News & Highlights

29.9.17 - TDA and big data fMRI review accepted! Out now IEEE Transactions on Big Data, you can find it here.

8.8.17 - SCM accepted and blog post! Our paper on the simplicial configuration model has been accepted on PRE. Stand by for the journal version! Also our paper on the shape collaborations got some attention.

24.8.17 - New paper out! Our paper on the structure of collaborations under the lens of topological data analysis is now available online.

11.8.17 - New paper in collaboration with Jon Cohen’s Lab from PNI and Intel on the limits of parallel capacity in neural networks. You can find it on the ArXiv here.

27.7.17 - We are going to Barcelona! We just got notified that Esther Ibanez Marcelo is going to present some of our results on the topological effects of LSD and psilocybin at Macfang this autumn. Stay tuned for a preprint later this summer.

Selected Publications

We study the structure of scientiļ¬c collaborations under the lens of topological data analysis. We show that it provides a natural extension of triadic closure which we find to be strong in the ArXiv, that the distribution of collaboration sizes is conserved across disciplines and that homological cycles bridge important portions of the underlying community structure.
In EPJ Data Science

The ability to learn tasks and generalize themis one of the most remarkable characteristics of general intelligence. So is the ability to perform multiple tasks simultaneously. Here, we show that these two characteristics are in tension, reflecting a fundamental tradeoff between interactive parallelism that supports learning and generalization, and independent parallelism that supports processing efficiency through concurrent multitasking.
in arXiv

We provide a review of recent research and development in multivariate and machine learning methods-based gait analysis that can be applied to big data analytics and describe how TDA can contribute to the field.
In J. Med. Bio. Eng.

Simplicial complexes are a popular way to describe multi-node interactions explicitly. We propose here a candidate to a principled null model allowing for easy comparison with empirical data. We provide an efficient and uniform Markov chain Monte Carlo sampler for this model and show its usefulness in a number of case studies.
Accepted in PRE Rapid Comm

We identify a fundamental entropy barrier for time series forecasting. For most diseases this barrier is well beyond the time scale of single outbreaks, the forecast horizon varies by disease and we demonstrate that both shifting model structures and social network heterogeneity are the most likely mechanisms for the observed differences in predictability across contagions.
In arXiv

We use a simplified network representation, called scaffolds, to show dramatic changes in the mesoscale homological structure of brain activity under the effect of psilocybin.
In JRS Int.

Recent Publications

More Publications


ISI Team