Meta-learning three-factor learning rules for reward-driven training of RNNs
Reinforcement learning plasticity rules for reinforcement training of RNNs
started as a postdoc independently (May 2025) 
Projects
Continual learning in RNNs
started in Methods for Computational Neuroscience 2024 Summer School - with collaborators who will be disclosed in the future 
Contextual reconfiguration of representations
with myself
started as a postdoc independently (April 2025) 
Multiple adaptation mechanisms account for the emergence of prediction error responses in V1
We identify possible plasticity mechanisms that explain the adaptation of responses to repeated stimulus presentations and the emergence of prediction error responses in a recurrent circuit model with multiple interneuron sub-types
started during PostDoc at TUM 
Shaping low-rank recurrent neural networks with biological learning rules
Biologically-plausible training of low-rank RNNs
with Pablo Crespo, Dimitra Maoutsa, Matt Getz
started during PostDoc at TUM 
Plasticity-driven circuit self-organization on spiking stabilized supralinear networks
How different forms of spike-timing dependent plasticity affect the functional properties of spiking SSNs
with Raul Adell Segarra, Dylan Festa, Dimitra Maoutsa
started during PostDoc at TUM 
Revealing latent stochastic dynamics from neuronal spiking patterns
Inference of latent stochastic dynamics through stochastic control
with myself
started as an independent project during my PhD 
From geometry to dynamics: Learning overdamped Langevin dynamics from sparse observations with geometric constraints
Introducing geometric inductive biases for learning stochastic systems
extended version of the last project of my PhD thesis 
Inferring synaptic interactions and transmission delays
Inference of neuronal interactions from spike train recordings from a geometric approximation of the inter-spike interval generating function of each recorded neuron
with Jose Casadiego*, Dimitra Maoutsa*, Marc Timme
before PhD 
ODEs for SDEs
Introduced a deterministic particle dynamics for sampling probability flows of stochastic systems (now known as Probability flow ODEs)
- with Dimitra Maoutsa, Sebastian Reich, Manfred Opper
PhD project .png)
Deterministic particle flows for constraining SDEs
Proposed a non-iterative formulation of Stochastic Path Integral Control
- with Dimitra Maoutsa, Manfred Opper
PhD project 