Health-AI Ethics Atlas - Requirements

[iamrmrishan]

Is there any specific reason to mention only D3 and Leaflet. I think we can use deck.gl. Only downside is heavy memory demand from clients’ machines to render a map resulting less mobile browser compatibility. Open to discussion.

[manascb1344]

Hi,
Has anyone who is interested in this project received any tasks?

[selenbw]

The project has not started yet, once it starts we will share the next steps.

[selenbw]

Please read carefully: #28

[kumar-piyush12]

Proposed Contribution: Use of 3D printed food products (esp. meat) (whose model is made via AI prompts) is set to reduce animal cruelty (particularly in meat sector). I want to showcase the rising trends (startups, manufacturing-trends) on a global map (interactive).
Please let me know your remarks.
Email: f20210815@pilani.bits-pilani.ac.in

[heymitali]

Hi, I am Mitali. I am very much interested in this project. Kindly guide me for the next steps.

[selenbw]

The project has not started yet, once it starts we will share the next steps.

[selenbw]

Is there any specific reason to mention only D3 and Leaflet. I think we can use deck.gl. Only downside is heavy memory demand from clients’ machines to render a map resulting less mobile browser compatibility. Open to discussion.

There is not specific reason to use only D3 and Leaflet. Plan is to decide based on the final group discussion and open to suggestions,

[selenbw]

Let's discuss: Web Scraping Task

• Your first task is to develop a web scraping script to compile a list of AI ethics guidelines and policies specifically related to healthcare or medicine. Focus on the following:

1. Identify and target reputable websites and online databases that host or reference such guidelines (e.g., WHO, IEEE, medical journals, government health departments).
2. Your script should extract key information including the name of the guideline/policy, the issuing organization, the publication date, and a brief summary or abstract.
3. Ensure your script is efficient and can handle different website layouts and structures.
   The goal is to create a comprehensive and accurate dataset through automated web scraping, demonstrating your proficiency in coding, data extraction, and handling web data.
   Please discuss here, ask questions and provide solutions.

[KYash03]

Creating a universal web scraping script wouldn't work because, as mentioned, each website has its unique structure and layout. I suggest we experiment with a combination of Beautiful Soup and spaCy/NLTK (NER) to extract the required information.

[Apurv428]

Hi @selenbw , do we have any targeted websites for this or we are planning to search the entire web?

[selenbw]

Publish or Perish and scrapping pubmed is definetely a good start.
Is there any way to compare their results given the same key words?
The websites we should include are below (open to other suggestions)
AI Ethics Guidelines Global Inventory
Linking Artificial Intelligence Principles” (LAIP) guidelines
Pubmed
Embase.com
Web of Science
https://www.aiethicist.org/
https://www.coe.int/en/web/artificial-intelligence/ethical-frameworks
Bonus:
Some Examples to ger inspired (hoping better dosing though :)):
https://playground.airespucrs.org/worldwide-ai-ethics
https://www.airespucrs.org/en/worldwide-ai-ethics

[iamrmrishan]

PoPCites.csv
I have uploaded the CSV here. we can get title, authors and compare results.

Publish or Perish gets a Google Scholar search (it includes the results from above mentioned sites as well). If we are looking for specific sites, we should consider a script. Also Publish or Perish got options for get a seperate result only with PubMed and Web of Science.

[201677I0318]

I agree with @KYash03 that we can't write just one script to crawl all websites.
But we can write different scripts for specific websites and put them in a GUI so that we can have something like Publish or Perish.
From this point of view, not only this GSoC project, we can do something just as interesting, that's a lot of work to be done, I've done the IEEE part, and if anyone is interested, welcome to contribute.

By the way Publish or Perish has a bug, the interface for Google scholar is obsolete and the following failure message appears when used.

[manascb1344]

Hello @selenbw,

I'd like to discuss how we can ensure that our web scraping approach prioritizes trustworthy sources for AI ethics guidelines and policies in healthcare or medicine. Are there specific websites or databases known for hosting authoritative content in this area? Additionally, should we concentrate on scraping particular websites or gather data from a wide range of sources to build a comprehensive dataset?

I've been exploring web scraping techniques for our task, aiming to collect AI ethics guidelines and policies related to healthcare and medicine. I experimented with BeautifulSoup and spaCy on platforms like PubMed, UNESCO, and IEEE. However, I encountered an obstacle - certain websites have measures in place to deter web scrapers.

