DUET accelerates Mamba-Transformer models by assigning prefill and decode phases to specialized packages, achieving 4x speed improvement.

Alish Kanani, Sangwan Lee, Han Lyu et al.

FreeTalk is an emotion-driven 3D talking head animation framework supporting arbitrary topology.

Federico Nocentini, Thomas Besnier, Claudio Ferrari et al.

RieMind constructs explicit 3D scene graphs (3DSG) for static indoor scene reasoning, outperforming previous methods with up to 16% improvement under ideal perceptual conditions.

Fernando Ropero, Erkin Turkoz, Daniel Matos et al.

Introduces Attention Residuals using softmax attention to replace fixed accumulation, enhancing model performance.

Kimi Team, Guangyu Chen, Yu Zhang et al.

Proposes STALL, a training-free video detector using spatial-temporal likelihoods, outperforming baselines with 0.85 average AUC.

Omer Ben Hayun, Roy Betser, Meir Yossef Levi et al.

Conformal gate routing controls surrogate violations distribution-free; on 35 OpenML tasks, τ=2 achieved 66% coverage and 12% violations.

Iqtedar Uddin, Mazin Khider, André Bauer

CausalEvolve adds causal scratchpads to program evolution, improving four open-ended tasks and reaching 38.89% on AIME.

Yongqiang Chen, Chenxi Liu, Zhenhao Chen et al.

VLA-Thinker boosts Vision-Language-Action models via dynamic image reasoning, achieving 97.5% success on LIBERO.

Chaoyang Wang, Wenrui Bao, Sicheng Gao et al.

AI learns scientific taste via Reinforcement Learning from Community Feedback, enhancing discovery efficiency.

Jingqi Tong, Mingzhe Li, Hangcheng Li et al.

DataEvolve employs evolutionary algorithms and sample evaluation to autonomously optimize pretraining data strategies, boosting model performance.

Tiantian Mi, Dongming Shan, Zhen Huang et al.

This study critiques questionnaire-based AI safety assessments, highlighting the divergence between model responses and real-world agent behaviors, emphasizing input and interaction differences.

Max Hellrigel-Holderbaum, Edward James Young

PDPS method exposes long-tail safety failures in LLMs, reducing computational cost by 33%.

Suvadeep Hajra, Palash Nandi, Tanmoy Chakraborty

OmniClone, Transformer-based humanoid teleoperation system, reduces MPJPE by over 66%, enabling versatile, high-fidelity control with minimal data on consumer GPU.

Yixuan Li, Le Ma, Yutang Lin et al.

Develops an exactly solvable n-gram-based recursive framework analyzing drift and selection effects on public text ecosystems.

Søren Riis

Proposes Tucker Adaptation (TuKA) for high-order tensor representation of multi-hierarchical navigation knowledge, enabling lifelong multi-scene VLN.

Xudong Wang, Gan Li, Zhiyu Liu et al.

AIM-SciQA employs large language models to automatically generate 13,672 multi-hop scientific QA pairs across documents, enhancing cross-document reasoning.

Seungmin Lee, Dongha Kim, Yuni Jeon et al.

Constructivist grounded theory analyzes how designers position LLMs as tools or teammates, affecting responsibility and decision authority.

Varad Vishwarupe, Ivan Flechais, Nigel Shadbolt et al.

PaIR-Drive enhances autonomous driving via a parallel framework, achieving 91.2 PDMS.

Zhexi Lian, Haoran Wang, Xuerun Yan et al.

Proposes GraspADMM, using ADMM to optimize multi-objective dexterous grasping, achieving 15% success rate improvement.

Liangwang Ruan, Jiayi Chen, He Wang et al.

RetimeGS introduces explicit temporal modeling in 4D Gaussian Splatting, significantly reducing ghosting artifacts and enabling smooth continuous-time scene reconstruction.

Xuezhen Wang, Li Ma, Yulin Shen et al.