Harsha Gurnani

@harshag.bsky.social

1369 Followers 815 Following

Neural dynamics, control, learning. Love talking about the brain, math, poetry, science+arts! Striving for kindness. (She/her) 🌎 Postdoc @ NYU (Prev: UW, UCL, IISc) 🤓 https://harshagurnani.github.io/

Cosyne schedule, Saturday, March 14, 1.15pm to 4.15pm, Poster number 3-161, Dynamics-based alignment across sessions reveals latent neural computation

outside manifold decoders have low controllability

three sets of trajectories on a flowfield. one trajectory is along the flowfield, another trajectory is against the recurrent flowfield, and a third trajectory requires the neural state to firrst be pushed out in a less controllable direction in state space

Equations for different RNN update rules. Equation 1 has a linear decay term plus a term with nonlinearity applied to summed inputs (sum all inputs - recurrent, external, bias). Eq 2 is its discretized form. Equation 3 has linearized dynamics of rnn state x, with a pointwise nonlinearity applied only to the recurrent input, that is, W_rec phi(x). Equation 4 is the same as eq 3 for hidden state or potential x, but the firing rates are additionally obtained after applying the nonlinearity to the hidden state.

Advertisement by Gatsby Unit and SWC in London about their joint PhD program, with applications due by 11 Nov 2024.