Poster Presentations
Poster presentations will occur in Canopy B on Tuesday, September 2nd from 4:30PM – 6:30PM.
| Presenter | Poster Title |
| Ada Fang | AutoScientists: Self-Organizing Agent Teams for Long-Running Scientific Experimentation |
| Binxu Wang | The Two Clocks and the Innovation Window: When and How Generative Models Learn Rules |
| Huihong Li | Firing Rate Homeostasis Promotes Computation and Learning Through Irregular Spiking |
| Panayiotis Ketonis | BrainDyn: A Sheaf Neural ODE for Generative Brain Dynamics |
| Hadi Daneshmand | How to Talk to Your Language Model? |
| Leo Kozachkov | Is All Learning (Natural) Gradient Descent? |
| Sahil Loomba | Are learning rules conserved across species? Reading synaptic weights and Hebbian models from comparative connectomes |
| Clarissa Lauditi | Spectral Dynamics in Deep Networks: Feature Learning, Outlier Escape, and Learning Rate Transfer |
| Indranil Halder | Pre-train Poisoning Scaling Laws for Large Language Models: Theory and Experiment |
| Francisco Acosta | Mechanisms of Timescale Learning in Task-Trained Recurrent Neural Networks |
| Billy Qian | Carving the solution space of recurrent neural networks with biologically-informed desiderata |
| Vicente Conde Mendes | A solvable high-dimensional model where nonlinear autoencoders learn structure invisible to PCA while test loss misaligns with generalization |
| Sunreeta Bhattacharya | Probabilistic inference and structure learning |
| Pierfrancesco Beneventano | Edge of Stochastic Stability: Revising Edge of Stability for SGD |
| Wenping Cui | Dynamic selection of behavioral motifs during motor learning in freely moving rats |
| Ling-Qi Zhang | Effective Dimensionality as Bias-Task Alignment in Animal Learning |
| Jacqueline He | Mid-Training Knowledge Distillation Trades Factual Recall for Reasoning |
| Andrei Mircea | Shared Gradients Capture the Development and Structure of Generalizing Mechanisms in LLMs |
| Amirabbas Kazeminia | Realized Function Geometry: An Anchor-Based Perspective on the Geometry and Dynamics of Neural Network Learning |
| Linran Wei | Learning to Aggregate or Avoid: Ecological Constraints Shape Collective Behavior in MARL-trained Zebrafish-like Agents |
| Behrooz Tahmasebi | Dimension-Free Scaling Laws for Invariant Score Matching |
| Benjamin Midler | Coupled and reciprocal evolutionary and learning dynamics |
| Mili Patel | Escape response reveals computations of network robustness |
| Sahil Moza | How a whole brain reconfigures with learning: context-gated reorganization of brain-wide dynamics during olfactory learning in C. elegans |
| Ann Huang | Functionally Equivalent, Representationally Distinct: Traversing the Solution Space Reveals Shared Optimization Bias |
| Itay Lavie | Phase Transitions in Attention: A Bayesian Theory of Copy Head Emergence |
| Phanisree Akshinthala | Judge Bias Across Training: When Does Evaluation Bias Emerge in Language Models? |
| Noor Sajid | Optimal neural recoverability after damage |
| Ella Tubbs | Shared and Multiplexed Belief Codes Support Near-Bayesian Inference in Recurrent Networks |
| Camille Rullán | Representational advantages of distributional reinforcement learning |
| Dries Rooryck | Feeding BabyLMs Macaroni: Investigating Code-Switched Curriculum Learning in Data-Constrained Multilingual Modelling |
| Eugene L.Q. Lee | Temporal stimuli traces represent adaptive uncertainty allocation allowing for population-level bet-hedging |
| Hadas Raviv | The First 1000 Days (1kD): Modeling Language Acquisition using Continuous Child-Centered Experience |
| Ann Huang | Training is Less Consistent in MoEs than Dense Models |
| Sonja Johnson-Yu | Active Electrosensing and Communication in MARL-trained Weakly Electric Fish Collectives |
| Luca Pesce | A spiked perspective on feature learning with gradient-based methods |
| Abhiram Iyer | The Marauder’s Map: Bézier Manifolds Reveal Hidden Surfaces for Model Merging and Ensembling |
| Shubham Choudhary | The Cost of Looking: Active Sensing Policies for Sparse Signals |
| Dongrui Deng | Sparse Selectivity Enables Sample Efficient Decoding |
| Anastasiia Halchenko | Learning Rate, Timing Plasticity, and the Emergence of Functional Connectivity in Adaptive Oscillator Networks |
| Yujun Li | Foliations of neural population manifolds |
| Saptarshi Sengupta | TS-Edit: A Model Editing Framework for Adversarial Defense in Time-Series Foundation Models |
| Yung-Ying Chen | Criticality and Stereotypy Characterize Optimal Dynamical Regimes for Learning |
| Isabelle Lee | Kinematics of Grokking: How Models Spin Up Features |
| Himani Sinhmar | Simultaneous Thinking and Acting: Resource-Aware Embodied AI for Highly Interactive Worlds |
| Ari Liu | Inductive biases from initialization and optimization determine when recurrent networks reuse dynamics instead of rewiring them |
| Dongsung Huh | Beyond Smooth Interpolation: The Missing Generalization Principle in Deep Learning |
| Sumedh Hindupur | Comparison as a Computational Primitive: Modularity of Implementation Across Systems |
| Hope Kean | The Logical Subspace of LLMs |
| Shaunak Bhandarkar | Investigating the Early Signatures of Feature Learning in Deep Linear Networks |
| Tushar Chauhan | Imperfect by Design: Insights for bio-inspired Autonomous Intelligence |