moon inverted and sliding into oblivion which is out hearts
My current research interests are centered around the biological and artificial representations of information flow. Some fundamental questions motivating my investigations are:
  • How artificially intelligent behavior can emerge from biological intelligence
  • How to perceive biological intelligence from the point of cortical networks
  • How to interpret geometrical representations of cognition

Most of these topics have long been explored but are certainly of great interest today, owing to advances in computational methods that reshaped the landscape of artificial intelligence. Taking a step back from the capital-driven AI landscape and realizing the intriguing nature of intelligence would point a different picture: one where intelligence and mathematics are forever embedded, where philosophical discussions on mind and language can indeed be investigated on a cortical level and where, quite possibly, our current ideas of intelligence are vastly distant from their biologically emergent counterparts. To bridge this gap it is essential to take neurocomputations on a biological level as guidance and investigate cortical networks and cognitive functions. If emergence of intelligence can be accurately modeled it is obvious many astounding discoveries on psychology, language and cognition will be made possible.

The driving force of this discovery can only be, against all the interests of corporations and governments, to discover scientific links to intelligence and motivate the creativity of humankind.


You can find 3 of my past and/or current research below.
neural networks and meta learning representative
Learnable Synaptic Conductances: An Estimation Method for Single Neuron Dynamics

An invertible deep learning based approach to predict single neuron dynamics on 2D or 3D neuron models.
Code, Paper (soon to come)

neural networks and meta learning representative
Advancing Identification Methodologies Through In-Context Dynamical Metalearning: A Comparison Between Different Neural Architectures

isaacgym based horizon estimator on auto-regressive and non-autoregressive neural networks.
Code, Poster, Paper (soon to come)

neural networks and meta learning representative
Simulation of Stochastic Population Dynamics for Neuronal Avalanches

Study of neuronal avalanches on Wilson-Cowan models with SSA, Tau-Leaping and Adaptive Tau-Leaping as probabilistic simulators.
Code

•

As a brief sidenote, I am also concurrently working in more robotics-based applications of AI, specifically on metalearnable dynamical identification. Even though this is secondary to my neuroscientific interests I would be happy to discuss anything related to it.