Explore tweets tagged as #RadAIchat
T5. CLEAR comes with an explanations, elaborations and examples paper to facilitate its use #RadAIchat
https://t.co/AnCBwKpVe2
0
2
6
Our work on a new software for AI, the Generally Nuanced Deep Learning Framework (GaNDLF) has recently been accepted at Nat Comm Eng. The focus of GaNDLF is to enable zero/low code model training for healthcare. Find out more at https://t.co/6QIR4ohm0s 🧙🏽♂️ #RadAIchat
0
3
10
T2. Not all reporting guidelines are suitable for AI research in medical imaging #AI specific challenges require tailored guidelines, such as #CLAIM to ensure accurate reporting #RadAIchat
https://t.co/J2sEAp1ybm
0
6
17
T3. #CLAIM focuses on #AI research in medical imaging, while @TRIPODStatement +AI deals with #AI powered prediction models Both certainly aim to improve reporting, but CLAIM includes specific aspects of medical imaging #RadAIchat
0
2
10
Ensure the AI models give the "uncertainty" measures could potentially be a good starting point! #RadAIchat
1
3
12
If we move completely to value based medicine, perhaps AI tools could be valuable from population health perspective #RadAIchat
1
2
14
T5. The #EUAIAct adopts a risk-based approach, and defines four risk levels for AI systems Most medical AI systems will be categorized as high-risk systems and would require conformity assessment and post-market surveillance #RadAIchat @Radiology_AI @myESR @EuSoMII
1
1
15
A5. The US and EU have approached #AI regulation very differently. A medium article summarizes this well. https://t.co/c6WWQpbyNj
#radaichat
1
2
9
T1. Reporting guidelines help to standardize methodology and ensure #transparency, #reproducibility and #quality in research #RadAIchat
@Radiology_AI @EQUATORNetwork @GSCollins
0
2
8
Agree! #RadAIchat
@shanlivahdati @Radiology_AI @AidenceVeye Something that we should be cognizant about is the energy that went into tuning the model hyperparameters. Usually that is way more expensive than final model training. Still a very good trend to share environmental footprints in research papers! #RadAIchat
0
0
10
T5. The METhodological RadiomICs Score (METRICS), a quality assessment tool for radiomics research based on the CLEAR guideline, is now also available @EuSoMII #RadAIchat
https://t.co/xHN7Zsmqud
0
3
9
Important point! #radaichat
T4. CLAIM 2024 replaces 'ground truth' and 'gold standard' with 'reference standard'. This change acknowledges uncertainty in medical data labeling, aligns with other reporting guidelines like STARD, and avoids implying absolute certainty in benchmarks. #RadAIchat
0
2
9
T5. No, CLAIM is not suitable for radiomics research #RadAIchat
0
1
6
T1. Of course, knowing *which* reporting guideline to choose for your manuscript is key! https://t.co/foJWoPfIBC
#RadAIchat
0
3
9
Unfortunately, #AI related tools can suffer from biases, including assumptions made by their developers. In developing #CLAIM, we not only sought experts from diverse specialties, but from #medicine, #DataScience, #statistics, and journal editors #RadAIchat
0
2
10
A1. Regulation of #AI should focus on safe, trustworthy and unbiased #AI. It is important to remember that #AI models are not infallible. #radaichat
0
0
9