Posters

Exact Gauss-Newton optimization in deep reversible networks surprisingly reveals poor generalization, despite faster training, challenging existing deep learning optimization theories.

New framework uses rough path theory to enable gradient-based training of SSNNs driven by rough signals, allowing for noise in spike timing and network dynamics.

Low Precision Ensembling (LPE) boosts large model accuracy using training-free ensemble creation via stochastic rounding in low-precision number systems.

EvolveDirector trains competitive text-to-image models using publicly available data by cleverly leveraging large vision-language models to curate and refine training datasets, dramatically reducing d…

E-NSDE, a novel time-aware sequential recommendation model, integrates neural stochastic differential equations and evidential learning to improve recommendation accuracy by effectively handling varia…

Evidential Mixture Machines (EMM) enhances multi-label active learning by deciphering label correlations for improved accuracy and uncertainty quantification in large, sparse label spaces.

Chess AI Leela Zero surprisingly uses learned look-ahead, internally representing future optimal moves, significantly improving its strategic decision-making.

Researchers introduce object-aware backdoors in Vision-and-Language Navigation, enabling malicious behavior upon encountering specific objects, demonstrating the vulnerability of real-world AI agents.

Event-3DGS: First event-based 3D reconstruction using 3D Gaussian splatting, enabling high-quality, efficient, and robust 3D scene reconstruction in challenging real-world conditions.

Spexphormer achieves significant memory reduction in graph Transformers by leveraging a two-stage training process that leverages attention score consistency across network widths to effectively spars…

DEVIL: a novel text-to-video evaluation protocol focusing on video dynamics, resulting in more realistic video generation.

This paper provides the first complete error analysis for diffusion models, theoretically justifying optimal training and sampling strategies and design choices for enhanced generative capabilities.

Pretrained neural networks surprisingly capture fundamental aspects of human cognition, enabling generalization in image-based learning tasks, as demonstrated by aligning neural network representation…

Shapley-based Cooperation Enhancing Multi-view Clustering (SCE-MVC) improves deep multi-view clustering by using game theory to fairly evaluate and enhance individual view contributions.

Deep random features enable efficient Euclidean distance compression, offering improved bit storage compared to linear methods for specific parameter ranges, thus significantly advancing high-dimensio…

ET-Flow, a novel equivariant flow-matching model, generates highly accurate and physically realistic molecular conformers significantly faster than existing methods.

New method estimates hallucination rates in generative AI’s in-context learning, improving model reliability.

This study develops a novel two-stage framework for accurately predicting conditional average treatment effects using both observational data and weak instrumental variables, overcoming limitations of…

New consistent estimators precisely track generalization error during robust regression’s iterative model training, enabling optimal stopping iteration for minimized error.

HyperDM accurately estimates both epistemic and aleatoric uncertainty using a single model, overcoming the computational limitations of existing ensemble methods.