LoTIS employs image-space localization of reference trajectories using Transformer architecture, achieving 94-98% success in diverse environments without robot-specific training.

Finn Lukas Busch, Matti Vahs, Quantao Yang et al.

Proposes bisimulation encoder for invariant visual representations, reducing latent size by 10×, enhancing robustness of JEPAs against background variations.

Leonardo F. Toso, Davit Shadunts, Yunyang Lu et al.

VIRAASAT enhances Indian cultural reasoning with Symbolic Chain-of-Manipulation, achieving up to 20% improvement.

Harshul Raj Surana, Arijit Maji, Aryan Vats et al.

Information-theoretic analysis plus Oracle RL and conditional-MI rewards improves CoT monitorability and reduces reward hacking.

Usman Anwar, Tim Bakker, Dana Kianfar et al.

SimVLA simplifies Vision-Language-Action models, achieving 98.6% success on LIBERO benchmarks with only 0.5B parameters, outperforming multi-billion models.

Yuankai Luo, Woping Chen, Tong Liang et al.

Proposes sinusoidal positional encoding to control symmetry breaking in graph flow models, balancing training speed and generalization.

Benjamin Honoré, Alba Carballo-Castro, Yiming Qin et al.

Gradient regularization (GR) mitigates reward hacking, improves reward model accuracy, outperforming KL penalty in RLHF tasks.

Johannes Ackermann, Michael Noukhovitch, Takashi Ishida et al.

Proposes MantisV2, using synthetic data and test strategies to close the zero-shot time series classification gap.

Vasilii Feofanov, Songkang Wen, Jianfeng Zhang et al.

Token Games evaluates language model reasoning by having models create their own puzzles, saving costs.

Simon Henniger, Gabriel Poesia

Grassmannian MoE (GrMoE) uses matrix Bingham distribution's concentration parameters to smoothly control routing entropy, preventing expert collapse.

Ibne Farabi Shihab, Sanjeda Akter, Anuj Sharma

Reverso model achieves efficient zero-shot time series forecasting using long convolutions and linear RNN layers, reducing parameters by 100x.

Xinghong Fu, Yanhong Li, Georgios Papaioannou et al.

WarpRec unifies academic rigor and industrial scale with a backend-agnostic architecture, supporting 50+ algorithms and integrating CodeCarbon for energy tracking.

Marco Avolio, Potito Aghilar, Sabino Roccotelli et al.

Introduced X-Value benchmark to assess cross-lingual values judgment in LLMs with 4,750 QA pairs.

Yukun Chen, Xinyu Zhang, Boyi Deng et al.

NeST uses activation probing and cluster-level training for efficient post-hoc safety alignment, reducing jailbreak success from 44.5% to 1.1%.

Sasha Behrouzi, Lichao Wu, Mohamadreza Rostami et al.

VETime achieves zero-shot time series anomaly detection through visual-temporal alignment and dynamic fusion, significantly improving localization precision.

Yingyuan Yang, Tian Lan, Yifei Gao et al.

Proposes Self-Supervised Semantic Bridge (SSB), integrating external semantic priors into diffusion models for spatially faithful unpaired image translation, outperforming prior methods.

Jiaming Liu, Felix Petersen, Yunhe Gao et al.

Proposes a feedback-loop model revealing that increasing recommendation adoption initially boosts but ultimately reduces individual diversity over time.

Gabriele Barlacchi, Margherita Lalli, Emanuele Ferragina et al.

MemoryArena evaluates agent memory in multi-session tasks, revealing gaps in current assessments.

Zexue He, Yu Wang, Churan Zhi et al.

NESSiE uses keyword matching to evaluate LLM safety, revealing models' bias towards helpfulness over safety with scores below 100%.

Johannes Bertram, Jonas Geiping

Training AI agents in Corecraft high-fidelity environment, GLM 4.6 task pass rate improved from 25.37% to 36.76%.

Sushant Mehta, Logan Ritchie, Suhaas Garre et al.