Posts by tiago@social.skewed.de
(DIR) Post #Amzr00tNsefbiZXz2e by tiago@social.skewed.de
2024-10-14T12:14:50Z
0 likes, 0 repeats
Periodic reminder that finding communities in networks using modularity maximization is not a serious methodology.https://skewed.de/lab/posts/modularity-harmful/
(DIR) Post #Amzr2bdKFWUKFOGvrc by tiago@social.skewed.de
2024-10-14T12:15:19Z
0 likes, 0 repeats
Some people still seem to believe that the so-called “no free lunch” theorem implies that all community detection methods are equivalent.This isn't true in the slightest—neither in practice nor in theory.https://skewed.de/lab/posts/free-lunch/
(DIR) Post #An2FGdmjpeVkBuot6W by tiago@social.skewed.de
2024-10-15T15:56:11Z
0 likes, 1 repeats
🚨Job alert!🚨Come join us at the Inverse Complexity Lab! @invcomplexity We’re hiring a post-doctoral researcher to join our group at IT:U, Linz, Austria.https://skewed.de/lab/call.html Deadline is 30 Nov 2024.We are looking for an early-stage or more advanced postdoctoral scholar who is interested in building on our ongoing projects, or developing their own research agenda related to inverse problems in network science, complex systems modeling, and/or connections to machine learning.This position is not bound to a particular research project, and the successful applicant will enjoy intellectual independence and freedom to choose research topics.We’re seeking applicants with a background in Physics, Computer Science, Statistics, Applied Mathematics, or related fields, and with an interest in interdisciplinary research.This position is guaranteed for 3 years. The gross salary range is € 66,532 to € 70,000 (indexed for inflation), depending on previous experience. The employment conditions in Austria include completely free health care (also for family members), social security benefits, 25 days per year of paid vacations, flexible working hours, and possibility of home office.In addition, IT:U will provide a KlimaTicket—a unified transport pass which gives free access to the entire transportation system in Austria, including trains and local public transport.LocationIT:U is a brand new public university in Austria, which began operations with its first cohort of professors in 2024, and is devoted to interdisciplinary research. It is co-located with the Johannes Kepler Universität in Linz, at the Science Park campus.Linz is the third largest city in Austria with a population of 200k, centrally located only 1h15m away from Vienna by train, and short connections to Germany, Italy, and other neighboring countries.For more information, see the call at https://skewed.de/lab/call.htmlSpread the word!#networkscience @academicchatter
(DIR) Post #AnWa0kNHB1uxigagVM by tiago@social.skewed.de
2024-10-30T07:09:57Z
0 likes, 0 repeats
Data science papers about political polarization, gender inequality, academic careers, social media, etc: too many to count.Data science papers about the most well documented genocide in history: zero.https://forensic-architecture.org/investigation/a-cartography-of-genocide
(DIR) Post #Anh5ZDHmdscM5vU4cy by tiago@social.skewed.de
2024-11-04T08:50:41Z
0 likes, 0 repeats
Reminder that we've got a 3-year post-doc position open at the @invcomplexity lab at IT:U, Linz, Austria!The deadline is Nov 30.Please get in contact if you have any questions about the position.https://social.skewed.de/users/tiago/statuses/113312253289238014
(DIR) Post #AnttKXRp74DuHOpcbg by tiago@social.skewed.de
2024-11-10T13:04:49Z
0 likes, 0 repeats
Specially good issue! Recommended!https://mathstodon.xyz/users/manlius/statuses/113448242387120699
(DIR) Post #Anw5yRYy0137NTt4pU by tiago@social.skewed.de
2024-11-11T14:35:54Z
0 likes, 0 repeats
Reminder that we've got a 3-year post-doc position open at the @invcomplexity, IT:U Linz, Austria, on complex systems and network science!The deadline is Nov 30.Please get in contact if you have any questions about the position.More information at: skewed.de/lab/call.html#complexsystems #networkscience @academicchatter @complexsystems @networksciencehttps://social.skewed.de/users/tiago/statuses/113312253289238014
(DIR) Post #AnyGdg9Y2w4SMIDdxY by tiago@social.skewed.de
2024-11-12T15:44:49Z
0 likes, 0 repeats
Mastodon developers should reflect why this is happening over at Bluesky and not over here... Despite the initial advantage that Mastodon had.They won't reflect or care enough to look past their ideological zeal. But they really should.https://fediscience.org/users/UlrikeHahn/statuses/113468988673812925
(DIR) Post #AnyIHMtbACdKAeFQx6 by tiago@social.skewed.de
2024-11-12T16:03:13Z
0 likes, 0 repeats
The way that this set up at Bluesky is obviously more intuitive, user friendly, and conducive to self-organization.It's built in the interface and everyone can curate their own list and share it.It's far better than what we have in the fediverse. These directories do not compare to starter packs.Stop blaming the users.https://social.chinwag.org/users/FediThing/statuses/113469817542409948
(DIR) Post #Ao1eyfIzNFXIBW3Rnk by tiago@social.skewed.de
2024-11-14T07:01:41Z
0 likes, 0 repeats
Some hard truths from @pluralistic: https://doctorow.medium.com/https-pluralistic-net-2024-11-02-ulysses-pact-tie-yourself-to-a-federated-mast-b2f89bb5b4d8It misses only one fact: Mastodon was born shitified!I still believe (maybe naively) that the free software principles are strong and people will eventually unshitiffy it. But this hurdle will cost it a unique moment in history.
