User:Khatami

From REU@MU
Revision as of 00:46, 13 June 2023 by Khatami (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

2023 Week 1: (May 30th 2023-June 2nd 2023)

Tuesday (5/30)

  • Atteded REU Orientation
  • Got miliar with mentors and cohort
  • Met with Dr. Madiraju to learn about my project Assignment
  • Contacted IT to activate Marquette ID and Marquette email

Wednesday (5/31)

  • Attended Good Research Practice session
  • Chatted with Dr. Madiraju about project assignment once more
  • Met Sajjad, the graduate student active on PEER SURE App and discussed how to build an annotated bibliography as well as important key words and topics which are most revleant to our current research.
  • Met with Christian, another REU cohort member, who will also be working on this project.
  • Reviewed the notes and slides from Dr. Brylow's Good Research Practice session
  • Some free writing to clear my mind on this project:
    • So far I have prepared a list of main priorities for my literature review and the necessary tools for this project
    • 1.Becoming more familiar with OUD field and studying the efficacy, the format and uses cases for Peer Support in clinical treatment, primary care and eHealth
    • 2.Reviewing the existing web applications and platforms that offer peer support for substance use disorders
    • 3.Searching for resources to learn how to use C# and generative AI tools and frameworks for web development

Thursday(6/1)

  • Initiated a literature review on generative AI for SUD to identify the best practices and the gaps in the field
    • Benefits of peer support groups in the treatment of addiction https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5047716/
    • Lived Experience in New Models of Care for Substance Use Disorder https://www.frontiersin.org/articles/10.3389/fpsyg.2019.01052/full
    • Qualitative Study of Addiction Peer Mentorship in the Hospital https://pubmed.ncbi.nlm.nih.gov/31512181/
    • These papers are not geenrally focused on web-based Peer Support. However, they were useful for understanding the definition, history, efficay, and formarts of Peer Support. Based on my readings, it appears that Peer Support has been historically based in clinical setting and only since recently (since 2011) there are a few papers mentioning web-based Peer Support. The overall approach to use Peer Support at eHealth and primary care setting is fairely new . I have gathered a few more focused paper on web-based forms of Peer Support which I will begin reading soon.

Friday(6/2)

  • Opening a page in Marquette Wiki for Project milestones and goals
  • Searching for tutorial for using Open AI API as well as how to How to connect to ChatGPT with C#

2023 Week 2: (June 5th 2023-June 9th 2023)

Monday (6/5)

  • Attended RCR Training with Dr. Brylow
  • Worked on online RCR modules

Tuesday (6/6)

  • Attended REU Group Collaboration
  • Continued and finished working on RCR modules
  • Began and finished DL2 modules

Wednesday (6/7)


Thursday (6/8)

  • Finished reading the papers with sepcific focus on web-based Peer Support from yesterday
  • Some general findings from other papers emerged after I finished and reviewed the papers from yesterday and last week. These could be useful for this project:
    • Peer Support mobile or web based applications use a survey periodically to check for symptoms of opiod (or alcohol) addiction).
    • Traditional forms of treatment for susbstance abuse face challegnes with engaging and retaining the patients through the treatments
    • Peer Support offers a solution by connecting peers with lived expreinces of susbstance abuse with those who are trying to manage and control their addition.
    • Additionally a sustained recovery from abused susbstances require constant check ups and following up with treatments which Peer Support groups may help with by providing a community and an oppurnity to seek advice, encouragment, proper treatment without the fear stigma.
    • Many patients experiencing OUD seek advice and information from online platforms and social media which can be harmful and lead them to misinformation
    • Additionally challegnes listed in Peer Support online spaces include:
    • Lack of effective moderation
    • Lack of common Security
    • Web or mobile based peer support applications, therefore, can fill these gaps by directly reaching for patients and provding them with accurate information while removing the harmful social isolation which brings many to susbstance abuse in the first place and can draw back those who are recovering to using again.
    • Other papers also acknowledge that there is very little known what kind of online Peer Services can be most helpful and therefore more research in this field is required
    • Some of these applications are more focused on moderated chat between the peers and mentors while others also offer modules and training courses to convey infromation about symptom management, path to recovery, and accurate medical advice.
    • PEER SURE App should have a detailed guidance documentation about the specific services it is offering.
    • Additionally a list of expected features and functionalities of the app will be helpful to focus the energy on research and development towards most important features.
    • Users of other Peer Support apps have reccomended the peer support application should monitor for:
    • Talks for recurrence of drug use
    • Medical advice
    • Language that indicates racism, homophibia, and cyber victimization and bullying
    • There is a lack of research on Peer Support applications, particularly on those that integrate NLP features. Though I found a research paper that focused on sentiment analysis to flag specific contents in text messages in peer support app. However, I could not find any paper that mentioned use of generative AI in Peer Support applications.
    • Generative AI can improve many exsisting application by making courses and modules more interactive (similar to how Khan Academy and Data Camp have implemented generative AI to help with completion of courses in their respective platfrom)
    • Additionally use of generative AI could be helpful to provide company and chat assitant to patients at times when a Peer Mentor is not present. Given that isolation and lack of timely access to human specialists are major catalysts for recurrence of drug use, perhaps there should be a feature for users to be able to access the chat bot on auto pilot. This will allow users to connect with a chat bot at any time during day or night (though this is just my idea. Dr.Madiraju has clarified a human should always be in the).
    • Generative AI can perform sentiment analysis and help with text moderation.
    • I hope these insights will assist with the development of the application and the fine-tuning of the language models for this project

Friday (6/9)

2023 Week 3: (June 12th 2023-June 16th 2023)

Monday (6/12)

  • Completed daily journal on Marquette Wiki
  • Began Building Systems with the ChatGPT API course: https://www.deeplearning.ai/short-courses/building-systems-with-chatgpt/
  • Continued watching videos on C#
  • Practised training ChatGPT Prompt Engineering for Developers course content
  • In future days, I will be meeting with Dr.Madiraju to decide if I should be focusing on practising with Open AI API or developing with C# at this moment
Retrieved from "https://reu.cs.mu.edu/index.php?title=User:Khatami&oldid=5627"