..

A PhD thesis plan

Introduction

It’s been 9 months already since I’ve started my PhD, and during a Spirals team seminar I had the opportunity to present my work during 30 minutes. Since I was at the time working on several projects in different thematics, unlike my colleagues, I decided not to focus on one of these projects, but instead to create a more coarse presentation, where I would talk about all ongoing projects of course, but also about my research career, past and future.

This scared me a lot since back then, I felt like I was not making any progression on any of my projects, a lot of those inherited from the last 3 years I spent as a research engineer in the team. Thematics, even if very interesting, seemed to have no link at all with each other, making me doubt really hard about the global direction of my thesis.

This blog post is a good occasion for me to link all of my work together, and hopefully feel better about all of it.

Plan

Incentives

Real world impact of technologies

Smartphones are the thing1: encompassing a number of features (phone book, alarm, music player…), they replaced mobile phones and are present in almost every aspect of our lives.

However, just like the “cloud”, such devices do not come at a free price, they require lots of resources to be produced, such as water, ores and rare-earth elements: these resources are usually extracted far from the view of final consumers, in conditions that do not necessarily respect human rights, and their extraction has a big impact on the ecosystems; we even know since 2010 that there will not be enough ores on Earth to continue living the way we’re doing it right now2.

Making smartphones has a direct impact on the planet.

In this context where we know that in some tens of years we won’t be able to produce as much hardware as we’re producing in 2023, several solutions:

  1. Use our devices longer (don’t buy the new iPhone or Pixel as soon as they’re released);
  2. Reuse our devices (reconditioning);
  3. Recycle our devices.

Recycling smartphones is something that’s very hard, costing lots of money to retrieve small quantities of ores, not to mention that recycled materials are very rarely used in building new smartphones.

Reusing is gonna be hard, because people usually keep their unused devices3: “65% of smartphones are renewed while they are still working, and 45% are kept instead of being immediately sold, given or thrown away.”

Only one remaining solution:

We MUST use our smartphones longer.

Compromised privacy

If you're not paying for the product, you are the product.

You probably heard this a few times, and this applies to the whole Internet nowadays: whether service you’re using, chances are high they’re displaying some custom ads and/or gathering personal information about you to get some money.

I personally do not think free internet should be supported by personal data trafficking, and neither does the European Union: in the last years, it’s been pushing for consent protection (i.e. platforms shouldn’t sell your data without you explicitly agreeing with that).

However, even if the GDPR is supposed to protect users from those doubtful depraved practices, Meta found a way to bypass the regulation by forcing its members to give their personal data, the only alternative being to pay.

And if you’re still willing to trust big corporations to handle your data, consider that as soon as your data is uploaded to their servers, it does not belong to you anymore; actually, their general terms of use certainly mention they can do whatever they want with your data, including accessing4 it “to improve their services”.

Where could we improve?

I do not want to rely on shady business practices that make money on my back. I want my privacy to be respected; as such, I would like every piece of data leaving my phone not to betray my privacy, and all of my data be tested accordingly: for instance, before sharing location data with a third-party service, my location data should not reveal any information about my home or place of work.

Regarding the energy consumption of our smartphones, as developers, we should not assume end users have powerful configurations; actually, Google releases an Android version that is dedicated to low powered devices, Android Go5. Even with powerful configurations, for the sake of keeping our smartphones as long as possible, we have to create applications that are as CPU/memory/storage/energy efficient as possible.

Even if this should not be an excuse to neglect potentially powerful tools such as machine learning (please don’t get me started on so-called artificial intelligence) to address mankind issues, we probably want to limit the development and usage of greedy and/or borderless useless technologies.

Research directions

For the sake of privacy protection and global energy consumption reduction, we want to bring data processing from cloud providers to as close as possible from end user devices, if not directly on end users’ smartphones. To do so, we plan to study and overcome the limits preventing:

  • big dataset storage;
  • heavy data processing;
  • no-Internet (device-to-device) data exchanges

on constrained devices.

  1. “Number of smartphone users worldwide from 2013 to 2028”, Statista, last accessed on Jun. 29th, 2023 ↩

  2. “Mining rush in the 21st century: how far will the limits be pushed? (French)”, Aurore Stephant, Jul. 12th, 2022 ↩

  3. “Infographic on cell phone reuse and recycling (French)”, e-recycle, Feb. 23rd, 2021 (last accessed on Jun. 29th, 2023) ↩

  4. “Google wrongfully bans father for taking naked pics”, Information Age, Aug. 23rd, 2022 (last accessed on Dec. 18th, 2023) ↩

  5. Android Go edition webpage: https://www.android.com/versions/go-edition/ ↩