(DIR) Post #Ao1gShZbbdEVrPiHtQ by tiago@social.skewed.de
2024-11-14T07:18:18Z
0 likes, 0 repeats
@mpldr @pluralistic The exact same features that keep people away and are driving the mass migration to BlueSky instead of here. This platform has too many restrictions, which makes it more difficult to have meaningful interactions.For example:1. The purely sequential timeline means you miss a lot of relevant posts, and it's much harder to keep track of the important conversations.2. The lack of quotes means you can't link or make references to what others say.This platform added federation, but unnecessarily removed a lot a features that existed on twitter, instead of improving on them.Instead of fearing "the algorithm", mastodon should have given the option for users and instances to control the algorithm of their own feed.Bluesky added the option to "detach" a quote from a post if it turns out it's getting unwanted attention. A simpler, more elegant solution to the (frankly, exaggerated) potential abuse problem, than simply not having quotes at all.In short, there's less friction there, and here it feels like a pain.
(DIR) Post #Ao1lxLUYFoKC7jmTCq by tiago@social.skewed.de
2024-11-14T08:19:52Z
0 likes, 0 repeats
@kawie @mpldr @pluralistic Of course I read the article.Can't you really understand my argument?I obviously don't mean that Mastodon is locked in, only that it's a shitty experience.That's a perfectly valid use of the term "enshitified", and underlies the general problem.
(DIR) Post #AoMm4SrnJSorLLCnmC by tiago@social.skewed.de
2024-11-24T11:30:06Z
0 likes, 0 repeats
In a space of a couple of weeks my number of followers in Bluesky overtook the number on mastodon — starting from about 200 or so.All academics are there.You post something there, people react.Here it's either silence or zealot, humorless bickering.The result of an engineered, restrictive platform that puts way too many railguards on how people are supposed to interact with one another.I still have hope in the fediverse, but what a colossal waste this platform has been so far...
(DIR) Post #AoN1nh82dGSxv5pXbk by tiago@social.skewed.de
2024-11-24T14:26:20Z
0 likes, 0 repeats
@philippsteinkrueger I'm leaving my account here. I still think that I at some point there will be a reasonable fediverse interface, possibly a forked version of mastodon, with the missing pieces implemented. Maybe this is where people will migrate to when bluesky gets shitified — although it took not only Musk but also a Trump victory to people finally move away from Twitter. I'm not holding my breath that something like this will hapen with Bluesky soon. Also, there's the possibility of fediverse integration with Bluesky — even threads did it to some extent.It's just so depressing that things could have been so much better...
(DIR) Post #ApFKKr4m37LlUy39NY by tiago@social.skewed.de
2024-12-20T19:08:24Z
0 likes, 0 repeats
🚨 Job alert! 🚨IT:U is hiring 5 professors (open rank) in Data Science, very broadly defined!Focuses include (among others): 1. Theoretical Foundations of Data Science2. Data Science in Biological Systems3. Data Science in Human Mobilityhttps://it-u.at/en/careers/professor-of-data-science/IT:U is a brand new public university in Linz, Austria 🇦🇹, seriously devoted to interdisciplinary research, and offering excellent resources for research groups.Spread the word!For questions, please get in touch! 2/2
(DIR) Post #AsGtkptVcUk41j4kBE by tiago@social.skewed.de
2025-03-21T06:56:16Z
0 likes, 0 repeats
Happy that this just came out on PRX: “Network Reconstruction via the Minimum Description Length Principle”https://journals.aps.org/prx/abstract/10.1103/PhysRevX.15.011065You can find the previous explainer thread here: https://social.skewed.de/@tiago/112377014076169811And the code here: https://graph-tool.skewed.de/static/doc/demos/reconstruction_indirect/reconstruction.html
(DIR) Post #AsGuBZ1GT6VRVcRpom by tiago@social.skewed.de
2025-03-21T07:01:09Z
0 likes, 0 repeats
New paper on ArXiv: “Uncertainty quantification and posterior sampling for network reconstruction”TL;DR; We present an efficient method to sample the entire ensemble of possible network reconstructions that are compatible with an indirect observation, e.g. a dynamics.https://arxiv.org/abs/2503.07736Explainer thread here: https://bsky.app/profile/tiago.skewed.de/post/3lk6jnqeut225Code here: https://graph-tool.skewed.de/static/doc/demos/reconstruction_indirect/reconstruction.html
(DIR) Post #AtRcP7Ru7x0LNNwJe4 by tiago@social.skewed.de
2025-04-25T08:55:40Z
0 likes, 0 repeats
@aleabdo It was the first time these kids saw perf!!
(DIR) Post #AyxqdwVsyYnIiSCpou by tiago@social.skewed.de
2025-10-07T11:26:39Z
0 likes, 0 repeats
I've been porting some code from C++17 to C++23. Many nice improvements in the language!A side note: Clang is so much worse than GCC in almost everything, from compilation times, runtime performance, to C++23 coverage... Staggering!GCC alone is a good reason to dump MacOS for GNU/Linux (among many others).
(DIR) Post #B0uuvVXO5Whe5qSRxA by tiago@social.skewed.de
2025-12-04T21:14:41Z
0 likes, 1 repeats
I wrote a blog post about the often stated but never explained assumption that communities in graphs should always be connected. This “connected cluster axiom” is inconsistent with statistical significance and null models that underlie the most widely employed methods.https://skewed.de/lab/posts/connected-clusters