Examination of Recommender Systems for Socially/Economically Disadvantaged Communities

This article is dedicated to the memory of the 1921 black community of Greenwood.

Memory:

Due to persistent barbaric and incendiary racial hatred and acts against them, many African Americans sought to create self sustainable communities of their own. One of those was Greenwood. In 1921, Greenwood was known as one of the wealthiest Black communities in the United States, and was also known as "Black Wall Street''. The community was self-sustainable with its own hospitals, lawyers, supermarkets, professional offices of every kind, dentist, schools, oil reserves, economic development projects, the list goes on.

But 99 years before the same week of George Floyd's death, African Americans in that community suffered a horrific pogrom (pogroms are not massacres. Its worse).

Sparked by mis-information spreaded in newspapers that a Black teenager attacked a white female, and fueled by envy, jealousy and racial hatred, the mob brutally killed, lynched and dismembered hundreds of black men, women and children and then burned the community to the ground. Thousands become homeless as the white mob aimed at ethnically exterminating black people from that community. Black Americans did put up a fight to protect the community, but were outnumbered once the US National guard and local police joined the white mob to overcome them. One guard officer reportedly called black people, "the enemy." as justification for what happened.

Many of the dead were then disrespectfully mass dumped into unmarked graves. To this day, those unmarked graves still exist and there have been no arrests and no one has been tried for any crimes. A picture of a black child carrying another dead black child after the riot shares a horrific dimension of this demonic attack.

Many African Americans who attended Tulsa public schools said they never heard about this monstrous nightmare in any of the schools and nothing was ever said about it.

HOC thank and support Michael Render, Ryan Glover and others for starting and raising 40 million for the black owned Greenwood digital platform named in memory of this horrific event (https://www.cnn.com/2021/06/24/business/greenwood-killer-mike-delayed/index.html)

https://www.cbsnews.com/news/greenwood-massacre-tulsa-oklahoma-1921-race-riot-60-minutes-2020-06-14/

https://theblackwallsttimes.com/2017/03/31/on-the-1921-greenwood-massacre-dont-call-it-a-riot/

HARMONIC ONE COMMUNITIES - COMMENTARY

December 2021 Edition

The problems and challenges which systematically plague historically disadvantaged communities have created an astronomically vast “universe” of data, knowledge and information.This volume of structured and unstructured data often express problems members of these communities persistently face across in every area of people's activities.

Due to the massive volume of information about these problems , those who wish to tackle the problem can possibly feel overwhelmed or experience “information overload” when attempting to address this dynamic web-like problem across so many areas. This is due, in part, to the fact that there are other efforts and programs which already exist and seem to work in silos. This may include a “sea” of social programs, institutional researchers, “special” initiatives, subject matter experts, social justice activist groups, political efforts, think tanks, multi-sided heated opinions, statistical information explosions, urban design challenges, and much more.

Intro to Recommender Systems

The gargantuan volume of information which exists to express or articulate problems in disadvantaged communities could justify using various forms of AI. One such approach, which can help is the use of specialized recommender systems.

Recommender systems are commonplace and many of us use them without even knowing it. Because we can't possibly look through all the products or content on a website, a recommendation system plays an important role in helping us have a better user experience. They are among the most powerful machine learning systems that online retailers implement in order to drive sales and to stand out significantly from competitors.

Many in the field of AI know how recommender systems work, but for those new or need a basic understanding: At its most basic level, recommender systems work by an iterative process of data collection and model training. In each iteration, the model training phase applies some form of statistical analysis on the collected data to “learn” the patterns that are present. Next, based on weighted numerical calculations, it would provide a list of items which are similar to the end user’s choice. The suggested list would theoretically improve on each iteration of the data-collection-model training cycle, hence adding more value to the user and hopefully peaking their interest to explore suggestions provided.

Some examples of recommender systems in action include product recommendations on online search engines, google, Amazon, Netflix,YouTube, online music engines, Facebook (Meta) , news networks, etc

Community Development Recommender Systems

Community based recommender systems can suggest information that is relevant to the advancement of a solution associated with the history , current state and predicted future of that community. AI algorithms, such as collaborative filtering or content-based filtering , could be utilized to build recommendation engines which can lead to more concise solutions for particular situations and synergistically respect the changing conditions of disadvantaged communities. Since recommender systems already exist for other purposes, its crucial for members of disadvantaged communities to understand these systems, their advantages and disadvantages as well as areas which require more research and resources to develop.

As with any system, it would be a good idea to assign or develop an indicator that could help measure or evaluate the effectiveness of initiatives found or developed from a recommender system. One such term, which can possibly serve this purpose, is Community Vitality.

