EmbeddingGemma is a lightweight 300M-parameter text embedding model surpassing larger models in multilingual and multi-task benchmarks, suitable for low-latency applications.

Henrique Schechter Vera, Sahil Dua, Biao Zhang et al.

Introduces LLM-Filter, leveraging pretrained large language models for state estimation, outperforming traditional and learning-based filters with strong generalization.

Shiqi Liu, Wenhan Cao, Chang Liu et al.

Contribution score computation in federated learning is fragile to architecture and attacks, requiring robust methods.

Balazs Pejo, Marcell Frank, Krisztian Varga et al.

GUIDE integrates global graph inference and diffusion policy, achieving 18.3% faster coverage and 34.9% fewer redundant moves.

Zijun Che, Yinghong Zhang, Shengyi Liang et al.

Proposes Online-Optimized RAG using online gradient updates to improve tool retrieval accuracy with minimal feedback.

Yu Pan, Xiaocheng Li, Hanzhao Wang

EgoBridge improves robot imitation learning success rate by 44% through domain adaptation.

Ryan Punamiya, Dhruv Patel, Patcharapong Aphiwetsa et al.

Terra: a hierarchical, terrain-aware 3D scene graph for large-scale outdoor mapping, combining LiDAR SLAM, CLIP-based semantic recognition, and multi-level clustering.

Chad R. Samuelson, Abigail Austin, Seth Knoop et al.

BiGraspFormer is an end-to-end transformer framework that directly generates coordinated bimanual grasps from point clouds, achieving high success and efficiency.

Kangmin Kim, Seunghyeok Back, Geonhyup Lee et al.

Proposes LoRD and B-LoRD models, low-rank doubly stochastic graph clustering with probabilistic interpretation, improving efficiency and structure control.

Wenlong Lyu, Yuheng Jia, Hui Liu et al.

Proposed a sim-and-real co-training framework using Unbalanced Optimal Transport, achieving up to 30% success rate improvement in real-world tasks.

Shuo Cheng, Liqian Ma, Zhenyang Chen et al.

APRIL accelerates RL rollout by over-provisioning and recycling incomplete responses, boosting throughput by 22.5%.

Yuzhen Zhou, Jiajun Li, Yusheng Su et al.

GD2P uses geometry-aware dexterous hands for nonprehensile pushing and pulling, enhancing manipulation efficiency.

Yunshuang Li, Yiyang Ling, Gaurav S. Sukhatme et al.

LAD-VF leverages formal verification feedback with AutoDiff for prompt optimization, achieving over 90% compliance without model fine-tuning.

Yunhao Yang, Junyuan Hong, Gabriel Jacob Perin et al.

UniPixel unifies pixel-level perception and reasoning, achieving top performance on 10 benchmarks with 3B/7B models.

Ye Liu, Zongyang Ma, Junfu Pu et al.

TempSamp-R1 enhances video temporal grounding via reinforcement fine-tuning, achieving significant accuracy improvements.

Yunheng Li, Jing Cheng, Shaoyong Jia et al.

Do Modern Video-LLMs Need to Listen? By integrating audio encoders, LLaVA-OneVision enhances cross-modal reasoning.

Geewook Kim, Minjoon Seo

Proposed MLLM-CD framework combines contrastive factor discovery and iterative counterfactual reasoning to uncover causal relations in multimodal unstructured data.

Jin Li, Shoujin Wang, Qi Zhang et al.

Qwen3-Omni uses Thinker-Talker MoE architecture to achieve state-of-the-art multimodal performance without degradation across text, image, audio, and video tasks.

Jin Xu, Zhifang Guo, Hangrui Hu et al.

SD-VLM enhances VLM's spatial understanding with MSMU dataset and depth positional encoding, outperforming GPT-4o and Intern-VL3-78B.

Pingyi Chen, Yujing Lou, Shen Cao et al.

DIWALI creates a dataset of 8,817 Indian cultural items across 17 facets to evaluate LLMs' cultural adaptation, combining automatic and human assessments.

Pramit Sahoo, Maharaj Brahma, Maunendra Sankar Desarkar