Introduces DeepONet within a GAN framework for high-resolution, long-range precipitation nowcasting, improving stability and physical consistency.

Mohammad Kian Golkar, Luciano Alves de Oliveira, Mohammad Khanjani

StreamScout enhances timeline views progressively for efficient streaming video understanding, reducing token usage by 59%.

Ce Zhang, Jing Bi, Jinxi He et al.

DUPIN combines unsupervised pretraining and few-shot learning for effective APT attack forensics.

Chanwoo Bae, Hailun Ding, Shiqing Ma et al.

Introduces invalidation contracts to enhance cache validity and compliance with version stamps and cacheability hints.

Michael Wu, Arquimedes Canedo

ReDeck enhances slide generation with step-level render feedback, improving spatial accuracy and design quality.

Muzhao Tian, Zezi Zeng, Yifan Yang et al.

Proposes Context-Aware Interleaved Batching for WhisperX, reducing WER to 8.2% and boosting speed by 12× by maintaining continuous context across segments.

Carlos Bain, Max Bain

BRF-GS employs 3D Gaussian splatting with hybrid BRDF kernels and spectral band selection for accurate multi-angle hyperspectral reflectance modeling.

Yiling Yao, Wenjuan Zhang, Bowen Wang et al.

DreamX-Creator employs Gated Cross-Modal Attention in a 7B model to achieve native 2K synchronized audio-video generation.

Jiashu Zhu, Yanhao Zheng, Ruitian Tian et al.

ECCBench evaluates vision-language models' memory beyond accuracy.

Shmuel Berman, Jia Deng

Proposes adaptive data cracking for LLMs, reducing reasoning costs by over 50% through dynamic structure extraction.

Milad Rezaei Hajidehi, Qitong Wang, Stratos Idreos

This study reveals that contrastive preference optimization (e.g., DPO) unintentionally transfers sycophantic behaviors from teacher models, with the log-ratio of teacher sycophancy predicting student behavior.

Camila Blank, Zhuofan Ying, Christopher Potts et al.

AutoSciRub automatically induces task-specific, verifiable rubrics, significantly improving autonomous research agent performance by 2.08 points on ResearchClawBench and 16.8 points on AstaBench.

Xuehai Wang, Haowei Qin, Tongxin Liu et al.

Knowledge-aligned SFT reduces hallucinations, enhancing factuality.

Arthur Becker, Jakob Kemmler, David Thulke et al.

Proposes an audio-driven adversarial defense method to preserve visual fidelity in 3D talking face generation.

Rui-Qing Sun, Chen-Hao Cui, Hui-Yang Zhao et al.

LightNav-0 leverages pretrained VLM's spatial reasoning via unified pointing and RVQ trajectory decoding, enabling multi-task embodied navigation without task-specific heads.

Shaoan Wang, Aocheng Luo, Fei Huang et al.

Zeva enables in-context causal learning, allowing robots to self-evolve without parameter updates, improving success rates over repeated attempts.

Fu Chen, Xin Ding, Bingjia Huang et al.

Proposes a state-based decentralized contingency MPC ensuring safety and convergence with limited sensing in multi-agent systems.

Max Studt, Georg Schildbach

AVENUE evaluates audio-video editing models' modality-selectivity using 1,291 clips and 7,957 instructions.

Hayeon Kim, Yoojin Jang, Jaejun Yoo

SocialReasonBench evaluates social reasoning via branching narrative videos, revealing LMMs' shortcomings in counterfactual reasoning.

Zheyu Huang, Zijing Shi, Haozhe Luo et al.

Liquid Gated Attention (LGA) enables continuous-time modeling with input-driven gating, achieving linear temporal complexity.

Yiheng Jiang, Yuanbo Xu, Yongjian Yang