“Go Deep or Go Home”
[Highlights of Indaba X Ghana]
Ghana Data Science Summit 2019 [IndabaX Ghana, @ghdatascience] has just ended and the program was loaded with information about data science, Machine Learning, and Artificial Intelligence. This year’s theme was dubbed “The promise of Data Science for Economic Transformation” at the AITI-KACE, Accra from Thursday the 25thof April to Saturday the 27thof April, 2019.
I must say I was blown away by the number of talent we have in the field of Machine Learning and AI and the various deployments of AI in the area of health, and agriculture, etc. In my opinion, this is one of the very well oraganized events I have attended where everything started and ended on time with the speakers well vexed in their area of expertise. A skeletal description of the conference entailed two days of exposés with lunch break, panel discussions and poster presentations.
On the top of my head, I could count over 250 participants and over 20 conference speakers, panelists and workshop session trainers. This article cannot surely detail the event but just a highlight of what transpired and what I observed.
What went on…
Day 1
Day 1 was centered on Data Management and Analytics Focus.
We had the welcome addressby Delali Agbenyegah [@DelaliAgbey] who is a Data Science Director at Express, Columbus, Ohio, USA and co-Founder of Wave-2 Analytics and also the Ghana Data Science Summit Chair.
We had the keynote address by Winnifred Kotin who is the country director for Superfluid Labs. She touched on “The promise of Data Science for Economic Transformation”.
John Aidoo, a Data Analytics Manager at Central Insurance Company, Van Wert, Ohio also treated “Data Management: The Foundation of All Analytics”.
We had Isaac Aidoo [@dataminergh] who is head of Customer Analytics at Zoona, a company based in South Africa speaking on Business Analytics: A Complete Overview of Data Driven Decision Making in a Quickly Changing Business Environment.
There was an interesting demonstration on Predictive Analytics in Healthcareby Danielle Belgrave [@DaniCMBelg], PhD, a researcher at Microsoft in Cambridge, UK.
Finally, Delali touched on “A Practical Exposition on Data Science in the Retail Marketing and Financial Services” before the panel discussion on “Overcoming Barriers to Analytics and its Application in Ghana” moderated by Akua Oseiwah Ahenkorah who is Operations Manager at EAI Information Systems, Ghana with ENN Nortey, PhD, Snr. Lecturer and Statistical Consultant at the University of Ghana and Anani Lotsi having same attributes as ENN and a lady volunteer who is a PhD student from the Mathematics Department of the Kwame Nkrumah University of Science and Technology.
Day 2
Day 2 centered on “Artificial Intelligence and Machine Learning Focus”. This Day had a gathering of great minds and intellectuals in the field who are practically engaging knowledge in AI.
The Keynote Address was by Moustapha Cisse [@Moustapha_6C], PhD, Head of Google AI Center, & Founder and Director of African Master of Machine Intelligence at African Institute of Mathematical Science (AIMS). He delivered a talk on Machine Learning and Artificial Intelligence — Current state of research and Applications in Africa and beyond.
Shakir Mohammed [@shakir_za], a Research Scientist at DeepMind, London, UK treated the topic “Introduction to Statistical Machine Learning and applications”.
In as much as most of the speakers have had some US or UK training, I was happy when other local players were given the opportunity to also share their thoughts on applications of AI.
Kwadwo Agyepong-Ntra[@KayO_GH] from Meltwater Entrepreneurial School of Technology expounded on Deep Learning Fundamentals — Architecture and Applications.
Ayorkor Korsah [@AyorkorK], PhD, Head of Department of Computer Science at Ashesi University talked about Introduction to Robotics and its applications.
Arguably one of the most intelligent guys we have in the field, Darlington Akogo [@darlingtinho], who found interest in coding when he was only 14 years of age and dropped out of school to pursue self-education in computer and its applications delivered his talk on Machine Learning Applications in Health Careand is also the Founder and CEO as well as Director of Artificial Intelligence, MinoHealth AI Labs.
A practitioner’s perspective on Building Machine Products in Real Lifewas by Richard Ackon [@esquire_gh], a Software Engineer at mPharma.
Finally, a Panel Discussion on Diversity, Inclusion, Building Careers and Creating opportunities in Machine Learningwhich was facilitated by Jeff Acheampong, Co-founder and CEO of Blytix and June Seif, Darlington Akogo and Samuel Mensah helped with the panel discussion.
Day 3
Day 3 was dedicated to Tutorials and special topics
Below are some topics treated with hands on tutorials
· Descriptive analytics and Data Visualization
· Mathematics of Machine Learning
· R Programming for Data Science and Machine Learning
· Python Programming for Data Science and Machine Learning
· Machine Learning Basics
· Introduction to Deep Learning and TensorFlow
· Machine Learning Applications in Predictive Analytics
· Introduction to Machine Learning Applications in Computer Vision.
Poster Presentation
The organizers did well by allowing for poster presentations which also came with some reward. The poster presentations were on display from the very first day and there was a network session at the end of day 1 and day 2 that allowed for interactions with the presenters on the projects presented.
One of the presenters Deborah Kanubala, an alumnus of AIMS presented a work she did on Rice farmers in Senegal and their ability to repay loans. Debbie won the contest with $50 cash price and a chance to showcase her work at the continental level.
John Bigiliko, did a poster presentation on common sense reasoning in Artificial Intelligence. John came third and won a cash price of $25 and we had other presentations as well on a power generator fuel system and another on the usage of AI to detect the number plate of car e.t.c. In sum, I felt a great display of brilliance and the use of Machine Learning and AI to optimize life in general.
