Our paper Simulation-based marginal likelihood for cluster strong lensing cosmology got accepted for publication a few days ago. The astronomy problem posed was about how we can use observations of the phenomenon of strong gravitational lensing by galaxy clusters to infer the values of cosmological parameters (or potentially the cosmological model itself). Each hypothetical universe would result in different galaxy cluster compactness and universe expansion history. Both consequences ultimately affect the Data: which galaxy clusters would be selected in a flux-limited survey, their mass, and their effectiveness as a gravitational lens.
Anyway, we hit a snag because one of the things you need to do this is to be able to say how likely you are to observe what you did under different hypothetical universe models (or something more specific, like the amount of dark matter) – this is the likelihood. Since we use large-scale computer simulations to model non-linear structure formation, we can’t write down a likelihood function – a common problem in fields like epidemiology, genetics, and geology as well.
Our solution ended up being something in the vein of approximate Bayesian computation: use summary statistics. Except in this case, the summary statistics have to be something you infer using the usual Bayesian approach and you end up with a joint PDF (rather than a set of scalar values) for a given dataset (whether real or simulated). Then instead of a kernel distance metric and threshold distance, you calculate the zero-shift cross-correlation of the summary statistics PDFs (ssPDFs) for the real and mock datasets – that’s your likelihood!
This paper is particularly important to me because
- it was my last as a professional astronomer
- it involved throwing away the original paper we had written, re-framing the question (away from the notions of tension and consistency with standard cosmology), and implementing creative ideas.
- we also had the best (anonymous) reviewer. We were both stubborn and pedantic but utterly respectful. This is how the refereeing process should be!
The idea is finally out there and I can now get on with several follow-up validation tests and applications in other fields of research. We could (and hopefully will) demonstrate the full posterior inference/model comparison in the kind of research problem where the simulations are relatively quick or emulators are available. There’s also much work to be done in the choice of summary statistics and inference of ssPDFs. Plus the logistics of calculating the cross-correlation with discretely sampled functions will get tricky when the ssPDFs are higher-dimensional, so any help with that would be greatly appreciated!
I left astronomy formally 18 months ago and recently finished up a computational modelling job in public health. With a little bit of time and perspective I now present to you my 10-step Program to Quitting Astronomy. Or any line of work. I’m not arguing that you *should* leave. This is simply about the process. Some of the details are specific to my aims, but I’m also incorporating a range of advice from my friends and colleagues, who each have their own unique goals. These steps don’t have to happen in exactly this order, but you’ll want or need to do most of these at some point. Good luck!
The 10-step program
- Acknowledge your feelings
- Identify your skills and interests
- Decide what to do next
- Find a mentor
- Go public
- Form a study group
- Re-train and re-brand
- Settle in
- Mentor others
So: you’re thinking about quitting astronomy. Why are you considering leaving? Are the problems specific to the job? Or a consequence of unusual circumstances? Could they exist in all workplaces? What future do you picture if you continue? What do you expect if you leave? What are the risks of leaving? What are the risks of staying? What are you afraid of? Do some honest soul-searching; talk to friends; read some quit-lit if you like. I had never planned to leave, so I struggled and this stage took me a whole year. At the end of the day, you might decide to stay. That’s OK. But your mental health is your health, so make that a priority whatever you decide.
If you make the decision to leave, then recognising your own skills and interests is crucial, especially when you come to re-branding (see Step 7). Make a mental list of intellectual interests and priorities (incl. workplace environment, autonomy, opportunity for progression, big picture, salary & benefits, flexibility & stability, family & relationships). Which of these are dealbreakers? My original list had astronomy near the top, and I had to gently but purposely delete it so that I could acknowledge its existence but then turn my attention to everything else.
Decide what to do next *generally* speaking, both in terms of your career and other life choices (based on Step 2)
. Take every opportunity to hear from friends and acquaintances outside your regular professional sphere about what they do and who they work with. Ask whether they hire people with your background and skills. My first resource was the JobsForAstronomers website, which – back in 2012 – introduced me to the now famous Insight Data Science Fellowship. Insight has since spawned several other satellite initiatives and similar data science programs are popping up everywhere. At that point, however, what mattered wasn’t the programs themselves, but rather realising that my PhD would really be useful and valued elsewhere, that I wouldn’t be wasting my achievements, and that a path existed somewhere to mediate my transition. At first I was drawn to data-science, but later decided (after going back to Step 1) that I still wanted to do some academic work as an applied statistician but in a different discipline; now I sit somewhere in between.
