RT @blake__bordelon: ICML this week! Come by. T PM @LauditiClarissa's work on muP BNNs . W AM, model of place field….

RT @KempnerInst: 'A Model of Place Field Reorganization During Reward Maximization'. @mgkumar138 · @blake__bordelon · Jacob A Zavatone-Veth….

First #ICML2025 conference proceeding (! We (Blake Bordelon, Jacob Zavatone-Veth, @CPehlevan) developed a simple model to better understand state representation learning dynamics in both artificial and biological intelligent systems! Comments appreciated!.

I'm heading back to Singapore for ICLR25! Hit me up for discussions or where to find good food!. #neuroai #home.

I am speaking at COSYNE 2025. Please check out my talk if you're attending the event! #cosyne2025 #cosyne25 - via #Whova event app

We're currently validating some of these model predictions by analyzing neural datasets. If you are interested in exploring this model further or have relevant datasets, please DM me! #Hippocampus #RL #NeuroAI.

Ablation studies in 1D and 2D environments show that inducing these biological representations improve policy convergence and facilitate learning new targets.

Noisy updates in place field parameters drive neural drift while compensatory plasticity mechanism maintains stable navigation behavior.

A reward maximization objective also causes place fields to grow in size and shift backwards towards the start location, suggesting the development of a reward predictive representation.

Place fields rapidly move closer to the target location to increase reward representation. The reorganization dynamics is modulated by the value of a location.

Our preprint with @blake__bordelon, @jzavathoneveth, @CPehlevan is now out! .We develop a simple reinforcement learning model that recapitulates 3 disparate hippocampal dynamics. Through ablation studies, these representations improve the speed and flexibility of policy learning.

The chapter introduces computational methods for biomedical sciences e.g. artificial neural network models can offer new hypotheses for understanding biological computations about learning and disorders, and algorithms to predict protein structures can minimize manual screening.

My first book chapter "Trends Innovations Challenges in Employing Interdisciplinary Approaches to Biomedical Sciences" is now published!. . #BiomedicalScience #neuroai.

RT @andrewtanyongyi: @patrickmineault Figure 1 in this computational study by @Ganeshk92 tries to explicitly conceptualize the proposed age….

The summer school was tons of fun because of @blake__bordelon @cogphilosopher @Adam_manoogian @nacloos @Z_Friedenberger @tyboydmeredith @jli_neuro @aesagtekin @BorromeanLink @fredpckuo @_Tobias_Thomas @ClaudiaMerger @schmutz_val @stefsmlab @Dongyu_Gong @milleire and many others!.

The Analytical Connectionism summer school at the @FlatironInst was a great intro to theory & comp methods to tackle questions in #neuroscience. Made new friends & we built an agent to model decision making in schizophrenic patients. Eager to explore similar topics soon! #neuroai.

Fantastic time at #CCN2024! Presented our poster on place field reorganization & policy convergence. Met new friends and gained new insights. If you are working in comp neuro, deep RL, or neural reps, let's connect! #NeuroAI

Link to our short paper: Stay tuned for the full paper!.

Our work bridges neuroscience and machine learning, offering testable predictions for future experiments and insights into why certain neural phenomena might be beneficial for learning and adaptation. Check out our poster B48 this Thursday, August 08 2024!.

We also found that drifting fields facilitate faster adaptation when target locations change. This implies that representational drift might aid in exploring state spaces when reward structures shift.