Despite this challenge, I made progress, particularly with PubMed. The script I've been refining retrieves HTML content from PubMed search results, utilizes BeautifulSoup for parsing, and extracts specific details such as guideline names, organizations, publication dates, and summaries.

To enhance the script's capabilities, I integrated spaCy for text processing tasks.

However, there are still areas where we can enhance the script. Do you have any suggestions on how we can navigate around websites that block web scrapers?

[loupdaniel]

Hello @manascb1344,

Could you please share platforms that restrict web scrapping?

[manascb1344]

Based on my scripts, it looks like both the IEEE and UNESCO websites have implemented measures to prevent web scraping. I'm currently exploring alternative methods to extract the desired data directly from the websites. Any suggestions on overcoming these restrictions would be greatly appreciated!

[selenbw]

The websites we should include are below (open to other suggestions)
AI Ethics Guidelines Global Inventory
Linking Artificial Intelligence Principles” (LAIP) guidelines
Pubmed
Embase.com
Web of Science
https://www.aiethicist.org/
https://www.coe.int/en/web/artificial-intelligence/ethical-frameworks
ALSO IEEE and UNESCO are great ideas too but I wonder about their restriction you mentioned.
Bonus:
Some Examples to get inspired (hoping better dosing though :)):
https://playground.airespucrs.org/worldwide-ai-ethics
https://www.airespucrs.org/en/worldwide-ai-ethics

[201677I0318]

Hello @manascb1344
I recently encountered the same problem when I was working on a scrapy on another project, spoofing a website by simulating browser behavior to bypass anti-scrapy settings often works, I tried to write a script that worked well in IEEE.

Here is my code, hope that helpful to you.

import csv
import time
from bs4 import BeautifulSoup
from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By
import tkinter as tk
from tkinter import ttk
from selenium import webdriver



# Set Chrome options
chrome_options = Options()
chrome_options.add_argument("--headless")  # Use headless mode
chrome_options.add_argument("--disable-gpu")  # Disable GPU acceleration
chrome_options.add_argument('user-agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36"')

# Create Chrome WebDriver
driver = webdriver.Chrome(options=chrome_options)

# Set URL
url = "https://ieeexplore.ieee.org/search/searchresult.jsp?newsearch=true&queryText="
HomePage = "https://ieeexplore.ieee.org"

# Create CSV file to save results
def create_csv(name):
    with open(name+'.csv', 'w', newline='', encoding='utf-8-sig') as f:
        writer = csv.writer(f)
        writer.writerow(['Cites','Tittle', 'Author Name', 'Publish Year', 'Publisher', 'Abstract', 'URL'])
# Def function to open webpage
def open_url(url):
    driver.get(url)
    time.sleep(2)
    wait = WebDriverWait(driver, 10)
    wait.until(EC.presence_of_element_located((By.TAG_NAME, 'body')))
    soup = BeautifulSoup(driver.page_source, 'html.parser')
    return soup
# Open ebpage
def search(keyword):
    soup = open_url(url+keyword)
    # Perform web scraping operations
    count = soup.find("h1", class_="Dashboard-header col-12").find_all("span",class_="strong")[1].text.replace(",","")
    count = int(count)//25 + 1
    print("Get " + str(count) + " pages of results")
    create_csv(keyword)
    for i in range(count):
        print("Processing page " + str(i+1))
        urltemp = url + keyword + "&pageNumber=" + str(i+1)
        soup = open_url(urltemp)
        items = soup.find_all('xpl-results-item', {'_ngcontent-sst-c145': "", '_nghost-sst-c144': ""})
        if items == []:
            print("No results found, reopend the url")
            urltemp = url + keyword + "&pageNumber=" + str(i)
            soup = open_url(urltemp)
            items = soup.find_all("div", {"_ngcontent-sst-c144":""},class_="hide-mobile")
            if items == []:
                print("No results found, skip this page")
                continue
        for item in items:
            tittle = item.find('h3', {'_ngcontent-sst-c144': "", 'class': "text-md-md-lh"}).text
            author = item.find('p', {'_ngcontent-sst-c128': "", 'class': "author text-base-md-lh"})
            if author is None:
                author = ""
            else:
                author = author.get_text(strip=True)
            publishInfo = item.find("div", class_="publisher-info-container")
            publishYear = publishInfo.find("span").text.split(": ")[1]
            publisher = publishInfo.find("span", class_="title").find_next_sibling('span').text
            abstractUrl = item.find("a", {'_ngcontent-sst-c144':"", 'class': "fw-bold", 'xplanchortagroutinghandler': "", 'xplhighlight': ""})
            if abstractUrl is None:
                abstractUrl = item.find("a", {'ngbtooltip':"HTML format", 'tooltipclass':"document-toolbar-tooltip"})
            abstractUrl = HomePage + abstractUrl['href']
            print("Getting the datil source of "+abstractUrl)
            soup = open_url(abstractUrl)
            cites = soup.find('div',string ="Cites in")
            if cites is None:
                cites = 0
            else: 
                cites = cites.find_previous_sibling('div').text
            abstract = soup.find("div", class_="abstract-desktop-div hide-mobile text-base-md-lh")
            if abstract is None:
                abstract = ""
            else:
                abstract = abstract.find('div', {'_ngcontent-puc-c169': "", 'xplmathjax': ""}).text
            with open(keyword+'.csv', 'a', newline='', encoding='utf-8-sig') as f:
                writer = csv.writer(f)
                writer.writerow([cites,tittle, author, publishYear, publisher, abstract,abstractUrl])
    # Close the browser
    driver.quit()

