ADVERTISEMENT

This Ghanaian just built a machine learning model that detects insults like WMT

Boredom can either make you creative or very lazy. For software engineer Joojo Dontoh, his magic lies in the former.

This Ghanaian just built a machine learning model that detects insults like WMT

One of Joojo’s works went viral on Ghana Twitter on Friday, January 15, 2021, when he built a machine learning model that detects insults. The machine learning model can detect insults like WMT and Ony3.

“Probably the silliest thing I’ve ever built. A machine learning model that detects insults. Fruits of the lockdown,” Joojo tweeted, accompanied by a video demonstration.

As of the time of filing this report, the said tweet has garnered 656 retweets, 1,400 likes and has over 80 replies. That is how far this learning model has travelled.

ADVERTISEMENT

Speaking to Pulse.com.gh in an exclusive interview, Joojo who is currently studying in Malaysia, said the idea came to him due to the boredom of staying indoors as a result of the COVID-19 lockdown in the Asian country.

“I work as a software engineer in Malaysia and due to COVID-19, we have a very strict lockdown,” he said. “Most times the situation creates a lot of boredom especially working in one confined space the whole time.

“So even though I have work to do on both my job and my master's degree, I wanted to engage in something fun, yet something I could learn from. Previously I'd seen many videos of how people would use machine learning in zoom video calls and stuff. A combination of these things birthed the idea to build such a model and then add a Ghanaian touch to it.”

ADVERTISEMENT

Joojo specialises in server-side applications, although he takes delight in building machine learning models.

His master’s thesis, however, centres on comparing machine learning models for predictability in bone cancer patients. To this end, he is not ruling it out as a future career in this field.

“I’ve been researching on machine learning models for one year now mostly because of my thesis,” Joojo explained.

“I’ve worked on python models (scikit-learn framework), TensorFlow and KNIME. To be honest, this was just a fun project I used to relax. Comparatively, I wrote very little code and it’s generally easy for an engineer to build. Though this isn't my speciality, I haven't totally ruled it out as a career path.”

ADVERTISEMENT

With his work going viral, Joojo plans embark on similar projects. He is not sure if his learning models can be monetized, but is ready to take it to the next level if a market presents itself.

“Hopefully, if I have the time I can embark on more interesting projects. If there is a chance and the time to monetize such a system, I will gladly take it.”

Joojo is currently revelling in the recognition that his learning model has brought him and he hopes many more Ghanaian engineers will spring up in this field.

ADVERTISEMENT

“To be frank, it’s nice to have some recognition for my work and I'm very thankful for that. Hopefully, we can have more and more engineers coming out of Ghana doing even greater work,” he added.

JOIN OUR PULSE COMMUNITY!

Unblock notifications in browser settings.
ADVERTISEMENT

Eyewitness? Submit your stories now via social or:

Email: eyewitness@pulse.com.gh

ADVERTISEMENT