However, the absolute differences between these model groups strongly depended on the measured concept, and we observed strong variance in performance among models of the same group.
We present results for over 50 metrics side-by-side to judge the opportunity costs of choosing one method over another. The results revealed strong variation across different groups of models. Overall, modern methods (transformers and generative AI) outperform the older, simpler ones.
Our project on comparing rule-based, classical machine learning, transformers and #LLM to measure debate quality on social media is finally out at Communication Methods and Measures! ✨✨✨
@ascor.bsky.social @nunetsi.bsky.social
Link here: www.tandfonline.com/doi/full/10....
Thanks my good friend for attending CCS in person on my behalf ;)
Hope this quick overview of data quality problem is helpful for the audience
Would be giving a talk on Data Quality Framework, online… Hopefully can see some of your faces
Hate to miss #CCS2025 :(
But I’m in the background working on bringing complex system to social sciences and humanities in the study of platforms with a new working group Critical Platform Studies at @uwmadison.bsky.social
Sign up to our mailing list (see comment) to join the community!
thinking of calling this "The Illusion Illusion"
(more examples below)
Happy to share this long-overdue project! We found that many real-world event sequences follow a surprisingly similar hierarchically structured pattern, and that multi-timescale memory mechanisms can explain this pattern. Feedback welcome!
arxiv.org/abs/2508.18281
A cool project just ended - very relieved and excited to see it out✨
Stay tuned!
New profile pic on NETSI website taken at the Royal Institution in London.
www.networkscienceinstitute.org/people/justi...
Title slide: "What drives the productivity of scientific labor?" a talk by Aaron Clauset at the Oxford Summer School on Economic Networks 2025
Summary slide arguing that prestige is a structural variable in the scientific ecosystem, and that it is differences in working environment (not pedigree) that explains why elite researchers dominate scientific discourse. Ends with a question: how exactly does environment do this?
Summary slide arguing that working environment (specifically, institutional prestige) explains differences in scientific productivity because there's just more available academic labor at elite places, and researchers use this labor to write more papers. Ends with a question: how much does it matter who you collaborate with? That leads into part 2 of the slides
Slides from my keynote lecture "What drives the productivity of scientific labor?" at the 2025 Oxford Summer School on Economic #Networks. Part 1 of 2: why and how do elite scientists dominate scientific discourse?
aaronclauset.github.io/slides/Claus...
On the side of publishing, i wonder if that is domain-specific problem or a general issue across fields? What I feel like is that most subdomains are created to cater the need for understanding particularities… but perhaps maybe in less niche journals, we do need more overarching theories.
Certainly - quality has little to do with model / theoretical complexity. I guess what i am trying to get at is that generalisation of theories requires testings on edge cases to ensure robustness of the theories themselves, meaning that particularities are only the means, not the ends.
@alexvespi.bsky.social on the simplicity of complexity - mentality (and approach) much needed in all disciplines.
Latest out in PNAS!! Comparative evaluation of behavioral epidemic models using COVID-19 data. Amazing collaboration with @ngozzi.bsky.social and @alexvespi.bsky.social www.pnas.org/doi/10.1073/...
Accounting for complexity. Moreover, even in establishing statistical laws, we are still interested in cases where such laws do (not) apply and therefore we require peculiarities to validate the universality of general models.
Absolutely agree that we need to find regularities and formalism in media effects research.
But not sure how conducive it is to reduce (model) complexity - my intuition is that the current methods are simply bad at accounting for nuance, but methods from complex system and statmec are good at …
Two posters and a bunch of amazing people at @netsciconf.bsky.social with @nunetsi.bsky.social
✨✨
Picture of the members of the Network Science Institute attending the Network Science Conference at Maastricht 2025
Picture of Esteban Moro presenting their work on Mobility Barriers in Cities during the Network Science Conference 2025
#NetSci2025 is a wrap—huge thanks to the organizers for a fantastic edition! So much energy around Network Science and its many applications. Great to reconnect with colleagues and make new (and hiMgher-order!) connections. See you all next year in Boston!
Thanks for sharing our work on the phenomenon of GIGO in the criminal network domain!
@nunetsi.bsky.social @oii.ox.ac.uk @riccardodiclemente.com @lambiotte.bsky.social
I have covered presidential politics for 40 years. This was the most juvenile display by a President and Vice President I have ever seen. Other presidents treated their enemies with more respect. This is a low point and a dark day for the US. Totally embarrassing.
Excited to kick off my fellowship at the Institute for Advanced Study (University of Amsterdam) with my talk: "Knowledge Silos as a Barrier to Responsible AI Practices in Journalism."
📅 Join me on 25/02 at 12:30 in Amsterdam!
More info and registration: ias.uva.nl/content/even...
🚨 New paper out!
We examine how knowledge silos in newsrooms create barriers to responsible AI adoption. When teams operate in isolation, AI tools can be introduced without proper ethical and editorial oversight, weakening accountability.
Read OA 🔓: www.tandfonline.com/doi/full/10....
If you're into network science and have done work on criminal networks/activity, consider sending your work to criminalcomplexity.weebly.com. Organising with Toby Davies, Cecilia Meneghini, Rafael Prieto-Curiel, Rick Quax, Huijuan Wang, Wang Ngai Yeung.
How do autocrats stay in power?
Our new research dives into the strategies of political control—repression, co-optation, and indoctrination—used across 229 autocracies from 1946 to 2010. What we found will change the way you think about authoritarian resilience. 🧵👇
Screenshot of the website of the "Data Donation Symposium 2025"
💥 4th Data Donation Symposium @LMU Munich
🗓 October 9th-10th 2025 (no conference fee)
👉 Empirial work employing data donations, method development, tools & infrastructures, etc.
🔍More info & CfP: data-donation-symposium.ifkw.lmu.de
#OSSEN2025 registration now open until 3 March. With @doynefarmer.bsky.social and Christian Diem from @inetoxford.bsky.social confirmed as speakers.
Without further ado, here’s the amazing speaker lineup for OSSEN 2025!
Experts from across network science and economics will cover topics ranging from financial networks to urban mobility and more!
Don’t forget to apply by 3 March:
www.maths.ox.ac.uk/events/summe...