Seek out a mentor or three. Find ex-astronomers who are already working in a field you’re interested in joining (preferably people who you can relate to), ask them about what they do day-to-day, good companies/institutes to work with, where/how to look for jobs, and CV/application/interview tips. People are almost always willing to step up: advice is free and it’s flattering to be asked. In the last few years, a few Facebook groups have been formed to help with this.
It’s OK to cold-call someone to ask for advice. I did.
At some point you need to let people know your plans and where you’re looking to go. Partly for logistical reasons (so you can ask for reference letters) and partly for your own sanity (you can stop pretending that your sole purpose in life is to get a faculty job). As a bonus, once people know you’re looking for a job, they might just tell you about potential mentors and any openings they hear about. So go public with positivity. There’s no need to be defensive: you’ve made a (big) choice for valid reasons.
Channel your inner career-transition spirit animal and explore new identities
Find a few other people preparing to make a similar move and suggest that you form a study-group that doubles as a support-group. I’ve met these people either in my own department or at astro-conferences (these days people discuss post-astro careers at every single event; thankfully this is no longer a taboo subject). You can swap tips and discuss all the practicalities of steps 7 & 8. You will find out about twice as many job opportunities. Most importantly, you will not be alone in this.
Re-training is simple; we do it all the time, like reading up for new projects or teaching ourselves new programming languages. But re-branding feels like a bit of a dirty word for a researcher. The fact is we’ve always been one-person companies: creating a product, marketing, consulting, HR, all wrapped up in a human package. Re-branding means consciously influencing your potential professional community’s perception of you, communicating your expertise, what values you stand for, and ultimately if and where you would fit in with their team. Both re-training and re-branding could involve applying to fellowships and internships, completing MOOCs and kaggle-competitions, attending short courses & workshops, conferences & hackathons, meetups & career seminars, blogging demo-projects, and taking to social media.
Also, take some decent professional photographs – you’ll need them.
Beginning the transition involves a few things: job hunting, re-designing your CV, learning new interview skills, and time. Your mentors and support-study group will be crucial here so make the most of them. A good recruiter can do wonders, and networking is key. I found social-media to be very useful for serendipitous job ads, although I rely more on job-listing websites and newsletters. Interviews can be quite different to what you’re used to, so treat each one as a learning opportunity. Are companies reluctant to hire astronomers? I can’t speak for all companies, but I can tell you that to improve your chances you should avoid being seen as a risky hire: do your research, adapt to speak their language and understand their stakeholders’ needs. So… go out and score that job!
Optional extra: transition between countries…
Adjust and settle into your new job & new life. Get used to answering the question: “So, what do you do?”. I still almost always mention astronomy. No job is perfect, but sometimes it’s not until after we’ve had one job outside our previous career that we realise what we’d taken for granted. That said, nobody I know regrets their decision to leave. It’s OK to miss your research topic. I miss gravitational lensing terribly – the research, my colleagues, and the conferences. I still like to share major science news with my family, friends, and bus drivers. But why not get involved with other science disciplines? Why not the tech industries and start-ups? Why not help shape policy? Now that I’ve freed myself from the shackles of lifelong career goals, I have the opportunity to be part of any of it. I am also free, it turns out, to continue doing research astronomy as a hobby, at my own pace, on whatever I’m inclined to do.
Congratulations! You’ve moved on – with some help. Now it’s time to pay it forward. From the moment you go public (step 5), you can expect requests for mentorship or at least a one-off chat. One year on, I was asked to speak on a panel to junior academics about alternative career paths (keep an eye on emails from your previous department and volunteer when you see calls for speakers). Your ability to help others transition will be useful for at least a few years, so go out and do your duty. Maybe write a blog about your experience.
The under-representation of women and minorities within certain areas of academia, business and education, particularly in high-paying jobs and managerial roles is well documented. Anti-discrimination laws can only do so much when unconscious bias is rife. This is where affirmative action comes in, controversial as it is, in the form of any of the following:
- scholarships for minorities
- prioritising minorities given equal qualifications
- diversity quotas.
The general idea is to account for the “selection function” that is privilege. In science, data is collected from objects which satisfy certain criteria. Often, the data we really want is convolved with the selection function. An example from astronomy is when you include only galaxy clusters brighter than a certain threshold in your sample; you are more likely to include objects: of higher mass; at low-medium redshift; that have cool-cores. If one wants to know the properties of an otherwise “fair” sample, you must perform some kind of deconvolution or correction (easier said than done). That’s what affirmative action attempts to do (also easier said than done). For the rest of this blog-post, I look at recent attempts to improve the representation of women, in particular, as well as responses to these .