Community Vitality Indicators

While there may be other definitions about Community Vitality on the web, HOC draws inspiration of definition from the wisdom of ancient civilizations found in Africa, Asia, India and South America. Within those cultures, vitality was often associated with something that is “energetic”, “alive”, “active” or “strong”. This was sometimes expressed as Chi (Yin and Yang), Ki, Prana, Ka etc.

Hence after surveying the great work and thinkers from these civilizations, HOC suggest that Community Vitality could mean the following:

“ the current state and ability of a community to develop, grow and live in a manner that is sustainable and progressive across all major people's activities in that community. This would include all micro-level (individual) and the macro level (groups) areas related to economics, education, housing, entertainment, health (mental and physical), business, technology, employment, environmental, law, politics, religion, sex and war. “

Within this definition framework, the micro-level and macro-level areas become measurable indicators which are integrated with a recommender system to help community members make decisions about problems in their community. The advantage of using this system is to leverage the basic advantage of many AI systems: to process large amounts of disparate data and instantly convert it to information to be used in human decision making. This system would allow problems to be addressed on many levels and many different areas of that community to help it become strong and sustainable and erase the title “disadvantaged”.

While there are a large number of possible AI applications which can help communities such as those in the African Diaspora, one must start somewhere.

Finally, we must respect the IceBerg effect of any attempt, which is: what you see on the surface may not fully represent that larger part of the problem which is submerged under the surface.”. Indeed, there is a bigger problem than what has not been presented in this article, but it's possible to explore it further with recommender systems.

A Framework for Implementing AI Projects Which Benefit Historically Disadvantaged Communities

Article Dedication: Isabella Faith Tichenor (2021)

This article is dedicated to 10 year old Isabella Faith Tichenor.

Prior to Isabella's suicide, Justice Department Investigators had found persistent failures by the Davis School district (Utah) to respond to reports of race-based harassment and violence targeting Black and Asian students.

Black students, who make up 1 percent of the student population, have been under attack by suspected white supremacists within the student body. Administrators are said to have failed to respond to hundreds of reports from students about being called slaves, the N-word, and threatened with lynchings if they didn’t go “pick cotton” as instructed by their white classmates."

Students are not born hating others. This is covertly and overtly taught.

https://news.yahoo.com/utah-girl-10-commits-suicide-001000669.html

HARMONIC ONE COMMUNITIES - COMMENTARY

November 2021

Article:

There is a gargantuan body of knowledge, research and information which depicts the complex challenges members of historically socially and economically disadvantaged communities have faced for hundreds of years.

Numerous statistical studies display how these communities are consistently and repeatedly disproportionately subjected to covert and overt forms of unnatural forces and phenomena. This has affected behaviors of all human activities to members of these communities. This includes economics, education, entertainment, health , business, technology, employment, environmental, law, politics, religion, sex and war.

The causes, reasons and situations for these problems are diverse and complex. Hence, from a holistic perspective, we must respect the Iceberg Effect of this situation: what we see and verbally articulate on the surface, may not fully include the larger hidden parts of the problem that is "under the surface" . Therefore, systems thinking allows us to have a healthy balance between the unknown and the known as we proceed with attempts to address or solve problems in historically disadvantaged communities.

But we must also remember that AI, with its inspiration from Mother Nature and her many examples of network-like phenomena, is just as complex as the community problems stated. While scientists did not (and still do not) fully understand the human brain, this incomplete view of the brain did not stop them from developing hardware and software to mimic reason and thought. This is why it’s called “Artificial” intelligence ..... it's not the same natural intelligence displayed by humans and animals. It's fake.

But synthetic cognition has a strength: the ability to do mathematical calculations on large amounts of data and instantly convert the raw data into information to make “a decision” (i.e neural networks). There are other advantages, but this basic attribute has made it a valuable tool in a society where information is becoming the new currency.

But, ironically, this widely used tool has another opportunity: to improve historically disadvantaged communities. The motivation scientists used to develop machines modeled after natural phenomena that they do not fully understand should also be used for communities such as those in the African American and African Diaspora.

Members of disadvantaged communities should be able to use AI to address problems they have personally experienced. Within this effort, there must be equal balance in the theory and application of AI to help members of these communities to solve their own problems. Too much or too little of either (theory vs application) would yield AI useless.

The question is where do we start?

There is no need to reinvent the wheel if methods exist already to address complexity (unless invention is needed upon application)

Enter systems engineering.

