Till N.
@tillnitus.bsky.social
Doing stuff with big images and small computers for oncology research
Pinned
Till N.
@tillnitus.bsky.social
· Jan 10
Tissue concepts: Supervised foundation models in computational pathology
Due to the increasing workload of pathologists, the need for automation to support diagnostic tasks and quantitative biomarker evaluation is becoming …
www.sciencedirect.com
Happy to share our work on sample efficient foundation models. „Tissue Concepts“ shows better (or same) performance on cancer subtyping tasks compared to other foundation models. All while only using 6% of the training data. Find out how —>
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
Reposted by Till N.
Check out this amazing PhD position in Sweden. The PI is a really amazing person. Do you want to continue to apply AI to pathology? This may be the PhD for you.
Karolinska Institutet hiring 1-2 doktorander i beräkningspatologi och medicinsk artificiell intelligens in Stockholm, Stockholm County, Sweden | LinkedIn
Posted 2:40:20 PM. Vill du bidra till medicinsk forskning av toppkvalitet? Att vara doktorand innebär att vara antagen…See this and similar jobs on LinkedIn.
www.linkedin.com
January 24, 2025 at 12:27 PM
Check out this amazing PhD position in Sweden. The PI is a really amazing person. Do you want to continue to apply AI to pathology? This may be the PhD for you.
Reposted by Till N.
It's somewhat funny reading hot takes that generative AI doesn't pose any real ecological challenges because the marginal energy cost of calling simpler trained models is low.
January 23, 2025 at 2:07 PM
It's somewhat funny reading hot takes that generative AI doesn't pose any real ecological challenges because the marginal energy cost of calling simpler trained models is low.
Reposted by Till N.
Our paper on Tissue Concepts is finally out in JCompBiomed 🥳: Supervised pre-training of foundation models needs fewer data and less compute/energy to analyze pathology whole-slide images.
How? Read more here: doi.org/10.1016/j.co...
🧵1/5
How? Read more here: doi.org/10.1016/j.co...
🧵1/5
January 10, 2025 at 11:43 AM
Our paper on Tissue Concepts is finally out in JCompBiomed 🥳: Supervised pre-training of foundation models needs fewer data and less compute/energy to analyze pathology whole-slide images.
How? Read more here: doi.org/10.1016/j.co...
🧵1/5
How? Read more here: doi.org/10.1016/j.co...
🧵1/5
Happy to share our work on sample efficient foundation models. „Tissue Concepts“ shows better (or same) performance on cancer subtyping tasks compared to other foundation models. All while only using 6% of the training data. Find out how —>
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
Tissue concepts: Supervised foundation models in computational pathology
Due to the increasing workload of pathologists, the need for automation to support diagnostic tasks and quantitative biomarker evaluation is becoming …
www.sciencedirect.com
January 10, 2025 at 8:48 AM
Happy to share our work on sample efficient foundation models. „Tissue Concepts“ shows better (or same) performance on cancer subtyping tasks compared to other foundation models. All while only using 6% of the training data. Find out how —>
www.sciencedirect.com/science/arti...
www.sciencedirect.com/science/arti...
Reposted by Till N.
Our medical image foundation model is recently published in Nature Computational Science: "Overcoming data scarcity in biomedical imaging with a foundational multi-task model" 🎉.
We developed a *supervised* foundational training strategy and pretrained a model (UMedPT) …
We developed a *supervised* foundational training strategy and pretrained a model (UMedPT) …
November 14, 2024 at 9:33 PM
Our medical image foundation model is recently published in Nature Computational Science: "Overcoming data scarcity in biomedical imaging with a foundational multi-task model" 🎉.
We developed a *supervised* foundational training strategy and pretrained a model (UMedPT) …
We developed a *supervised* foundational training strategy and pretrained a model (UMedPT) …