Posters

IRE framework expedites the discovery of flat minima in deep learning, enhancing generalization and convergence. By decoupling the dynamics of flat and sharp directions, IRE boosts sharpness reduction…

Boost equivariant model training by strategically relaxing constraints during training, enhancing optimization and generalization!

Deep RL agents often suffer from instability due to the ‘chain effect’ of value and policy churn; this paper introduces CHAIN, a novel method to reduce this churn, thereby improving DRL performance an…

Constrained Parameter Regularization (CPR) outperforms traditional weight decay by dynamically adapting regularization strengths for individual parameters, leading to better deep learning model perfor…

Boosting machine learning model interpretability, this paper introduces cluster-based and tree-based Sparse Explanation Values (SEV) for generating more meaningful and credible explanations by optimiz…

Context-aware preference modeling improves language model alignment by resolving ambiguity through a two-step process: context selection followed by context-specific preference evaluation. The approa…

AI systems are made safer by ‘circuit breakers’ that directly control harmful internal representations, significantly improving alignment and robustness against adversarial attacks with minimal impact…

Boosting adversarial robustness fairness in deep neural networks, Anti-Bias Soft Label Distillation (ABSLD) adaptively adjusts soft label smoothness to reduce error gap between classes.

Over-parameterized gradient descent dynamically adapts to signal structure, improving sequence model generalization and outperforming fixed-kernel methods.

This paper significantly improves our understanding of multiclass PAC learning by reducing the sample complexity gap and proposing two novel approaches to fully resolve the optimal sample complexity.

Training high-quality diffusion models efficiently is now possible, thanks to novel sample complexity bounds improving exponentially on previous work.

Linear ensemble sampling achieves a state-of-the-art regret bound of Õ(d³/²√T) with a logarithmic ensemble size, closing the theory-practice gap in linear bandit algorithms.

A novel algorithm, D-FTBL, achieves improved regret bounds for bandit convex optimization with delayed feedback, tightly matching existing lower bounds in worst-case scenarios.

Researchers enhanced diffusion samplers by developing a novel exploration strategy and a unified library, improving sample quality and addressing reproducibility challenges.

New modeling of bid space in multi-unit uniform price auctions achieves regret of Õ(K4/3T2/3) under bandit feedback, improving over prior work and closing the gap with discriminatory pricing.

A novel fully dynamic algorithm achieves a (4+ε)-approximate solution for the k-center clustering problem with outliers in general metric spaces, boasting an efficient update time.

Researchers developed novel methods to improve the generation of adversarial examples against safety-aligned LLMs, achieving significantly higher attack success rates compared to existing techniques.

Improved few-shot jailbreaking techniques efficiently circumvent aligned language models and their defenses, achieving high success rates even against advanced protection methods.

This paper significantly improves Bayes regret bounds for hierarchical Bayesian bandit algorithms, achieving logarithmic regret in finite action settings and enhanced bounds in multi-task linear and c…

A new algorithm, AOGS, achieves significantly lower regret in two-sided matching markets by cleverly integrating exploration and exploitation, thus removing the dependence on the number of arms (K) in…