Proposes MindGPT-4ov, a multi-stage post-training framework integrating data synthesis, curriculum fine-tuning, and reinforcement learning, achieving SOTA performance.

Wei Chen, Chaoqun Du, Feng Gu et al.

ReVSeg uses reinforcement learning to optimize explicit reasoning chains, achieving state-of-the-art video segmentation with interpretable decision trajectories.

Yifan Li, Yingda Yin, Lingting Zhu et al.

RULER-Bench evaluates video-generation rule reasoning; SOTA reaches only 48.87% rule coherence.

Xuming He, Zehao Fan, Hengjia Li et al.

DeepSeek-V3.2 employs sparse attention and large-scale reinforcement learning to match GPT-5's reasoning and agent capabilities, significantly narrowing the open-source gap.

DeepSeek-AI, Aixin Liu, Aoxue Mei et al.

Proposed Temporal Dynamics Enhancer boosts SNNs' temporal modeling, achieving 57.7% and 47.6% mAP50-95.

Fan Luo, Zeyu Gao, Xinhao Luo et al.

Chain-of-Ground enhances GUI grounding through iterative reasoning, achieving 68.4% accuracy on ScreenSpot-Pro.

Aiden Yiliu Li, Bizhi Yu, Daoan Lei et al.

Models Transformer attention as an interacting particle system; analyzes mean-field limits revealing multi-cluster formation and phase transitions.

Philippe Rigollet

Introducing reasoning overconfidence in LLMs; Long-CoT reduces it via iterative reflection, based on the cognitive rigidity hypothesis.

Jiannan Guan, Qiguang Chen, Libo Qin et al.

SPARK combines VLMs, DiT, and differentiable kinematics for single-image sim-ready reconstruction, reaching 0.3959 Chamfer Distance.

Yumeng He, Ying Jiang, Jiayin Lu et al.

FRAMER utilizes frequency-aligned self-distillation with diffusion priors to enhance image super-resolution, significantly improving PSNR and SSIM.

Seungho Choi, Jeahun Sung, Jihyong Oh

Structured Spectral Reasoning (SSR) decomposes multimodal signals into spectral bands, employs band masking, hyperspectral fusion, and contrastive regularization to boost recommendation accuracy and robustness.

Wei Yang, Rui Zhong, Yiqun Chen et al.

AI-Trader benchmarks autonomous multi-market trading, revealing models' limited profitability and risk control in real-time environments.

Tianyu Fan, Yuhao Yang, Yangqin Jiang et al.

Flow matching (FM) with paths (OT/VP) outperforms diffusion models in tabular data synthesis, achieving high utility with low function evaluations.

Bahrul Ilmi Nasution, Floor Eijkelboom, Mark Elliot et al.

Establishes non-asymptotic convergence bounds for three discrete diffusion models using Euler discretization, applicable to finite and infinite state spaces.

Giovanni Conforti, Alain Durmus, Le-Tuyet-Nhi Pham et al.

FR-TTS uses filling-based reward to improve NTP image generation, significantly boosting quality.

Hang Xu, Linjiang Huang, Feng Zhao

Introduces reflected Langevin diffusion algorithm, proving convergence in Wasserstein-2 distance.

Tarika Mane, Amine Boukardagha

WMAct combines reward rescaling and frequency annealing to enable LLMs to perform efficient world model reasoning via active multi-turn interaction.

Bao Shu, Yan Cai, Jianjian Sun et al.

Combining PPO and DQN in multi-agent RL enables P2P energy trading in dairy farms, reducing costs by 14.2% and peak demand by 55.5%.

Mian Ibad Ali Shah, Marcos Eduardo Cruz Victorio, Maeve Duffy et al.

SpaceMind employs camera-guided modality fusion, achieving SOTA in RGB-based 3D spatial reasoning benchmarks.

Ruosen Zhao, Zhikang Zhang, Jialei Xu et al.

DiskChunGS achieves large-scale 3D Gaussian SLAM via chunk-based memory management, completing all KITTI sequences successfully.

Casimir Feldmann, Maximum Wilder-Smith, Vaishakh Patil et al.