I’ve been inspired by reading Tom Stafford’s 2016 research review. The amount of tangible outputs this year has not felt in proportion to the amount of effort being put in. This is obviously going to happen when projects take a while to complete, nevertheless there can be frustration and feelings of being ineffectual. I can recommend doing a research review like this, it was quite cathartic.
Theme 1: Attention / Bayesian cognitive models
My focus has shifted to Delay Discounting (see below), so I’ve restricted my work on Attention to a few big things. I’m certainly not ending my research here, but it has taken a back seat for a while.
- Upcoming workshop at Carnegie Mellon University in 2017: I have been working behind the scenes with a group of awesome people to set up a workshop in Vancouver that will hopefully result in an important position paper.
- Thinking: I made very slow progress on a nascent idea which potentially has important implications for the field of attention research in general. Hopefully I will get time to bring this to fruition in 2017.
Theme 2: Discounting and Valuation
A little while ago I decided to switch focus of my Bayesian cognitive modelling approach, and apply it to higher-level decision making tasks. The focus is on temporal discounting, exemplified by the marshmallow task which examines how people’s valuation systems trade off smaller sooner rewards vs larger but later rewards. The strategy was to contribute 2 methods papers to the field, which would then aid me (and others) with empirical research questions.
- Paper 1 (+ software): The methodological paper Hierarchical bayesian estimation and hypothesis testing for delay discounting tasks was published in Behavior Research Methods, and was covered in a blog post by the Psychonomics Society.
- As part of that paper, I made the Discounting Toolbox (coded in Matlab) available for other researchers. It’s had a good level of interest, and I’ve been in touch with people from India, Europe, USA and Canada, working with them on various fixes and improvements.
- Paper 2 (+ software): I’ve been working on a very very cool methods paper with Tom Rainforth (Oxford). This is currently in stealth mode, but we are looking forward to submitting this to a pre-print archive, then to a journal in early 2017.
- Paper 3: We are very close to submitting an empirical paper on the effects of fasting on delay discounting. We’ve got some really cool empirical results, and have done some pretty pleasing Bayesian data analysis.
- Paper 4: We have some very exciting results + theoretical contribution to discounting (in my opinion). This is currently in stealth mode, and we have another experiment to run, but I’m looking forward to getting this submitted in 2017.
- Collaborations: In January this year I was fortunate to be invited to a research away day with Neuroscience at University of Dundee. This was remarkably fruitful and have 3 concrete collaborative projects which will kick off in 2017.
- New staff member secured: Mid 2017 should see the arrival of a new member of faculty in the Division of Psychology. We have strong overlapping interests, so I’m grateful to all the behind the scenes work put in to securing another position in our Division.
- I’m very happy to have started a collaboration with William Tipples, and we will be submitting a Stage 1 grant proposal early in 2017.
- There are some other collaborations which are in the early stages, so hopefully more to report on in 2017.
Learning & Skill development
- Matlab: My conviction to leave Matlab, in favour of Python, basically did not happen yet. I’m not beating myself up about this. Previously I was having a bit of a “grasses are greener” syndrome. While it does have it’s issues, it is a solid and reliable language where you can get serious work done, especially when you get into the more advanced aspects of the language (e.g. Object Oriented Programming, and unit testing).
- R: I’ve now got basic competence with R, and have been using R and RMarkdown to conduct some pretty funky Bayesian data analyses.
- Julia: Like lots of other people, I have become rather interested in this new programming language. In short, this is a high-level language which is very similar to Matlab, but with the performance of C, and with a number of really advanced features such as meta programming. I’ve not spent a lot of time on this, this is more of a longer term skill development thing. When it comes to starting another project from scratch, I’m still not decided between Matlab, R, Python, Julia.
- Bringing software engineering approaches into scientific computing. I’ve been learning things such as version control, unit testing. Sometimes learning this stuff has felt a bit like a procrastination activity, but on the whole, I see this as investing in the quality and rigour of the work I can produce. This is especially the case for any scientist who works on advanced data analysis, cognitive modelling, toolbox creation, etc.
Other fun things:
- Submitted a Fellowship proposal, but failed to win it.
- External presentation at Neuroscience, UoD.
- Workshop on Decision By Sampling at the University of Warwick. This was really bloody good.
- I’m half way through a couple of MOOC’s. The Mind is Flat by Nick Chater, and Julia Scientific Programming which is good, but perhaps more suitable for someone new to programming.
- Attended ViiHM meeting on human and machine vision in a lovely hotel in Bath. Also attended a ViiHM satellite meeting where 12 academics discussed issues around salience in human and machine vision.
- Internal PhD examiner for Dr Clemens Speth.
- Convened a PhD viva for Dr Glenn Williams.
Overall, I am very happy with progress this year. Looking back at it, I’m actually quite surprised at all this, given personal challenges I’ve faced in 2016 and that all the teaching and admin duties got done. Although I am now shattered and having to think about some work/life balance issues. Thanks very much to my current and embryonic collaborators, including my dissertation students and undergraduate research assistants!