ML & Community : 2019 in Review

Source: Deep Learning on Medium

ML & Community : 2019 in Review

“I wanted a perfect ending. Now I’ve learned, the hard way, that some poems don’t rhyme, and some stories don’t have a clear beginning, middle, and end. Life is about not knowing, having to change, taking the moment and making the best of it, without knowing what’s going to happen next. Delicious Ambiguity.”Gilda Radner

I am incredibly grateful for 2019. It was a pivotal year in my ML career. I had some ups, many downs and many many waves of uncertainties. Most of the events that happened this year are things that I didn’t plan for.

Mid 2019, I joined Duke Energy’s Digital Labs working from the incredible Duke Energy Innovation Center in the North Davidson area of Charlotte. This move has proved to be a phenomenal experience for me. I have been privileged to work on projects that lie in the intersection of machine learning and renewables while innovating at a fast pace. For more context on why this is important to me, I was one of the second set of cohorts admitted into the Summer Institute for Sustainability and Energy some years back. This program is sponsored by the University of Illinois at Chicago, Argonne National Laboratory, the Clean Energy Trust among others.

Later, I completed a professional certificate in Energy Innovation and Emerging Technologies from Stanford University. This program examines emerging technologies that will transform how we obtain, distribute and store energy. I took courses on solar cells, smart grids and planning for a sustainable future. Fastforward to 2019, I have had the opportunity to work on Machine Learning projects for solar and wind farms predictive and prescriptive maintenance. I have also developed data-driven strategies, time series forecasting models and energy research methodologies to track the carbon footprint of our large customers. In addition to this, I have had a remarkable opportunity to mentor other data scientists and develop strategies to grow the data science team.

In 2019 and serendipitously, I became a Google Developer Expert in Machine Learning. This program gave me the opportunity to actively contribute and support the developer and startup ecosystems around the world by giving more conference talks, writing articles and mentoring machine learning enthusiasts. I’d like to give thanks to Hannes Hapke for putting in the good words for me and to Karl Weinmeister, Chris Fregly and the Google Developers Experts Program team for the discussions and support. If you’re interested in becoming a GDE, check this, this and this to help you started. You can also reach out and I’ll connect you to the right people.

Projects, Writings and Technical Contributions

  1. Parliament-AI: In this paper, we present results from ongoing research on the categorization of bills introduced in the Nigerian parliament since the fourth republic (1999–2018). For this task, we employed a multi-step approach which involves extracting text from scanned and embedded pdfs with low to medium quality using Optical Character Recognition (OCR) tools and labeling them into eight categories. We investigate the performance of document level embedding for feature representations of the extracted texts before using a Bidirectional Long Short-Term Memory (Bi-LSTM) for our classifier. The performance was further compared with other feature representation and machine learning techniques. I worked with Olamilekan Wahab on this project and it was accepted at the NeurIPS 2019 Workshop on Machine Learning for the Developing World (ML4D). This is the paper and code for the project.

2. FurniSure: I was finally able to complete the write-up about the project I worked on as an Insight Data Science Fellow. I describe how I built an object detection algorithm (FurniSure), using a convolutional neural network-based algorithm called “You Only Look Once” to identify, classify, and localize different types of furniture in images and videos. For the project, the focus was on automating furniture detection so that these items can be identified for advertising purposes.

Conferences and Talks

  1. Machine Learning for All Conference: I gave a talk on “The Creative Art of Engineering Bespoke Features” at the ML4ALL conference in Portland, Oregon. Check out the Video, Blog Intro and Slide Deck.
  2. Google Developers Machine Learning Summit ’19: I gave a talk on “Using Natural Language Processing to Understand Parliamentary Bills from Low Resource Countries” at a Summit which was held at Google’s office in Cambridge, MA. It was an incredible experience. The video was posted on Google Developers’ YouTube page which has about 2million subscribers.

3. AI4D, Deep Learning Indaba: I presented a remote talk on our grant-winning proposal “Using Artificial Intelligence to Digitize Parliamentary Bills in Sub-Saharan Africa”.

4. Google Developer Expert Summit: I attended the 2019 GDE Summit at Sunnyvale, CA. It was a rewarding experience and I had fun. I met most of the ML GDEs and learned a lot.

Grant, Awards and Collaborations

  • Parliament-AI Grant: To extend our work on digitizing parliamentary bills and our high-level goal of building a knowledge representation of all parliaments (bills, debates, votes and proceedings) in Sub-Saharan Africa, we received a grant from the Artificial Intelligence for Development (AI4D) network. This grant will enable us to leverage on the intensive acceleration of computer vision research to build our custom OCR and we also plan to extend the methodology to other countries like Kenya, Ghana and South Africa. Excited to be selected as one of ten winners for this grant.
  • Neural Machine Translation for Low-Resource Languages: According to Sebastian Ruder, one of the biggest open problems for NLP is NMT for low-resource languages. NMT suffers a language diversity problem and growing up in a multi-lingual community with about 300 languages and thousands of dialects, I decided to work on NMT for the second largest Afro-Asiatic language after Arabic — Hausa Language. Hausa is also the third largest trade language across a larger swathe of West Africa after English and French. I have started working on this and the results are pretty good so far. I’m currently collaborating with scholars from the Niger-Volta Language Technologies Institute and working with some starter notebooks created by the Masakhane community. Expect a paper on this in 2020.

Community Roles

  • ML4D: I was a reviewer and program committee for the NeurIPS 2019 Machine Learning for the Developing World (ML4D) workshop. I reviewed three papers and the experience was exhilarating. Also, the acceptance rate for this workshop is around 30%.
  • Rise Networks: In the last quarter of 2019, I was invited to be a Professional Judge and Mentor for the Rise Labs AI Ideathon. The Rise Labs is led by Toyosi Akerele-Ogunsiji, who is an extraordinary entrepreneur and leader. I had fun reviewing the submitted projects and some were incredible.
  • Blossom Academy: I became an Advisor and Technical Mentor at Blossom Academy. Blossom creates a pathway where talents across Africa with diverse academic backgrounds can build the knowledge, skills, and abilities to reach their potential as data science professionals and leverage those talents to contribute to the economic growth of their respective countries. At blossom, I advice the technical team and also do what I love doing the most — teach machine learning.

The Most Important Achievement of 2019

In 2019, I got married to the most important person in my life. The ceremony was beautiful and I really appreciate those that came to merry and celebrate with us — from an uncle that I hadn’t seen in almost two decades who flew in from London to my mom’s last-minute flight across the Atlantic Ocean because of visa delay.


In 2019, there were highs and lows, all sandwiched inbetweet many moments of deep introspection. I realize that I still have a lot of learning to do and a lot of growth to achieve. I look forward to 2020 and unlike other years, I will not be setting outrageous goals for myself. The only thing I have in mind is working more on Neural Machine Translations for Low-Resource Languages, mentoring more people in ML and possibly writing a book. I will wait for every other things to come naturally.

Follow me @waleakinfaderin if you are on Twitter. According to this interview by Omoju Miller on Andrew Ng’s DeepLearning.AI platform, I’m one of the i̶n̶t̶e̶r̶e̶s̶t̶i̶n̶g̶ people to follow if you want to learn more about ML. 🙂