3 minute read

The first step in any research cycle is to choose a topic. This time around, I used a decision matrix to rank a list of topics by their impact, neglectedness, feasibility and how well they fit my interests, skills and lifestyle.

Since I was unsure of some of the scores I used in my matrix, I took my top five topics and spent a day or two learning more about each one.

  1. AI for translating unseen languages
  2. Extraterrestrial technosignatures
  3. Recognising AI sentience
  4. Biosignatures
  5. AI misuse: pathogenic DNA

This process helped me to see where I could contribute the most with the time, skills and resources I have available.

Using machine learning and the scientific method to solve an important problem is what I am aiming for.

Evaluating research topics

The topic that best matched my background (biology and machine learning) was preventing the misuse of AI for generating pathogenic DNA. I ruled this topic out, however, since I want all my work to be open and in this case that would probably do more harm than good.

AI for translating unseen languages came out top in my decision matrix, partially due to the impact it would have on non-human animals, which I place high value on. There is not yet enough data available for me to make a contribution to this field using machine learning.

I enjoyed learning about artificial sentience and still think that it is a very important topic, given how much suffering not recognising it would incur, but I could not see any concrete problems for me to tackle right now.

The topic that best matched my interests was exoplanet biosignatures. I have always wanted to work on finding and modelling life on other worlds and studied astrophysics initially before switching to biology, hoping to move into astrobiology at some later point in time.

Technosignatures are related to biosignatures. Whereas different biosignatures are generated by different forms of life, different technosignatures are generated by different forms of civilisation. Technosignatures may be controlled and/or used intentionally (e.g. for communication with other civilisations) – i.e. they require intelligence – whereas biosignatures are a by-product of any living process. The search for exoplanet technosignatures is therefore the search for signs of intelligent life elsewhere in the universe.

Focus on technosignatures

I have decided to focus my research efforts on technosignatures initially and here’s why:

Technosignatures is a neglected topic

From my limited reading around the topics, the technosignature field appears to be much more neglected than biosignatures. These days it is mainly funded by private individuals. Biosignatures and finding habitable planets have much broader support from space agencies. This gives technosignatures higher priority in my decision matrix.

Finding intelligent life elsewhere would be high impact

Detecting any life elsewhere would have enormous consequences for how we see life on our own planet and may change the course of our longterm future as a spacefaring species looking for new planets to settle on. Finding intelligent life would have an even bigger impact. Communication with another civilisation would be an incredible milestone in human history.

Radio signals are available for analysis now

Data for detecting biosignatures will come soon and there is a lot of groundwork that can be done before that, but technosignatures (in particular narrow band radio signals) could be hidden in existing data and machine learning could help uncover them.

Detecting technosignatures is a great problem for machine learning

Finding anomalies and classifying signals in large data sets is something machine learning algorithms are very successful at and I spent several years developing deep learning models to do just that on other types of data. So this topic is a great match for my current area of expertise.

Other discoveries and learning opportunities are possible

Due to the overlap, work on technosignatures may also help with the hunt for biosignatures and habitable planets. In addition, any anomalies detected that are not technosignatures might lead to the discovery of other astronomical phenomena. All of these areas might be future research topics for me and I plan to learn as much as possible about them while working on technosignatures.

Astrobiology sparks joy

Finally, curiosity and passion are key for success in research and astrobiology is a field I’ve wanted to work in for many years. I am looking forward to getting started!

Next steps

A proper literature review of the topic is the next step in my research process. I will spend several weeks on this and will post updates as I go along.