Vishwa Parekh
@vishwa_parekh
Followers
49
Following
82
Media
0
Statuses
16
Our #sustainability🌎& #radiology #AI paper is out in @radiology_rsna! Consider energy when using LLMs @Rads4SF Kudos to a fantastic team @dharmam_savani @AdwayKanhere @ruthcarlosmd DrAnupam Joshi @umbccsee @PaulYiMD @vishwa_parekh Thank you for the highlight @DrRichardDo🙌
Nice work @flo_doo ! Optimal Large Language Model Characteristics to Balance Accuracy and Energy Use for Sustainable Medical Applications | Radiology
1
8
35
Check out this study that shows how smaller LLMs used in medical applications may have a reduced carbon footprint when compared with more general-purpose larger LLMs. https://t.co/JNwI6jBs3H
@UM2ii @flo_doo @dharmam_savani @AdwayKanhere @ruthcarlosmd @PaulYiMD @vishwa_parekh
pubs.rsna.org
This study explores the balance between accuracy and energy use in different-sized large language models for a medical application, highlighting the importance of model selection to balance perform...
0
3
9
👏🏻 Congratulations to @AdwayKanhere, Software Engineer at the @UofMaryland, for winning the Best Poster Award at the @CVPR 2024 Workshop: Data Curation and Augmentation in Medical Imaging! 🌼 🚀 Workshop: https://t.co/bBsCUeKrgr 📖 Paper: https://t.co/Y6kOgKBKMe
0
3
11
Want to learn more? Check out our paper on arXiv ( https://t.co/bjrilWOc7C) or visit our Github repository to access the code ( https://t.co/OOsq7BDlYQ). And be sure to follow us @UM2ii for the latest updates and news on Text2Cohort!
lnkd.in
This link will take you to a page that’s not on LinkedIn
0
1
3
Excited to announce the release of Text2Cohort to allow users to interact with the imaging data commons (IDC) using natural language! Simply ask a query like “download all male brain MRIs for patients under the age of 25 across all relevant IDC collections” @UM2ii @CancerIDC
1
1
3
Another one from the amazing @SamSantomartino! More to come from this rising ⭐️. @UM2ii @flo_doo @vishwa_parekh @PeterKamelMD
Code-free #DeepLearning platforms performed well in binary CXR classification, but need further development before clinical implementation https://t.co/KUwNo2N2S8
@SamSantomartino @NimaHNejad @PaulYiMD #chestxray #CFDL #MachineLearning #ECR2023
1
4
13
Code-free #DeepLearning platforms performed well in binary CXR classification, but need further development before clinical implementation https://t.co/KUwNo2N2S8
@SamSantomartino @vishwa_parekh @UM2ii #CFDL #AI #MachineLearning
1
1
6
Great way to close out #RSNA2022 with an awesome talk by @UMMCRadiology resident @EBeheshtian on evaluation of bias and fairness in radiology #AI! Check out the publication in @radiology_rsna! @RSNA @UM2ii
0
6
47
Fantastic start to #RSNA2022 with a talk by med student extraordinaire @SamSantomartino on clinic readiness of #AI in #radiology! Look out for this future ⭐️! @UM2ii @flo_doo @RSNA @vishwa_parekh
1
6
34
Check out our research demo @awscloud at #RSNA22! Learn about @UM2ii’s work on imaging progressive resolution for #AI and #ML! @chafey @AdwayKanhere @itsPranavK @vishwa_parekh @flo_doo
0
4
9
Very excited to partner with @awscloud to develop the next generation of medical image streaming technologies. https://t.co/GZJ33NtDzp
@PaulYiMD @vishwa_parekh @PeterKamelMD @AdwayKanhere @itsPranavK
aws.amazon.com
Store, analyze, and share medical images in the cloud at petabyte scale.
0
4
8
Introducing @UM2ii's official website✨. Check it out at https://t.co/vy4lnY1XLP now! @PaulYiMD @vishwa_parekh @PeterKamelMD @AdwayKanhere @itsPranavK
0
6
8
Phenomenal work by @UM2ii SWE extraordinaire @itsPranavK! More to come on #federatedlearning from our lab directed by @vishwa_parekh and @PaulYiMD.
1
2
6
Swapna Burman, 6 toes in both legs, no money for special shoes. Father a rickshaw puller, mother working in a tea garden, running for her country in the final event of the Heptathlon with a bandaged jaw. And winning it for India This country certainly has no shortage of heroines
670
15K
42K
It's World Immunization Week! #VaccinesWork to save lives. Vaccines prevent >2 million deaths every yr – @WHO
https://t.co/04bmXBfpt2
7
301
307