Data Science MS Capstone Class

ERIC TOPOL AND PRANAV RAJPURKAR

FEB 02, 2025

https://shorturl.at/tgdDo

“Our op-ed in today’s NY Times explores an unexpected finding. A series of recent studies compared the performance of doctors with A.I. versus A.I. alone, spanning medical scans, diagnostic accuracy, and management reasoning. Surprisingly, in many cases, A.I. systems working independently performed better than when combined with physician input. This pattern emerged consistently across different medical tasks, from chest X-ray and mammography interpretation to clinical decision-making. In some of the studies … the gap for performance favoring A.I. alone was large.”

 

 

 

 

 

February 1, 2025

Sharath Chandra Kovi: Initially, we were all excited about creating a large language model (LLM) for our project. But after looking into the training costs and other complexities, it seems like it might be too ambitious for us. What if we focus on creating a small language model instead
Naresh Narla: I agree, Sharath. The training costs for an LLM are indeed high, and we might face challenges with computational resources. A small language model is more feasible within our timeframe and resources. We can still make a significant impact with it.
Sri Lakshmi: Good point, Naresh. Plus, a smaller model will be easier to manage and fine-tune. We can focus on specific applications in the medical field, like diagnostic support or analyzing medical literature. This way, we can still achieve our goals without the added complexity.
Sharath Reddy Badam: I think we’re all on the same page. Creating a small language model sounds like a practical and effective approach. Let’s make sure we clearly define our objectives and roles so we can work efficiently towards our goal.

February 1, 2025

Venkatapraveen Reddy Adireddy, Durga Narasaiah Naidu Kilaru and  Sangepavu Jaswanth Kumar, have formed a group, and are working together on the “Sign Language Application” project. 

This week, we focused on finding “reliable and useful data sources” for our project. We came across multiple datasets that could be helpful. However, since they are quite large, we are still analyzing them to choose the one that best fits our needs. 

We will continue working on this and update once we finalize our dataset choice.

 

This is a posts page where any of us can post items of interest, such as links to data sets we have found, comments on, or suggestions for, other’s projects, progress reports, or issues arising from analysis.