Researching information and brains at the Cambridge Consciousness and Cognition Lab and the Imperial MIND lab. Adult figure skater. Creator.https://bsky.app/profile/keenandown.bsky.socialhttps://bsky.app/profile/keenandown.bsky.social/post/3mgrtmd3tps23Ready for the biggest review into Partial Information Decomposition ever? We've gone through the literature and collected every PID measure and property we could find and filled in the gaps! Thanks to Alberto Liardi for his superhuman work here! Check it out below! 👇 arxiv.org/abs/2603.0667811 Mar 2026 12:01 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3mgrtmd3tps23https://bsky.app/profile/keenandown.bsky.social/post/3mgp3fhe4ls2cOne of the best things about consciousness research is that you are forced to appreciate the value of your own experiences, good and bad Being human enough to regret taking the saltiest pretzel is already a miraculous thing10 Mar 2026 09:42 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3mgp3fhe4ls2chttps://bsky.app/profile/keenandown.bsky.social/post/3mfxe73neg223Just got tearful over a commutative diagram (happy tears, which is somehow worse) Time to go to bed28 Feb 2026 23:16 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3mfxe73neg223https://bsky.app/profile/keenandown.bsky.social/post/3mfcpbvi77c2uHappy to announce this new preprint! In it, we use info decomp (ΦID) on fMRI in Alzheimer's and MCI to explore how info-dynamic representations change. AD saw big decreases in synergy ('deductive' information) and increases in redundancy. Check it out here: 👇 https://www.biorxiv.org/content/10.64898/2026.02.18.706630v120 Feb 2026 18:08 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3mfcpbvi77c2uhttps://bsky.app/profile/keenandown.bsky.social/post/3mbtvmeobic2nDidn't post about this at the time, but since we're going back to adult nationals again this year, here are some pictures from the recent Cambridge Classic open! PB for the national series so I'm quite happy 🎉⛸️ Thanks again to coaches James and Hannah!!07 Jan 2026 16:36 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3mbtvmeobic2nhttps://bsky.app/profile/keenandown.bsky.social/post/3maq3up4gp22kHappy to announce that the first major work of my PhD is now available in full in IEEE Transactions on Information Theory! 🎉 We push a set-theoretic version of info theory to new levels, describing 'sets of information' in a bold new way, with surprising insights! https://ieeexplore.ieee.org/document/1127138824 Dec 2025 10:52 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3maq3up4gp22khttps://bsky.app/profile/keenandown.bsky.social/post/3lxhonxxfzs23Went down to visit some family in my homeland (east Kent) Feat. the cliffs south of Dover, at Samphire Hoe (the leftover dirt from the channel tunnel) France was super clear on the horizon while I was down over the weekend!28 Aug 2025 13:48 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lxhonxxfzs23https://bsky.app/profile/keenandown.bsky.social/post/3lwtky4nk322lOne reason why multivariate information theory is hard: When we find the MI between two variables, we use probs drawn by an implicit third variable (the observation) - at least in the frequentist sense. E.g. we find probabilities for different eye colours by aggregating over *people* 🧍 (1/2)20 Aug 2025 13:49 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lwtky4nk322lhttps://bsky.app/profile/keenandown.bsky.social/post/3lwriczkohc2pTried out the new Rainbow Rocket North yesterday! Awesome having a new climbing gym closer to home!19 Aug 2025 17:56 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lwriczkohc2phttps://bsky.app/profile/keenandown.bsky.social/post/3lwoj3a7hrc27I think we need some more statistics on which subjects drink the most caffeine https://archive.thetab.com/us/2017/04/19/ranked-the-majors-that-drink-the-most-coffee-in-america-65418 I refuse to believe that philosophers drink the most (My bet was bioinformaticians or something of that sort)18 Aug 2025 13:31 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lwoj3a7hrc27https://bsky.app/profile/keenandown.bsky.social/post/3lwmqnfyb222lLucky to have caught this while it’s in London for 3 shows Was super refreshing having such an enthusiastic audience!17 Aug 2025 20:41 