Q-DAPS estimates question difficulty by computing the entropy of plausibility scores, excelling on four QA datasets.

Jamshid Mozafari, Bhawna Piryani, Adam Jatowt

SafeManip uses LTLf to evaluate temporal safety in robotic manipulation, revealing task success does not equal safe execution.

Chengyue Huang, Khang Vo Huynh, Sebastian Elbaum et al.

MedHopQA evaluates biomedical QA via multi-hop reasoning with 1,000 expert-curated question-answer pairs.

Rezarta Islamaj, Robert Leaman, Joey Chan et al.

Proposes Linearized Graph Sequence Models (LGSM), decoupling information propagation from nonlinear processing to enhance long-range dependency learning.

Joël Mathys, Basil Rohner, Saku Peltonen et al.

Real-time whole-body teleoperation using Virdyn IMU motion capture, validated on Unitree G1 robot with Sim2Sim and Sim2Real.

Hamza Ahmed Durrani, Suleman Khan

Proposes Executable Agentic Memory (EAM) using knowledge graphs to enhance long-horizon GUI planning, achieving 19.6% success improvement and 6× cost reduction.

Zerui Qin, Sheng Yue, Xingyuan Hua et al.

Introduced a sequential supersample framework using sequential CMI to control generalization gaps.

Futoshi Futami, Masahiro Fujisawa

Dynamic Cognitive Reconciliation Decoding (DCRD) predicts and mitigates context-memory conflicts, achieving state-of-the-art performance across six QA datasets.

Yigeng Zhou, Wu Li, Yifan Lu et al.

MoLA uses pretrained multimodal inverse dynamics to convert imagined future videos into executable robot actions, boosting success and generalization.

Yajie Li, Bozhou Zhang, Chun Gu et al.

MoCam employs structured denoising dynamics in diffusion models to unify geometry and appearance for robust novel view synthesis.

Haofeng Liu, Yang Zhou, Ziheng Wang et al.

Introducing perceptual entropy to prevent diversity collapse in flow-based RLHF, achieving a score of 0.734 and diversity of 0.989, surpassing baselines.

Xiaofeng Tan, Jun Liu, Bin-Bin Gao et al.

BARISTA employs dense scene graphs in egocentric videos for multi-task understanding, covering 185 real coffee-making videos with over 3.6 million annotations.

Patrick Knab, Orgest Xhelili, Inis Buzi et al.

SkillSafetyBench evaluates agent safety against 155 adversarial cases across 47 tasks, focusing on skill-mediated non-user attack surfaces.

Chang Jin, An Wang, Zeming Wei et al.

BadSKP is a multi-stage optimization backdoor attack targeting graph-to-prompt interfaces in KG-enhanced LLMs, achieving over 90% success rate.

Xiaoting Lyu, Yufei Han, Hangwei Qian et al.

Chronicles-OCR evaluates VLLMs' cross-temporal visual perception of Chinese character evolution using a Stage-Adaptive Annotation Paradigm with 2,800 images.

Gengluo Li, Shangpin Peng, Xingyu Wan et al.

Proposed a method using multi-view latent priors for action manifold learning, outperforming on LIBERO dataset.

Junjin Xiao, Dongyang Li, Yandan Yang et al.

Proposes energy-to-token model using Leontief function; emphasizes energy consumption as key constraint; recommends reporting Joules/token.

Xiang Liu, Shimiao Yuan, Zhenheng Tang et al.

Proposes COMPOSE framework: training preserves object geometry, inference dynamically composes slots, achieving state-of-the-art generalization in continual few-shot learning.

Phu-Quy Nguyen-Lam, Phu-Hoa Pham, Dao Sy Duy Minh et al.

D-PACE uses dynamic position-aware weights to improve parallel speculative decoding, boosting speed and output length.

Tianyu Wu, Yu Yao, Zhenting Qi et al.

PointForward uses point-aligned sparse 3D queries for fast, multi-view consistent driving scene reconstruction, outperforming pixel-aligned methods.

Cheng Chi, Xianqi Wang, Hongcheng Luo et al.