Sebastian Morel

Postdoctoral research fellow at HKU Institute of Data Science | Complex Networks, Bayesian Inference, Data Science

[Alec Kirkley @captainkirk1041]

Excited for NetSci-X 2027 in Hong Kong! Send your abstracts by Oct 1st (possible deadline extension coming soon), and see https://t.co/ZzS5NvPQJQ for more details.

[SEES Lab @sees_lab]

Can simple closed-form mathematical models predict human mobility as well as deep learning? In a new paper in @NatureComms we show that the answer is YES

Human mobility is well described by closed-form gravity-like models learned automatically from data https://t.co/5LaWeXB4ih

[Alec Kirkley @captainkirk1041]

**Deadline change!** The abstract submission deadline for SINM has been changed to **February 20th** to give more time for decisions prior to the early-bird registration deadline. Looking forward to seeing your submissions!

[Sebastian Morel @sebastianmorelb]

9/9 And in the PANINIpy package:
https://t.co/EVcXiioVk1

[Sebastian Morel @sebastianmorelb]

New preprint out in collab with @captainkirk1041, where we develop a Bayesian methodology to infer partial rankings (rankings with ties) from pairwise interaction data.

"Learning when to rank: Estimation of partial rankings from sparse, noisy comparisons"
https://t.co/MfHLpBhxi2

[Sebastian Morel @sebastianmorelb]

8/9 Code for the partial ranking algorithm can be found here:
https://t.co/MPGP47e8xB

[Jean-Gabriel Young @_jgyou]

Last week, PhD student Simon Lizotte won the best poster award at @NetSciConf for our work on symmetry-driven embedding of networks in hyperbolic space.

We uploaded the paper to @arxiv as the conference ramped up, and you can find it here: https://t.co/w3NCZvDyv1

[Sebastian Morel @sebastianmorelb]

Code for the greedy agglomerative algorithm can be found here: https://t.co/0upbFjGkVg

[Sebastian Morel @sebastianmorelb]

Our results show that while SBMs can achieve significant data compression, they can also present considerable degrees of spatial discontiguity at low spatial aggregation scales, making them mostly unsuitable for spatial regionalization in this regime.

[Sebastian Morel @sebastianmorelb]

To make things even better, our proposed algorithm is not limited to SBMs but can be applied to a large family of objective functions (including most community detection methods), making it suitable for a wide range of geospatial applications where spatial contiguity is required.