PEHT integrates LoRA-enhanced Transformer with urban mobility and congestion data, achieving state-of-the-art network traffic prediction with over 90% parameter reduction.

Abdolazim Rezaei, Mehdi Sookhak, Mahboobeh Haghparast

RiVER leverages score-based optimization with instance ranking and winner-heavy rewards, improving LLMs on algorithmic tasks by 8.9%-9.4%.

Yingyu Lin, Qiyue Gao, Nikki Lijing Kuang et al.

Error-conditioned neural solver (ENS) uses PDE residuals as direct inputs, achieving 10× accuracy improvement and avoiding costly optimizers.

Haina Jiang, Liam Wang, Peng-Chen Chen et al.

Introduces two activation regularizers for Top-k sparse autoencoders, enhancing interpretability and robustness without sacrificing reconstruction quality.

Nathanaël Jacquier, Maria Vakalopoulou, Mahdi S. Hosseini

This paper establishes the algorithmic equivalence between Blackwell approachability and Gradient Equilibrium (GEQ), unifying online decision frameworks.

Brian W. Lee, Nika Haghtalab, Michael I. Jordan et al.

Introduces decision-alignment framework for uncertainty quantification (UQ), critiques traditional metrics, and proposes prior-weighted utility (PWU) metrics, validated through experiments showing superior decision utility alignment.

Annika Schneider, Tommy Rochussen, Joshua Stiller et al.

Introduces Red Queen Gödel Machine (RQGM), a co-evolving framework with learned evaluators, achieving 1.35-1.86× efficiency gains in code, science, and proof tasks under non-stationary utilities.

Alex Iacob, Andrej Jovanović, William F. Shen et al.

This study reveals the asymmetric dynamics of rule survival in pretraining models, with support frequency predicting rule fate and demonstrating irreversibility via editing interventions.

Juliana Li, Diya Sreedhar

Grad Detect leverages layer-wise gradients for single-pass hallucination detection, outperforming confidence baselines with 94-99% accuracy.

Anand Kamat, Daniel Blake, Brent M. Werness

Proposes Success Visitation Matching (SVM) rewards, transforming sparse rewards into dense signals, doubling RL finetuning speed in robotic tasks.

Raymond Tsao, Andrew Wagenmaker, Sergey Levine

Proposes graph-bound execution-state capsules for low-latency, small-batch on-device AI, enabling byte-exact snapshot and restore with sub-millisecond GPU performance.

Liang Su

Proposes Piecewise Polynomial Interpolation-based Gradient Descent (PPI-GD) achieving oracle complexity of O((p/ε)^{d/(2ℓ)}) for data dimension d=O(log^{0.49}(n)), outperforming classical GD/SGD.

Dongmin Lee, William Lu, Anuran Makur

Proposes STARE, a surprisal-guided advantage reweighting method, stabilizing policy entropy and improving accuracy by 4%-8% on models from 1.5B to 32B.

Haipeng Luo, Qingfeng Sun, Songli Wu et al.

Proposes a zero-shot active feature acquisition framework using LLM-derived discriminative statistics and MaxEnt closure, significantly improving IBD diagnosis accuracy.

Binyamin Perets, Natalie Mendelson, Shiran Vainberg et al.

Proposes LoopWM, a parameter-shared transformer with iterative latent refinement, achieving 100× parameter efficiency and stable long-horizon environment prediction.

Hongyuan Adam Lu, Z. L. Victor Wei, Qun Zhang et al.

Introduces Kolmogorov PDE-based diffusion policies with dimension-independent convergence, improving long-horizon control in robotics and manufacturing.

Lekan Molu

This paper analyzes the sparsity and geometric structure of on-policy distillation (OPD), revealing small, coordinate-sparse updates that are spectrally concentrated and deviate from source principal directions.

Guo Yu, Wenlin Liu, Yulan Hu et al.

Proposes Bebop with TV loss and rejection sampling to stabilize MTP acceptance rate, achieving up to 95% and 1.8× RL training acceleration.

Yucheng Li, Huiqiang Jiang, Yang Xu et al.

This study compares xLSTM, Mamba-2, and Gated DeltaNet architectures, demonstrating xLSTM's superior performance in complex sequence tasks due to its robust state tracking and memory accumulation.

Anamaria-Roberta Hartl, Levente Zólyomi, David Stap et al.

Proposes Latent World Recovery (LWR), a robust multimodal learning framework that aligns modality-specific embeddings in a shared latent space, handling missing modalities without imputation.

Hui Wang, Tianyu Ren, Joseph Butler et al.