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lwmqnfyb222lhttps://bsky.app/profile/keenandown.bsky.social/post/3lwjvmc74ik2lTrying some combos! I almost always find myself doing a salchow rather than a flip after an euler Will have to practise it more16 Aug 2025 17:32 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lwjvmc74ik2lhttps://bsky.app/profile/keenandown.bsky.social/post/3lwhpv5fs222rI find it genuinely so interesting how nobody really talks about the problem of quantifying shared information I feel like it deserves a video or *something*, but it’s flown so under the radar The maths influencers are letting us down15 Aug 2025 20:44 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lwhpv5fs222rhttps://bsky.app/profile/keenandown.bsky.social/post/3lwet7jlly22oJust saw a single sunflower in a field of weeds Struck me in a way I wasn’t expecting14 Aug 2025 17:06 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lwet7jlly22ohttps://bsky.app/profile/keenandown.bsky.social/post/3lwcak4nwd22yThe quickest way to make life more interesting is to hunt for unfamiliar qualia - me 15 minutes ago before people started to stare at me on the street13 Aug 2025 16:26 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lwcak4nwd22yhttps://bsky.app/profile/keenandown.bsky.social/post/3lw7xp7hijk2mIn the space of all phenomenal experiences, are isomorphic experiences the same? Do phenomena live in a skeletal category? Maybe someone with more category theory experience than me could tell me if this is interesting/a thing or not12 Aug 2025 18:43 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lw7xp7hijk2mhttps://bsky.app/profile/keenandown.bsky.social/post/3lw4drf2pm22vMathematics is all about conveying arguments. Because of that, elegant maths relies heavily on narrative structure and exposition. Mathematics is prose first, complexity second. I will defend this until I die11 Aug 2025 08:08 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lw4drf2pm22vhttps://bsky.app/profile/keenandown.bsky.social/post/3lvzzmbpd7k2zLogarithmic Decomposition is great for understanding distinctions in qualia and their representations! Some colourblind folks can play along! For more information on the method, check out the papers: arxiv.org/abs/2409.03732 https://www.mdpi.com/1099-4300/27/2/15110 Aug 2025 10:01 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lvzzmbpd7k2zhttps://bsky.app/profile/keenandown.bsky.social/post/3lvy4fyfgts24Something people don’t realise about figure skating in the UK is how many exams there are Start to finish there are 36 tests/levels to do, most of which need to be done in front of a judge in an exam or passed in competition09 Aug 2025 15:46 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lvy4fyfgts24https://bsky.app/profile/keenandown.bsky.social/post/3lvvn6pdtf22gSome of our emergence measures might be suffering from being overly dependent on our choice of variables - most notably some PID measures. See my recent poster on the problem which I presented at Embracing Complexity here: 👇 https://www.ccc-lab.org/uploads/9/8/3/6/98367296/frontiers2025poster.pdf08 Aug 2025 16:08 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lvvn6pdtf22ghttps://bsky.app/profile/keenandown.bsky.social/post/3lvszxpfjis2oI've been having loads of fun teaching once more at Downing College's (@downingcollege.bsky.social) international summer programme again this year! All students on the programme take an online maths course with yours truly before coming to Cambridge for two weeks! https://www.dow.cam.ac.uk/international-programmes/downing-college-international-programme-15-18-years07 Aug 2025 15:19 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lvszxpfjis2ohttps://bsky.app/profile/keenandown.bsky.social/post/3lvqjillmjk22Is there a natural way to construct a (signed) measure on a simplex? Turns out there is - entropy is a great choice! Check out our preprint on this here 👇 arxiv.org/abs/2409.0373206 Aug 2025 15:19 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lvqjillmjk22https://bsky.app/profile/keenandown.bsky.social/post/3lvo4p36r2227Trying some new spin positions and failing at axels05 Aug 2025 16:24 +0000at://did:plc:g7wr72hg7ghj2xtg6dkma4ee/app.bsky.feed.post/3lvo4p36r2227