SMASH integrates onboard egocentric vision and scalable whole-body skill learning, enabling humanoid robots to perform continuous outdoor ping-pong without external sensors.

Junli Ren, Yinghui Li, Kai Zhang et al.

Multi-LLM revision pipelines decompose second-pass gains into re-solving, scaffold, and content effects, enhancing MCQ and code generation tasks.

Jingjie Ning, Xueqi Li, Chengyu Yu

CGCMA uses conditionally-gated cross-modal attention to achieve event-conditioned asynchronous fusion, improving Sharpe ratio to +0.449 on CryptoMI dataset.

Yunxiang Guo

Proposes Bench2Drive-VL, integrating DriveCommenter for real-time closed-loop autonomous driving evaluation with multimodal reasoning.

Xiaosong Jia, Yuqian Shao, Zhenjie Yang et al.

UniMixer introduces a unified recommendation scaling architecture with parameterized TokenMixer, combining attention, TokenMixer, and FM advantages for improved efficiency.

Mingming Ha, Guanchen Wang, Linxun Chen et al.

HabitatAgent achieves 95% accuracy in housing consultation using a multi-agent system.

Hongyang Yang, Yanxin Zhang, Yang She et al.

Introduced TPC-CMA method, reducing modality gap by 82.3% with only 4.84% accuracy drop.

Hongyuan Liu, Qinli Yang, Wen Li et al.

DreamControl-v2 trains guided diffusion models directly in robot space, integrating diverse datasets to enhance scalability and automation for humanoid skills.

Sudarshan Harithas, Sangkyung Kwak, Pushkal Katara et al.

Using perturbative analysis, the paper proves that the Grothendieck constant K_G exceeds Davie-Reeds' lower bound by at least 10^−12, marking a first quantitative improvement since the 1980s.

Chris Jones, Giulio Malavolta

C-TRAIL integrates LLMs with trust mechanisms via a closed-loop Recall-Plan-Update framework for autonomous driving trajectory planning.

Zhihong Cui, Haoran Tang, Tianyi Li et al.

Proposes a granular ball-based square superpixel method enabling end-to-end deep learning integration, improving efficiency and structure.

Shuyin Xia, Meng Yang, Dawei Dai et al.

PRISM leverages multi-view videos and a 3D knowledge ontology to enhance embodied VLM's spatial, physical, and action understanding in retail environments.

Amirreza Rouhi, Parikshit Sakurikar, Satya Sai Reddy et al.

Multi-Layer Memory Framework enhances LLMs' long-term context retention, achieving 46.85% success rate and 56.90% six-period retention.

Sunil Tiwari, Payal Fofadiya

GenSplat enhances view generalization in robotic policy learning using 3D Gaussian Splatting.

Sen Wang, Huaiyi Dong, Jingyi Tian et al.

LatentPilot internalizes future visual dynamics via privileged supervision, achieving SOTA in VLN benchmarks with 66.3% success rate.

Haihong Hao, Lei Chen, Mingfei Han et al.

SparseDriveV2 leverages dense static vocabularies with trajectory decomposition, achieving 92.0 PDMS, challenging the necessity of dynamic proposals.

Wenchao Sun, Xuewu Lin, Keyu Chen et al.

GISTBench uses IG and IS to test whether LLM user profiles are supported by behavior; survey alignment reaches ρ=0.67.

Iordanis Fostiropoulos, Muhammad Rafay Azhar, Abdalaziz Sawwan et al.

MEDiC unifies CLIP distillation and pixel reconstruction, reaching 73.9% kNN and 85.1% fine-tuning accuracy on ImageNet-1K.

Konstantinos Georgiou, Maofeng Tang, Hairong Qi

StreamingVLA achieves 2.4× speedup and 6.5× halting reduction via Action Flow Matching and Adaptive Early Observation.

Yiran Shi, Dongqi Guo, Tianchen Zhao et al.

Proposes AgentTrace, a visualization framework that enhances user understanding and traceability of high-authority AI agents' actions, improving risk awareness and post-hoc auditability.

Zifan Peng, Mingchen Li