If systems thinking is about how you organize your “thinking”, systems engineering is about organizing what you “do”. While there are many fancy definitions of systems engineering, it is basically a structured approach to convert an idea into a product or solution within a fixed time period.

Like systems thinking, systems engineering is not new and has (and is continued to be) been utilized by companies and large institutions for years to build very complex machines and structures. It's an approach that has been used to build many complex products…. From the iphone to an electric car.

Hence when developing an AI project that can address complex problems in socially economically disadvantaged communities, we can use the same (or similar) tools and methods for building solutions for complex problems.

The major steps of this method can be tailored to develop solutions for disadvantaged communities. An outline of this adaptation is as follows:

STEP 1: DEFINE THE PROBLEM/NEED - clearly state what is the situation, person or thing that needs attention and needs to be addressed or solved.

STEP 2: FEASIBILITY STUDY – Determine if the project is worth doing and all of the resources needed exist for the effort to be a success.

STEP 3: PROJECT PLANNING- Plan out who are the major stakeholders in the project, the roles, titles and a functional timeline with risk mitigation plans to complete the project.

STEP 4: SOLUTION VISION – Provide a high level view of how the solution or product should work.

STEP 5: HIGH-LEVEL DESIGN – Create high level design from the solution vision.

STEP 6: DETAILED DESIGN – Create a detailed design as preparation for construction.

STEP 7: CONSTRUCTION – Use the design to build the solution. Some elements during construction of an AI solution

may include the following:

7a. Data Understanding

7b. Data Collection

7c. Data Preparation/Preprocessing (Standardize and Normalise the data)

7d. Exploratory Data Analysis

7e. Data Modeling/Algorithm Selection/Development

7f. Model Evaluation, feature engineering and model metrics

7g. Prep model for iterative testing.

7h. Model Final Evaluation

STEP 8: TEST – Test the solution/product in a controlled environment to confirm functionality or make adjustments.

(note in this step, you are testing the system as a whole, not the exclusively model)

STEP 9: DEPLOYMENT – Begin deploying and integrating the solution in the live environment.

STEP 10: OPERATION AND MAINTENANCE – Monitor and maintain as needed

STEP 11: CONTINUOUS REVIEW AND IMPROVEMENT – Conduct iterative agile reviews of the solution to see if goals are met OR anything needs to be improved to enhance results.

Reading Nature to Harness Technology for Social Economically Disadvantaged Communities

Article Dedication: This article is dedicated to the memory of Jelani Day (2021) and Emmett Louis Till (1955). We stand with Jelani's mother, Carmen Day, his fraternity (Omega Psi Phi) and celebrities, demanding answers

https://a-z-lyrics.lyricsaz.be/jelani-days-mother-clarifies-autopsy-report-that-claimed-his-internal-organs-were-missing/

Article:

There is a strange concept in nature called entelechy. An entelechy is something complex that emerges when you put a large number of simple objects together. One example is a brain neuron. By itself, its a boring (but important!) piece. But looking at a super large mass of billions of neurons, yields a brain, something that is different from a neuron.

As stated in last month's newsletter, it's the power of systems thinking which allows one to view new things which develop into something that is greater than the whole of its parts. Mother nature has practiced this principle for billions of years. Often, many of our modern AI principles draw inspiration from her. For example, there are principles, tools and algorithms (too many to name here) within the AI space of deep learning which harnesses this principle to create classification systems which make decisions on the fly.

Another principle found in nature is "self-organization". Once an entelechy is formed, it begins to take on properties not seen at the level of the "simple part". For example ,when a person gets a minor cut on the skin, the skin will "self-organize" into nature's version of a "bandaid" (scab). Or when a woman gets pregnant, how the egg and sperm combine in her womb and self organize itself into a beautiful living being. This is a living testament that the balance union of a man and woman culminated into the first breath we all took - through a woman.

There are other lessons from nature, but many know that scientists (including those in data science, medical science, genetics etc) have tried to "copy" nature's beautiful and effortless genius to develop mind blowing creations. And in doing so developed new creations - some good, others terrible.

Hence, we must maintain awareness of how mother nature can influence the design of new applications or algorithms which can solve problems in our communities. Thinking outside the box while developing AI applications inspired by mother nature could definitely help make a difference in historically disadvantaged communities. We must be prepared to design, develop, test, implement, and demonstrate how AI can be used to tackle the most difficult societal problems.

Harmonic One Communities seek to use machine learning and related tools in data analytics to make positive impacts on communities which for decades have faced challenges with making progress in almost every facet of society.

Feel free to explore Harmonic One Communities website to learn more about projects under development.