This year we assembled for the 3rd Deep Learning Indaba. For those of you who haven’t heard of the Deep Learning Indaba, you can check out my open letter from last year, and have a look at the website. Keep an eye out for their videos to be posted
Dear Deep Learning Indaba Team,
This year there was something extra afoot. The prior years were the kindling, and this year, I believe we saw the flame. The sort of flame that burns with strength and spreads despite being strangled from oxygen and fuel and adversity. It is no longer an AI conference, but a movement — robust and tenacious and always pushing forwards.
There are some key differences between the Indaba and other machine learning conferences, which I’ve highlighted before. But a couple new ones stood out this year from our prior Indabas. I can barely touch the surface of all the amazing things that happened this year. Every symposium, workshop, poster, practical, talk deserves it’s own detailed post
Out of the comfort zone
In the commitment to diversity and pan-Africanism, you left the comfort of South Africa and took us to Kenyatta University in Nairobi, Kenya. Without having the same country ties, this challenge you took on fully, accepting the hardships and embracing the uniqueness of a new environment. I truly believe that it was worth it. Africans rarely visit other African countries. You facilitated learnings about each other, and we’ve become closer that way.
Having lived in Nairobi for 7 months, I’ve always said South Africa is “Africa lite”, and Nairobi is true African city. One of the first things that stands out about Nairobi is how the old colonial building can’t hold back the the wildness of the environment — it’s overwhelmed and eaten by the magnificent vegetation. A key example of this wildness, was our daily lunches at the Indaba. A hawk would fly down and literally steal the food from our plates. And instead of having places to sit inside, an outdoor chill zone was setup, with music and air and space for discussion.
Filling in the Context
Last year, the Indaba started to contextualize machine learning research in the larger ecosystem with the symposium on Ethics and Fairness, as well as the “How to write a research paper session”. This year, you went even wider. For example, I joined in on the Ethics and Fairness Symposium, only to realise that I was participating in a policy research session. I started to learn about how AI practitioners can affect policy and participate in policy research. Additionally, you ran an interactive workshop about writing up PhD Research Applications, and two workshops dedicated to running a startup, and another to running machine learning in production systems. Even the keynotes focused on the impact of the work we do, and it’s potential problems
This full view of supporting and informing machine learning practitioners, not only in terms of algorithms and data, but about where there work sits in the world and how they can move forward is vital. And I guarentee you, no other conference is dedicated to it in this fashion.
The Strength of our Community
- In 2017, we were shy with each other. Intimidated by other researchers and other attendees. I found we kept to our bubbles, afraid of what other’s might say
- In 2018, we starting forging connections. We started finding collaborators. These connections started small and grew. I met my friend and collaborator Laura Jane Martinus there. In December, we attended NeurIPS. I found the conference generally overwhelming, but what stood out was that we had an Indaba crew — a big one. A crew that by the end of it many international researchers had started joining — wanting to be a part of what we had. The community was showing it’s face
Leading up to the Indaba 2019, Avenger’s GIFs were shared regularly on twitter:
The community is so strong that we all feel we’re fighting on the same team for the same things. And coming together at the DL Indaba is a cumulation of our feelings of empowerment. There are individuals from across the continent who I’ve engaged with online, but never met. When I finally saw them at the Indaba, we literally ran towards each other and embraced.
This year, when many attendees in Sudan and Nigeria did not have funds to attend, the community funded them. International ML researchers and Indaba allies contributed funds which funded flights and accommodation of 25 Sudanese and 5 Nigerian researchers.
Diversity to new heights
The DL Indaba seems to be only AI conference in Africa where the attendees and speakers look like the world outside. Having been involved in some of the behind the scenes action, I know the time and effort that the Indaba committee put into seeking brilliant diverse speakers who have not been placed in the spotlight before, as well as ensuring diversity within the attendees. This is something I spoke about last year extensively, and something I rant about at every other conference. I now use the Indaba as a benchmark for what other conferences on the continent should look like. I’ve had arguements with other conferences, where they ask what diversity of speakers looks like. I simply point them to the Indaba. Every year, the line-ups are more and more diverse with absolutely no decrease in quality.