With the success of the program organization and execution, I had a few observations during the program which I think are pertinent to the subject of Machine Learning and Data Science.
Diversity
I was very much intrigued when a priest who is an attendee shared a story. He was given a training in Austria during his student days at the seminary. The training involved some statistics but he was weak in statistics and the program was designed such that he was allowed to take a supplementary course in statistics so he can be at par and this marked the beginning of his interest in statistics. Later on, he was allowed to pursue PhD. However, when he came to Ghana and applied to teach statistics, he was rejected with the excuse that he doesn’t have enough statistics background simply because his first degree wasn’t in Statistics. I could tell at that point he felt left out and betrayed by the course he gave so much love and attention.
I used this story to illustrate the exclusion of people from other disciplines in the field of AI and Machine Learning. When I enrolled in the Microsoft Professional Program in Data Science, the key thing the instructors stressed on was one’s mathematical and or statistical background. It almost feels that if you’re not from any of the Science, Technology, Engineering, or Mathematics (STEM), you can’t do Data Science. In fact, John Aidoo was blunt about it when he explained even why those from Computer Engineering are those mostly admitted into MSc. Data Science programs or allowed in the field of Data Science and Machine Learning. In as much as I agree with him, persons from other disciplines such as Sociology, Economics, and other Social Sciences feel left out of the revolution. What gives me the advantage is that I am a hybrid of the STEM and Social Sciences having a degree in Mathematics and Economics. The only defense for diversity I heard from Darlington was that when building Data Science team, there should be diversity in order to have diversity in thinking. If a team is only made up of Engineers and Statisticians and Mathematicians, then your thinking is narrow but with diversity, ideas are sourced from different areas that makes the business complete.
In my opinion, I urge others who have weak statistics background to take up the challenge to learn Data Science. We leave the field for the STEM guys because some bit of coding is required and it’s also easier if you’re a software developer when you can build modules and deploy them. This shouldn’t scare anyone. I answered a question posed by two of the attendees and I explained that, if you don’t know coding, you can still do Data Science. The only difference is that you might have high time trying to decipher the way algorithms work. Your aim then will be to know what each algorithm does and how to interpret the results. At some level, especially with data visualization, some knowledge of visualization in python or R is needed. You can also use excel but in cases where you have to match labels with features, the better way is python or R. Companies such as Microsoft has made it easier for non-coders to use Machine Learning Algorithms to data. You can spend time learning Azure which is pretty easier and has a way you can deploy your model without having to build an API for your clients. However, it is not free but Microsoft has a free subscription which comes with $200 credit which is enough for one to learn Azure.
Women Inclusion
There is still a huge gap in the number of males compared to females in the field of AI and Machine Learning. There is still a minority representation of women in tech. Even at the conference, we had only three women speakers, one panelist, one volunteer panelist, and one facilitator and the host (Aseda [@AsedaAD]) compared to about seventeen men speakers and panelists (per my counting). I cannot for sure articulate the reason for this phenomenon but with less women in tech, it feels like dealing with class imbalance in a classification problem.
Perhaps it’s just psychological that technical areas are left for men and care and hospitality is left for women. Just as we have more representation of women in the area of hospitality, so it is in tech but the other way around. I dunno if it is out of shear disinterest or women just don’t want to bothered with computer.
The Exposure
One trend I noticed among the speakers is where they attended college and graduate School aside a few ones. This is one of the problems I listed in my immediate previous article about the state of Data Science in Africa. I indicated that it is still difficult for [Junior] Data Scientists to find job and grow after their training. In as much as one may be familiar with Machine Learning algorithms, he will lose it if his skills are not used.
Majority of the conference speakers attended grad school in the US and the UK and found jobs there after graduation which is an indication that their skills are sorted after by corporate US and corporate UK since they know the value and importance of having a data science team. A part for South Africa which is doing better in offering internship opportunity to students interested in Data Science, its neighbors are doing close to zero in developing talents in the field. This is not a case of “supply creates its own demand”.
One solution I propose is, such conferences should not be for only enthusiasts or students alone but the corporate folks should also be invited to see for the themselves the transformation happening that they are missing out. This will help bridge the gap between skilled labor and corporate Ghana. Else, the only solution is for one to carve his own niche. The field still is not known and much will have to be done to create awareness.
How about those who have taken the lead but in other disciplines not STEM …?
One question that was raised that I would like to touch on is “What of those who have taken the lead but in other disciplines not STEM” and how they can become data scientists.
This is bothering a bit because it will depend on how far you’ve gone and if you can come back to study Mathematics or Statistics and hence Data Science. Some sites I have visited don’t give age limit on the opportunities they offer to junior Data Scientists but others do put a cap on age. Some opportunities I saw in a certain South African website put the age cap at 30 years and others are somehow generous enough to take it to 35.
I surely do not have a say in what age they should limit to but in my opinion and in the African context, the field is now gaining grounds and yes opportunities should come with some age cap but it shouldn’t be too limiting in order to give chance to those with or without STEM but graduated before the tern “Data Science” was coined.
I will conclude by saying that Data Science, Machine Learning, Deep Learning which forms the building block of AI have come to stay and we need to harness and deploy them. Deep Learning is seen as the core drive of AI and so… you either “Go deep of Go home”
It will be unfair to conclude without honoring the sponsors of the program. Big thanks to the main sponsors — Microsoft, Wave 2, Deep Learning Indaba and Google.
Without Data, you’re just another person with an opinion — W. Edwards Deming
You can follow me @joshkobeng, Joshua Kobina Obeng (LinkedIn), Nii Joshua (Facebook). I offer private tuition in python for data science, PowerBI, and Azure Machine Learning.
