.hk/press/news_detail_29076.html)