Photo by Blake Connally on Unsplash

Collaborator Insights: Beyond AI Skills; How I Grew As A Human Being At Omdena

From me to teamwork, from pointless web scrolling to doing Good for others from my laptop.

Alexis Carrillo Ramirez
Published in
5 min readApr 6, 2020

--

I had the privilege of participating for the first time in an Omdena project and it was a beautiful growth experience. It all began when Michael Burkhardt, Founding Member of Omdena, and I connected and he invited me to apply to a project with the #AIforGood promise: “Develop Machine Learning models to help others in need”. The challenge attempted to impact people who suffer Post-Traumatic-Stress-Disorder (PTSD) by developing a “therapist” chatbot.

A near-death experience leading to PTSD

The project started with the Kick-off call were we met Christoph von Toggenburg who proposed the challenge. During the call, he talked about victims of violence who cannot reach medical attention and have to deal with untreated mental health issues, especially PTSD. When he talked about his motivation with this project, he shared with us his brave own story when he and his teammates miraculously survived an ambush to a humanitarian convoy while helping refugees in Central Africa. Because of that near-death situation, Christoph started to manifest PTSD symptoms.

With this in mind, a group of 30+ AI experts and enthusiasts started to work on the project with the constant guidance of Omdena Founder Rudradeb Mitra. In the following, I want to share my personal experience without specific details about the challenge.

Technical results from the project

You can read articles from Colton Magnant and Albert Lai about the development of the project. Sara wrote about PTSD during the first weeks. Anam and Sabina were part of the project and gave us their point of view about dealing with PTSD. On the more technical aspects, Sylvia tried lineal classification on the dataset and Natu’s approach for the classification problem was Neural Network architectures and an ML backend that helps the chatbot developed by Petar.

From this challenge, I decided to share personal lessons.

My personal lessons learned

#1 There is no “I” in a “team”

At the first stages of the project, we needed data to work on and through community collaboration, we could find a great dataset. The problem was there was no information about PTSD, only thousands of lines of transcripts from therapy sessions.

Colton Magnant made a subset with 48 transcripts on whose lines included the term “PTSD”. I found out at this point, that my Psychology background was more helpful for the project. As a subject expert, I was able to identify manifestations of PTSD on the lines of the sessions. I took the dataset and developed an annotation system for the PTSD criteria, thinking on the input for supervised ML models for text classification. Because we could not find more suitable datasets, all the other teams were waiting for the marked transcripts and I was the only one who could do that.

I felt like the whole weight of the challenge on my shoulders

Until one day during an update call the other team members told me: “We are aware that you need help with that” and asked me “How can we help you?” At that moment, I stopped to think about “me” and started to think about “we”.

The solution was a guide about PTSD symptoms, how to identify a manifestation on the transcripts and how to annotate it on a spreadsheet. With that in mind, I did a video tutorial about the DSM-5 criteria for PTSD and how are marked on the dataset. More than 15 people helped with the annotation process passing from seven to the complete subset of 48 transcripts in a matter of days. We, as a team, solved the bottleneck. The other tasks teams were working and I was happy about that.

It is easy to say, “I’m a team player” but when you have been quite some time on the way of self-learning you get used to working on your own. I solve the problems in my way because there are no other perspectives or means to solve them. Without noticing, I entered the challenge with an individualist mentality, but the team helped me to change that. The enthusiasm of the team drove me to be more flexible, encouraging me to try things that I would discard on other situations and solve a big problem for the challenge. The team with their drive, energy motivation, and kind encouraging words changed me.

Now, from my new perspective, collaboration is about seeing the broad picture, thinking about others, being reachable, and asking for help all the time.

#2 Working for the greater good = a better future

I worked on the project for 1–2 hours a day. It is easier than it sounds. I reduced the scrolling time in social media reading toxic comments, fake news, or seeing perfect life photoshopped facades.

Instead, I worked on something really meaningful that gave me a sense of purpose. I devoted my commute time to chat with my teammates and the Binge-watching hours doing my part on the challenge. Eight weeks later I went back to normal with the satisfaction of the work done. What I did can help people who really need it and I feel good for dedicating my time to that. Today we can do so much for others just being in front of our computers, in the comfort of our homes.

I believe that when we help others we spread a message: “Just like you, when got help, do the same with others when they need it”. By doing so, it possible that someone else gives you a hand when you need it because that is all we do, we help each other.

More about Omdena

Solving Challenges Through Collaboration

Omdena is an innovation platform for building AI solutions to real-world problems through the power of bottom-up collaboration.

--

--

Alexis Carrillo Ramirez
Omdena
Writer for

Machine Learning, Statistics, Psychometrics, NLP, Python, R