Achieving diversity in a world that does not know how to celebrate it takes effort. One needs to seek far beyond one’s own network
Masakhane — We Build Together
As per our theme from last year, this year we lived “Masakhane” — We build together. The planning of Indaba seemed to involve even more people across the continent than ever before — from the many workshops, to the registration, to the Baobab app, to the wonderful practicals. Indaba organisers, we know letting control of your Indaba child was difficult to do, but in allowing us on the continent to get involved, we feel ownership of our ML community.
The last series of IndabaX’s reached 27 countries. That’s 27 organising committees dedicated to strengthening Machine Learning within their country. That’s 27 countries hosting machine learning events, often despite severe adversity (Somalia, Sudan, and DR Congo for example).
Collaborations are emerging between universities and individuals and organisations. We feel empowered through each other. This year, we saw the strengthening of many other movements: The many AI Saturdays in Nigeria, Data Science Nigeria, Zindi (African Kaggle for Africans by Africans), and Sauti Yetu (an African NLP Unconference).
Sauti Yetu — Our Voice
The theme to this year’s Indaba was “Sauti Yetu”. Our Voice. In prior years, Africa fought the tough battle against our societal imposter syndrome — whereby international researchers would come and tell us we are good enough. This year, we’ve been conquering our imposter syndrome, and as evidence, African voices finally stood up.
Our favourite day was the African Symposium day, filled with the voices of Africans and their research, followed by the simply magical poster sessions under the trees — something you will never see anywhere else in the research world. The energy of that day was wild and empowering. My heart was so full with love and excitement and inspiration. Each and every researcher on the continent truly is an inspiration. The quality of research is unbelievable and the content is truly representative. Research world, watch out, expect delegations of African researchers at your conferences. We’re coming!
The Critical Yeast
DeepMind Researcher and Co-founder of the DL Indaba movement, Shakir Mohammed, speaks of the “critical yeast”.
This is the bread we smelled earlier. Yeast has the effect that, when making bread, starts with a small amount but quickly multiplies to become the key ingredient that changes water and flour into something textured and delicious. That is what is asked of us: who are the groups of people that when put and held together, begin to act like yeast, multiplying exponentially, and as that group rises and grows, they lift-up the people around them, and transform the world around us.
Indaba committee, you are that critical yeast and it’s reached that point of multiplying exponentially and the people are rising and lifting up those around them. I used to be sad when the Indaba was over, since I knew it would be a year until it happened again. Now the spirit of the DL Indaba continues every day. Not a day goes by where I am not exchanging ideas or working on new projects with friends and collaborators I’ve made because of the Indaba. Right now, we’re busy organising workshops at major conferences, running research projects that involve Africans from every part of the continent. It is heart-filling and energizing.
I’ve been struggling to write this article because it’s hard to phrase what empowerment feels like. Personally, I went from feeling severely inadequate two years ago, to feeling empowered last year, to feeling empowered to the point where I can empower others. But it’s not just me, the entire ecosystem has changed. Two years ago we felt alone in our work and our research. Now, we know we have hundreds of people we can turn to for advice and support.
What’s even better is the world has started to take notice. International tech-writer Dave Gershgorn specifically requested asked to cover the Indaba. The word on the street is that ML researchers around the world all know about the Indaba and the work being done. Our voices are being heard and the world is paying attention.
Thank you ❤
This article is dedicated to the DL Indaba team. That team is no longer just the committee and organisers. It’s you, it’s us, it’s everyone who attended, who funded people to attend, who sponsored, who did research, who shared that research, who wrote about the Indaba. It’s us and our community
Thank you to the committee. Words don’t make sense here. We know the endless effort it took to put this event on and can’t wait till we next meet ❤
Thank you to the sponsors who understand our vision, enable this event to happen, decidate time and funds and support us throughout our journeys
And thank you to the local organisers, speakers and attendees. You are the DL Indaba now. Let’s see how we can change the world by Indaba 2020 in Tunisia ❤