Alibaba open-sources AI model that can detect cancer and nearly 150 conditions

plaidfuji•about 15 hours ago

It is 2026. How are we still publishing articles on medical diagnostics data science and using area under the ROC curve as the primary metric of success. ROC-AUC of 0.9 under severe class imbalance (almost always the case in diagnostics) could still mean something like 4/5 predicted diagnoses are wrong (false positives). Precision-Recall curve + mAP or GTFO.

Science article in question: https://www.science.org/doi/abs/10.1126/science.aec6129

Also, the most interesting result here is that the CNN-based feature encoder significantly outperformed a vision transformer encoder backbone…

steve-atx-7600•about 10 hours ago

Teaching students how to interpret evidence must be way undervalued still. I went to one of the top CS schools 20 years ago and you could get a degree without even taking a single probability or stats class of any kind.

chrisjj•about 14 hours ago

> How are we still publishing articles on medical diagnostics data science and using area under the ROC curve as the primary metric of success.

Waste avoidance.

Bullsh*t is more than sufficient to convince an AI-gulled target audience.

Zaraif13•about 13 hours ago

I don't agree with the flak Chinese labs get. If it's really that easy to distill and compete with frontier models, why aren't other countries anywhere near this AI race?

felixgallo•about 11 hours ago

Because distillation is a friendly term for industrial espionage, and most other countries are not willing to become international pariahs in the eyes of the west.

yogthos•about 7 hours ago

meanwhile in the real world https://www.science.org/content/article/china-tops-world-art...

ElProlactin•1 day ago

I much prefer Gemini. It tells me I'm very smart and almost always right.

vrighter•about 16 hours ago

you hit the nail right on the head!

t0bia_s•about 23 hours ago

I'm wonder when there will be model for predict future. Could be named "Crystal bAill"

mdp2021•about 22 hours ago

> model [to] predict future

The attempt is already there: "predicting the future" is called "intelligence".

pizzly•about 22 hours ago

Isn't detecting cancer predicting a possible future? Hopefully one you can change.

westurner•1 day ago

From https://news.ycombinator.com/item?id=44693991 :

> EPS3.9 also had significant anti-tumor effects in the mice with liver cancer and activated anti-tumor immune responses

“A Novel Exopolysaccharide, Highly Prevalent in Marine Spongiibacter, Triggers Pyroptosis to Exhibit Potent Anticancer Effects” (2025) DOI: 10.1096/fj.202500412R https://faseb.onlinelibrary.wiley.com/doi/10.1096/fj.2025004...

"A Gemma model helped discover a new potential cancer therapy pathway" https://news.ycombinator.com/item?id=45604231 :

> eCPMV VNPs + EPS3.9 + [...]

"Scientists are discovering a powerful new way to prevent cancer" https://news.ycombinator.com/item?id=45474404

Notes re: Kidneys not Livers: https://news.ycombinator.com/item?id=47460486 ; gh/topic/healthcare-ai

K0balt•1 day ago

Anyone know how many parameters?

UnboundedContex•about 22 hours ago

Approx ~131 million total. Made up of estimated:

102M - modified BERT-base-Chinese text encoder

26M - 3D U-Net-style vision/anatomy encoder

2.8M - projection layers, anatomy-specific projections, query tokens and attention layer

Written to run on something like an H100 though- as the CT Scan data is quite large.

anon373839•1 day ago

The model weights are only ~5GB, so this is small.

westurner•1 day ago

ScholarlyArticle: "Large-scale AI-guided liver malignancy diagnosis: multicenter study and a single-arm trial" (2026) Nature Medicine https://www.nature.com/articles/s41591-026-04589-y

Source: https://github.com/alibaba-damo-academy/damo-radar

Model: https://huggingface.co/radar-generalist

Alibaba open-sources AI model that can detect cancer and nearly 150 conditions

⚡ Community Insights

Discussion (21 Comments)Read Original on HackerNews