Xpress D/T

1. Problem Description-In times of pandemic situations & looking at the 2nd Covid wave In India & other countries; in addition to ensuring safety & security of near & dear ones; as Governments/medical institutions; it becomes imperative to ensure everyone gets adequately tested; so as to avoid infection spread. Governments are committed to testing but due to large populations who require testing; don't have enough resources/procedures(which are faster, accurate). 

2. Proposed Solution - The proposed solution makes use of Artificial Intelligence algorithms to test samples of potentially affected people for a number of infectious diseases including Covid19.  The algorithms can be trained over multiple datasets to increase their accuracy. The testing becomes agile thereby ensuring that most of the population gets tested in a relatively shorter time period & is cost effective.

3. If Scaled Globally - Faster - Testing times (for infectious diseases), medication start, less spread of disease.      

1. The problem being addressed is around speeding up the detection of infectious diseases (like Covid-19, Malaria, Dengue, etc) using AI algorithms for earlier treatment and preventing disease spread. In pandemic like situations & having experienced one during the 2nd Covid Wave in India - testing speeds & diagnosis is of utmost importance especially for larger populations.  This is important not just to contain the spread of disease but also start of treatment for those infected.

2. Scale of Problem (globally):

a. WHO Covid19 Dashboard - June 12, 2021 -> Globally174,918,667 confirmed cases, 3,782,490 deaths. 

a. World Malaria Report - Apr 2021 -> Globally 229 million malaria cases in 2019. The estimated deaths at 409000 in 2019. 

c. WHO Dengue & Severe Dengue - In 2021, and as of 15 May 2021, the Pan American Health Organization (PAHO) reported 323 729 suspected and confirmed dengue cases and 79 associated deaths, in the Americas region. The five countries reporting most cases are: Brazil (229 436), Peru (21 587), Nicaragua (16 465), Colombia (12 710), Paraguay (10 524).

The factors relating to my solution (and an ideal solution) are around being agile and being accurate with limited resources and manpower.  

To help address the problem mentioned above:

- The solution is a set of AI models which using deep learning process (and using available clinical data sets) gets trained and is able to identify the infectious disease parasites (malaria, Covid19, Dengue, etc) present within the body (blood smear, X-Ray, etc). If the parasite is not present then it concludes that the person is uninfected. If the parasite is present then it concludes that the person is infected; in which case the appropriate treatment needs to start. 

- The AI models once trained would be faster (as compared to human beings) and can do the job with a margin of error. Once trained properly - It also would minimize the occurence of human error which might become prevalent in pandemic kind of situations.    

- For such a model couple of parameters are of utmost importance – accuracy, size and portability. The aim of this model is to keep the same accuracy level as is being currently achieved, a smaller and computationally more efficient model and one which can be deployed easily to the edge and IOT things to help the solution reach remote areas.

1. The target population in this case (whose lives could improve through this solution) can be an individual who is:

- affected by the disease (malaria, Dengue, Covid19, etc)

- reside in a city or even a remote corner of the world/village which could be hard to reach.

 2. For various infectious disease kind of situations - containing disease spread, diagnosis and faster medication start is the need of the hour. The solution caters to such needs of the people in the following ways:

- To detect the disease in a faster manner (in a few minutes/hours as compared to days currently) so that it can be confined and does not spread.

- Once detected  - Appropriate medication can start as soon as possible. 

3. Currently the solution is in the prototype stage and is making use of datasets which are publicly available. The above category of people are not directly benefitting from these solutions due to various regulations and policies mostly government controlled. The governments are becoming aware of these technological advancements and once this gets rolled out - this can benefit all such people's lives directly by providing them with:

- faster diagnosis

- cost effective

- faster medication start

- Draft Better confinement procedures to control further spread

The problem, solution & target population which I have described above are well-aligned to the challenge. 

While the challenge is around Covid19 (an infectious disease which has nowbecome a pandemic); there are other infectious diseases (like Malaria, dengue, etc) too which fitsin this category (not really pandemics but are big killers - infection & death statistics given above). All these diseases if not detected, contained & treated faster have the potential to become pandemics. The solution whichis talked about uses AI models to combine agility alongwith accuracy which are 2 ofthe most important parameters (inmy opinion) to help diagnose/contain spread.

I selected this stage of development because to make a solution move from Prototype to Pilot especially in the health & security areas - government policies and procedures play a pivotal role. Since the current policies don't really allow this type of diagnostics hence it remains in the Prototype phase. At my levels and in my capacity; am trying to create awareness at various levels about this kind of diagnostics & treatment available; especially useful during pandemic situations like the ones we faced in India & Delhi -my hometown currently. 

The prototype is currently tested using various publicly available datasets only for improving the accuracy. In the coming years as more datasets become available the accuracy can be enhanced and government can become more aware about alternates to speed up the testing & diagnostic facilities to prevent spread and start treatment earlier of ones affected. 

