Flow Expander (FE) employs verifier-constrained entropy maximization via mirror descent to expand pre-trained flow models’ valid design space.

Riccardo De Santi, Kimon Protopapas, Ya-Ping Hsieh et al.

The study reveals limitations in VLMs' spatial reasoning for non-textual elements, with text-recognition pathways outperforming visual pathways.

Yuval Levental

GLM-5 integrates DSA and asynchronous RL, achieving SOTA in long-text understanding and code generation with 20% improvement.

GLM-5-Team, :, Aohan Zeng et al.

Flowers employs multi-head pointwise warps for PDE solution operators, outperforming similar-sized models with only 17M parameters, achieving state-of-the-art results.

Till Muser, Alexandra Spitzer, Matti Lassas et al.

Req2Road employs LLMs and VLMs with retrieval-augmented generation to automate SDV test artifact creation and vehicle execution, boosting testing efficiency.

Denesa Zyberaj, Lukasz Mazur, Pascal Hirmer et al.

Introduces event-level augmentation protocol for IoT anomaly detection, evaluating 14 models under simulated real-world perturbations, revealing regime-dependent robustness.

Dmitry Zhevnenko, Ilya Makarov, Aleksandr Kovalenko et al.

Study reveals visual preferences of vision-language models using visual prompt optimization; experiments show significant shifts in choice probabilities.

Manuel Cherep, Pranav M R, Pattie Maes et al.

This paper introduces size transferability of Graph Transformers using convolutional positional encodings, enabling models trained on small graphs to generalize to larger ones.

Javier Porras-Valenzuela, Zhiyang Wang, Xiaotao Shang et al.

BindCLIP combines contrastive learning with diffusion-based pose generation, enhancing interaction sensitivity and out-of-distribution generalization in virtual screening.

Anjie Qiao, Zhen Wang, Yaliang Li et al.

Colosseum framework detects collusion in multi-agent LLM systems via action regret and communication scores.

Mason Nakamura, Abhinav Kumar, Saswat Das et al.

EditCtrl achieves real-time video editing with local/global control modules, improving efficiency by 10x.

Yehonathan Litman, Shikun Liu, Dario Seyb et al.

AnchorWeave significantly improves long-term video generation consistency using retrieved local spatial memories.

Zun Wang, Han Lin, Jaehong Yoon et al.

A2G-DiffRec enhances item-side fairness in diffusion recommenders using adaptive autoguidance, effective across three datasets.

Zihan Li, Gustavo Escobedo, Marta Moscati et al.

DriveFine integrates masked diffusion with self-correction, achieving robust and precise autonomous driving decisions, with PDMS reaching 91.8 and EPDMS 89.7 on benchmarks.

Chenxu Dang, Sining Ang, Yongkang Li et al.

Proposes Query-as-Anchor framework with hierarchical encoders and joint contrastive-autoregressive training for scenario-adaptive user representations, achieving state-of-the-art results.

Jiahao Yuan, Yike Xu, Jinyong Wen et al.

The Interspeech 2026 Audio Reasoning Challenge uses MMAR-Rubrics to evaluate audio reasoning models, with agent systems leading in quality.

Ziyang Ma, Ruiyang Xu, Yinghao Ma et al.

GeoEyes enhances UHR remote sensing image understanding with AdaZoom-GRPO, achieving 54.23% accuracy on XLRS-Bench.

Fengxiang Wang, Mingshuo Chen, Yueying Li et al.

Proposed a low-cost sim-to-real transfer method by learning force direction, significantly improving contact-rich task success rates.

Yifei Yang, Anzhe Chen, Zhenjie Zhu et al.

Proposes knowledge profiling revealing near-saturation encoding but persistent recall bottleneck; thinking improves recall significantly.

Nitay Calderon, Eyal Ben-David, Zorik Gekhman et al.

DeepXiv-SDK offers a structured, progressive, and cost-aware scientific literature interface, enabling efficient retrieval and evidence extraction from the ArXiv corpus.

Hongjin Qian, Ziyi Xia, Ze Liu et al.