Posters

Deep neural network training reveals asymmetric loss valleys, impacting model fusion and federated learning; sign consistency between noise and convergence is key.

Deep state space models (SSMs) gain adversarial robustness through an adaptive scaling mechanism, improving performance without overfitting issues.

ExRAP: A novel framework boosts embodied AI’s continual instruction following by cleverly combining environment exploration with LLM-based planning, leading to significantly improved task success and …

LEADS: a novel algorithm learning diverse skills through successor state representations for robust exploration in reward-free environments.

Boost RL performance by solving a series of simplified MDPs before tackling the complex real-world one!

RECURVE: A novel boundary detection method leveraging representation trajectory curvature, surpassing state-of-the-art techniques by accommodating both gradual and abrupt changes in time series.

LLM quantization, while improving efficiency, creates a security risk: attackers can craft seemingly benign models that exhibit malicious behavior only when quantized.

PCRIL, a novel prompt contrastive recovery approach, significantly boosts cross-modal hashing accuracy, especially when dealing with incomplete labels by progressively identifying promising positive c…

D2DMoE boosts Transformer efficiency by up to 60% via smart activation sparsity and dynamic expert selection, outperforming existing methods.

Eigen-SAM significantly boosts generalization in deep learning by directly addressing SAM’s limitations through explicit top Hessian eigenvalue regularization.

New method, STRIPE-X, powerfully detects ’encoding’ in AI explanations—a sneaky phenomenon where explanations predict outcomes better than their constituent parts alone would suggest.

This paper introduces a model-agnostic algorithm that uses natural language predicates to make statistical model parameters directly interpretable, significantly improving explainability.

This research introduces a novel, automated protocol translation framework for self-driving labs, tackling the challenge of converting human-readable experimental protocols into machine-interpretable …

Expected Probabilistic Hierarchies (EPH) offers a novel, scalable approach to hierarchical clustering by optimizing expected scores under a probabilistic model, outperforming existing methods on vario…

This paper introduces Expectation Alignment (EAL), a novel framework and interactive algorithm to address reward misspecification in AI, aligning AI behavior with user expectations.

SNELL: Sparse tuning with kerNElized LoRA achieves state-of-the-art parameter-efficient fine-tuning performance with drastically reduced memory usage.

Exogenous Matching learns optimal proposals for efficient counterfactual estimation by transforming variance minimization into conditional distribution learning, outperforming existing methods.

Exo2Ego-V generates realistic egocentric videos from sparse exocentric views, significantly outperforming state-of-the-art methods on a challenging benchmark.

Boost RL efficiency in continuous action spaces by masking irrelevant actions using three novel continuous action masking methods!

This paper provides the first exact minimax-optimal mechanisms for locally differentially private sampling, applicable across all f-divergences.