        

# Get input from user
def get_input():
    keyword = input("Please input the keyword: ")
    search(keyword)
    print("Finish searching")
    print("The result has been saved in "+keyword+".csv")
    print("The browser has been closed")

# get_input()
get_input()

AI ethics guidelines.csv，this is the results file.

[AtharvSabde]

Hello everyone, Myself atharv sabde and I am very excited to be part of this project.
@selenbw what next steps should I take to get selected in this project. please guide me. thankyou

[Shreeman5]

Hello All,
My name is Shreeman Gautam and I am a PhD student doing visualization research at the University of Utah.
@selenbw, I looked at the project, read the contributor guidelines and I think this a good fit for me since I have a background in CSS, HTML and Javascript(D3). I have one question: If the project goes well and we finish all the tasks, will we write a conference paper about it?
Thank you!
Shreeman

[KarthikDani]

Hello @selenbw! I have scraped one of the toughest platforms ethically and it's public data from Linkedin.

With 2 crawler's I was able to fetch nearly 200,000 company profiles along with each of their 16 different company details, all done in two steps. Find my repo on the same Linkedin Company Directory Scraping System

Well I am new to GSoC, and would like to know if I still have time to submit proposal for this project? I LOVE ETHICS!

[Khyati9505]

Hello everyone,

My name is Khyati Bhat, and I'm currently a first-year student majoring in Biological Engineering at the Indian Institute of Technology, Madras (IITM). I am a beginner to opensource however I have been working on web development, focusing on front-end development using HTML, CSS, and JavaScript along with JavaScript libraries like React. Additionally, I have worked on projects involving machine learning utilizing various Python libraries like Numpy, Pandas, OpenCV, Matplotlib, and Seaborn as well as Python Frameworks like Tensorflow and Keras. 
After reviewing many of the project ideas, I find myself very intrigued by the Health - AI Ethics Atlas project as this project has the potential to become a tool that help solve a significant problem that the medical field is facing right now, and is a project I have a personal stake in.  
   @selenbw I'm working on the proposal right now, would it be possible to send it to you by email for feedback?

Email: Khyati9505@gmail.com

[kruti107]

Hello,
I am Kruti Pandya currently in last semester of MTech AI journey.I would be happy to be the part of the project "Health-AI Ethics Atlas" @selenbw
email: kruti.pmtai22@sot.pdpu.ac.in

[selenbw]

This project wasn't selected for GSoC, but my team will continue working on it. We're aiming for a conference or journal paper and plan to finish in 3 months. It will require consistent commitment. If you're seriously interested in joining the team, please contact me at selen.bozkurt@emory.edu.

[Zawa-ll]

Hello! I am interested in contributing to the Health-AI Ethics Atlas project. I just sent you an email, looking forward to your response for further discussion.