- The solution provides a new or significantly improved approach to the problem as it uses AI to speed up testing - not being used much currently. 

- The solution is a single compact trained model (about 20 MB in size) - thereby have the ability to be pushed to the edge (good feature to have in case of remote testing and diagnostic needs where electricity, manpower and other resources could be a challenge).  

- Faster to train – less time to train the model & requires fewer parameters, memory/hardware.

- Because it is compact - it can be pushed onto edge devices thereby improving the reachability to far off locations.

This is catalytic as once this gets trained using more datasets - its accuracy improves and with more liberal government policies around usage of AI technologies in healthcare sector it is only a matter of time that this kind of technology starts producing results - helping mankind not just in containing the spread of diseases but even to start treatments early.   

1. Current - Prototype - 0

2. 1 Year - 500,00,000

3. 5 Years - 100,000,0000 (as much as possible - if resources permit)

- Availability of datasets and procuring them to improve accuracy scores

- Creating awareness amongst doctors, government bodies, publically posting on forums & government initiatives sites; with my limited resources and capacity

- Participating in MIT Solve to further give this initiative a push, create awareness and to bring about some legitimacy and focus towards this way of diagnosing and testing. If successful, then this has the potential to become a revolutionary initiative and an alternate and faster way to perform disease detection and diagnostics to contain the spread of infectious diseases and help populations turn over a pandemic situation.

There are currently 3 of us who are working around various aspects right from improving the model's accuracy to helping create awareness at various levels, marketing, etc. 

We are well positioned to deliver this as our team (including self) throughout in our careers have been focused primarily around creating small ecosystem start-ups from business propositions and then helping them scale to larger self-sustainable individual businesses. There have been several examples of such small startup ecosystems which we have envisioned and help scale. Add to above are the following factors (which we feel would accelerate the journey further along with a push from the Solve team & platform)

1. The prototype is ready and is pretty accurate. 

2. It just needs to be marketed (taken to technicians and doctors) and bundled into a more appreciable look and feel output format (something which I have tried to brief in the 60 month approach earlier).

3. Our network consists of a vast number of individuals and organizations who can step in & possibly help (if required).  

Solve can help me overcome:

1. Helping my solution/prototype reach a wider and like minded audience. 

2. I enjoy networking and collaborating with individuals and organizations. 

3. Solve with its mentorship, direction, network connect, alumni network will be in the best position to help me reach out to people in need especially in the affected countries and help this solution to reach out to the masses when they need it the most - currently being a pandemic situation. 

1. Product/Service Distribution - Since the prototype is ready - With a partner who has an established network -  time to market can be cut down drastically. Hence it will be great if I can find a partner who can distribute the product/service using his network

2. Funding/revenue model - The partner can also help fund some part of the go to market strategy

3. Marketing, media & exposure - This kind of solution to reach to masses require significant amount of media exposure and marketing. The partner can help with execute the plan that I have in this regards and can also suggest improvements.   

1. MIT faculty - to help check the business model & suggest improvements

2. Organizations 

a. to help scale remote working operations 

b. marketing and media initiatives

c. partly fund the solution as highlighted earlier

In current pandemic situation  - There are heroes all over the world who are putting their health and safety at risk to treat those who are infected and to save lives. 

I feel I am qualified for this prize as my solution aims to help such heroes by acting as 1st line of defence in an infectious disease outbreak/pandemic like situations. Let the AI algorithms do the 1st level of analysis and then these heroes can step in to perform a higher degree of analysis/diagnosis what the algorithms are not able to pick up. All this will; give such heroes a breather/rest from their tough daily routines and will also eliminate the potential for any human errors which might creep in because of these heroes being too tired or too busy handling the pandemic.

In current pandemic situation  - There are heroes all over the world who are putting their health and safety at risk to treat those who are infected and to save lives. 

I feel I am qualified for this prize as my solution aims to help such heroes (could be women) by acting as 1st line of defence in an infectious disease outbreak/pandemic like situations. Let the AI algorithms do the 1st level of analysis and then these heroes (women) can step in to perform a higher degree of analysis/diagnosis what the algorithms are not able to pick up. All this will; give such heroes a breather/rest from their tough daily routines and will also eliminate the potential for any human errors which might creep in because of these heroes being too tired or too busy handling the pandemic.

The solution makes extensive use of AI technology to perform disease detection and diagnostics thereby helping control the pandemic situation. Having witnessed the  2nd Covid wave in India & Delhi specifically - i feel this kind of solution would have helped our governments, people immensely and would also have acted as first line of defence mechanism to help the already overburdened healthcare infrastructure.

The solution helps last mile health workers  by taking some load off their schedules using its accurate diagnosis and detection technique. If this kind of system would have been in place in India during the 2nd Covid wave - things would have been different altogether and would have helped contain the spread.