DAYMN — 10 Oct 2021
Data Articles You Might Need. Subscribe & follow here!
With daylight savings kicking in, and yet another mental arithmetic challenge upon us, every time we need to talk to someone in a different time zone — here are the top 5 articles from this past week. Please do share your feedback & thoughts!
1. SimBot — A New AlexaPrize Competition
Alexa Prize has launched a new challenge, Alexa Prize SimBot Challenge - focused on helping advance development of next-generation virtual assistants that will assist humans in completing real-world tasks by continuously learning, and gaining the ability to perform commonsense reasoning.
For those who might remember the famous BellKor algorithm for the Netflix Prize almost 12 years ago now, we can hope for something as significant! While it is aimed only at universities, it’d be interesting to follow the research and quality of submissions here over time.
Call out? Amazon has open-sourced a valuable TEACh (Task-driven Embodied Agents that Chat) dataset that has some gold mine of human to human dialogues for those looking for a quality dataset for your NLP algorithms.
As a big fan of acronyms and good quality open-source datasets, I’m sold!
2. Quantified Self — Made Much Easier
For those who have read my Love, Life and R article, you’d know my passion for quantifying and analysing our personal data to derive better understanding of our own selves — and find ways to improve ourselves. This great article from Dr. Gregor Scheithauer does a wonderful job of not only showing how to do it, but also how to do it for “everyone” without having to worry about knowing how to code.
https://towardsdatascience.com/create-beautiful-art-from-your-personal-data-9dc0abfeeaf
Call out? In general, I use Python and R to create charts and visualization of data. But I recently discovered a free online service with the name of rawgraps.io. It allows you to create a vast number of different data charts (depending on your type of data). As I understand it, it works in your browser, but the data is not uploaded to their backend; everything works in your browser on your computers.
3. Building and managing ML models in smaller organisations — the challenges
Often times, when people think of ML models and running them at scale — the thoughts immediately go to Big Tech, with the likes of Netflix, Google, Facebook, Amazon, AirBnB and others. And as a corollary, the practices and concepts are then “inspired” from these companies. However, as someone who has worked for much smaller organisations and trying to stand up ML functions — there are some unspoken, unwritten massive challenges — the least of which is not a mismatch between expectation vs reality.
This research paper from MIT does a great service of shining a light on this selection bias, from real interviews of real people.
Why does it resonate with me? I’ll leave the theme headers here, and you be the judge of it!
- “It’s Tough.” Tensions Between Expectations & Feasibility
- “A Hotbed of Bias.” Efforts to Assess, Prevent, & Mitigate Bias
- “You can poke and prod black box models, right?” Black Boxes & Overconfidence
- “Data Literacy Is Not a Silver Bullet.” On Communication & Collaboration
- “GDPR Doesn’t Affect Us.” Assessing Tensions Between Privacy & Growth
4. Bill of Rights for AI-powered world
There has been a lot of chatter in the public policy making spheres about the harmful and useful impacts of AI-driven decision making in a lot of our daily lives. While absolute evil like Facebook algorithms pervade our lives, there are also subtle not-meant-for-evil-but-has-bad-outcomes kind of algorithms that we encounter every day as part of our modern lives.
So how does one i.e. the government/wider citizenry make sure that algorithms we use are not being misused or misguided? It needs a wider public debate and a public policy reckoning that includes conversations at all levels of the society — not just regulators, governments or Big Tech. This good article on the Wired talks about the need for a Bill of Rights for AI in the US.
https://www.wired.com/story/opinion-bill-of-rights-artificial-intelligence/
Did you know? Only governments from 6 countries/nation blocs have come forward with a clear stance or guidance on fairness in AI and algorithms — China, Australia, EU, OECD, Japan and Singapore. While no one has yet passed legislation or regulation on AI usage, it’s not too far away in the foreseeable future!
5. How to waste your career — one comfortable year at a time
For those of who experiencing a mid-career crisis or wondering if what you clock in everyday has a significant impact, this wonderful article from Apoorva Govind does a great job of shining a light on the how to ask that question of yourself, what needs to be done to measure it — so that it is not just a feeling — and how to remedy it. Spoiler — it’s not just about quitting to find a new job, but doing something different where you are.
Without giving away too much, this is a great read below.
https://apoorvagovind.substack.com/p/how-to-waste-your-career-one-comfortable
Call out? If you are ambitious about your career and want to achieve specific goals, you can’t get there by random accident. You need to be deliberate and keep your objectives in mind every step of the way. Engineering jobs are plentiful in the current tech era, engineers are in high demand. There are only wrong reasons to stay in a role for too long and grow complacent.
Have a wonderful week ahead everyone, hope at least one of these articles is exciting reading material for you, and made you think for a moment
