Team: 26
School: Justice Code/International/Harrison
Area of Science: Computer Science
Interim:
Our project problem is an inquiry about whether or not it is possible to code an A.I. that can accurately identify and diagnose diseases. Would it be plausible to create or code an A.I. that can accurately recognize and diagnose disease in a medical setting? A.I. for this purpose does exist but is incredibly hard to integrate into the rigorous environment of a hospital, especially to the point where it can diagnose diseases faster than a human being. We hope to build an A. I that is capable of diagnosis after being given symptoms, medical history, ethnicity/race, and other variables. To code the AI needed to accurately identify and diagnose a wide range of ailments our team will have to compile a reliable and unbiased data set to "train" the artificial intelligence with. According to Plat AI having “clean†and unbiased data is essential for creating a good AI. (Melkonyan) Unbiased data is more important than the actual AI itself as biases in a medical environment are life-threatening. The language our team will use to create this AI is Python as it is easy to learn and understand and has many functions that would be useful for coding a working AI. We think that AI is the next step in making healthcare accessible to everyone. According to a study called The potential for artificial intelligence in healthcare, “There are already a number of research studies suggesting that AI can perform as well as or better than humans at key healthcare tasks, such as diagnosing disease. Today, algorithms are already outperforming radiologists at spotting malignant tumors, and guiding researchers in how to construct cohorts for costly clinical trials.â€(Davenport and Kalakota) Those who cannot afford to see a doctor can simply input their information into an AI, this would streamline the treatment process and make helping patients cheaper and less time-consuming. As of now, this project is still in the research phase. According to an article written by Zulaikha Lateef.
This will be a reactive machine AI. This means that it will only be able to make decisions based on the information given to it and cannot use this information to make informed decisions in the future; this would protect patient confidentiality among other things. This is expected to be a multiple-year project because our team is still in the process of learning Python. We have taken a few lessons but are still beginners in the language. Near the end of this project, we hope to have a cohesive model that can identify a range of diseases varying in difficulty to diagnose. For example, according to Magmutual, cases of Appendicitis are notoriously difficult to diagnose despite being a life-threatening disease, on the other hand, something like the common cold can be diagnosed by the average person. In an article published by AMCC, “A number of rare diseases don’t even have diagnostic criteria†(Breining) Several factors must come into consideration when we try to create a model for this purpose. Beyond the demographics listed earlier, these factors would include things like user error, and disease type( like the difference between a viral and bacterial disease). We would need to either find or compile an accurate database that the AI will use to find the disease that most accurately matches the symptoms through a process of elimination. We also plan to make the A.I. compatible with software commonly used in hospitals.
Works Cited
AAMC. Apr. 2015, www.aamc.org/news-insights/rare-diseases-difficult-diagnose-cures-hard-come. Accessed 21 Jan. 2023.
Davenport, Thomas, and Ravi Kalakota. "The Potential for Artificial Intelligence in Healthcare." Future Healthcare Journal, vol. 6, no. 2, June 2019, pp. 94-98, https://doi.org/10.7861/futurehosp.6-2-94. Accessed 21 Jan. 2023.
Edureka. 3 Jan. 2023, www.edureka.co/blog/types-of-artificial-intelligence/. Accessed 21 Jan. 2023.
Magmutual. www.magmutual.com/learning/article/top-ten-hard-diagnose-diseases/. Accessed 21 Jan. 2023.
PlatAI. Feb. 2022, plat.ai/blog/how-to-build-ai/. Accessed 21 Jan. 2023.
Team Members:
Mekhi Bradford
Kingsley Walker
Chinyere Offor
Noel Emma-Asonye
Sponsoring Teacher: Caia Brown