HullFT employs convex reconstruction and gradient caching for efficient test-time fine-tuning, improving speed and quality tradeoff in large language models.

Alaa Khamis, Alaa Maalouf

Proposes COMPOSE, a dual-graph framework combining citation and formal theorem graphs, generating plausible future theorems with 108K training pairs and 47K future papers tested.

David Busbib, Michael Werman

Proposes MedCase-Structured, a pipeline combining LLMs and terminology validation to generate HL7 FHIR R4 clinical datasets for diagnostic reasoning, with an 82.5% success rate.

Valentina Bui Muti, Eugénie Dulout, Ziquan Fu

Proposed dual-path block architecture enhances compute and capacity in LLMs at fixed FLOPs.

Markus Frey, Behzad Shomali, Joachim Koehler et al.

Combining multi-objective genetic programming with survival tree optimization, this study enhances predictive accuracy and interpretability in survival analysis, validated on two real-world datasets.

Thalea Schlender, Peter A. N. Bosman, Tanja Alderliesten

SGMD introduces gradient-based pseudo-score distillation with dual potentials, tripling training speed and enhancing motion dynamics in 4-step video diffusion.

Zhuguanyu Wu, Ruihao Gong, Yang Yong et al.

On Navier-Stokes benchmark, ConvAE-Relay achieves 38.34% error under 10x Reynolds shift, outperforming FNO's 46.68%.

Jianing Shi

DirectorBench uses personalized multi-agent diagnosis and finds cross-shot transitions average only 0.256.

Jiamin Chen, Qianben Chen, Jiawen Zhang et al.

A multi-agent system with semantic communication and checkpoints enhances adaptive method selection.

Geremy Loachamín-Suntaxi, Robert Lazar, Dimitrios G. Giovanis et al.

EarlyTom performs early-stage token compression inside the vision encoder, reducing TTFT by 2.65× and FLOPs by 61%, maintaining accuracy.

Hesong Wang, Xin Jin, Lu Lu et al.

Proposed a causal intervention method for continuous variables, applied to verb bias, validating its role in in-context learning.

Zhenghao Herbert Zhou, R. Thomas McCoy, Robert Frank

Multi-horizon horizon-aware GCN emulator improves long-term ice sheet forecasts, reducing RMSE by 25% over 20-year simulations.

Zesheng Liu, Maryam Rahnemoonfar

Introduces PLANAHEAD framework, systematically evaluating natural language plan representations' impact on multimodal LLM web agent performance.

Alejandra Zambrano, Sara Vera Marjanovic, Imene Kerboua et al.

Proposes Cayley-Table Completion with tensor factorization and flatness prior to recover discrete algebraic structures via continuous optimization.

Dongsung Huh

Quantifying and optimizing simplicity via polynomial representations enhances generalization in image and text classification.

Tianren Zhang, Xiangxin Li, Minghao Xiao et al.

LFQ method enhances low-bit quantized LLMs' generation quality by minimizing cross-entropy in the final block.

Jung Hyun Lee, June Yong Yang, Jungwook Choi et al.

Rec-Distill employs decoupled training to distill 24B-parameter teacher models, achieving over 60% transferability in industrial recommendation tasks.

Haoran Ding, Wenlin Zhao, Yuchen Jiang et al.

UKA introduces user-aware active knowledge acquisition using ToM uncertainty, boosting emotional support dialogue quality.

Mufan Xu, Kehai Chen, Jiahao Hu et al.

GODeform framework leverages geometry-guided feature modeling to achieve significant improvements in monocular 3D shape recovery.

Yiyao Ma, Kai Chen, Zhongxiang Zhou et al.

SCOPE integrates a frozen LLM with an open-set plugin classifier, achieving 91.05% open-set detection accuracy and 96.63% anomaly correction in ATC readback monitoring.

Qihan Deng, Minghua Zhang, Yang Yang et al.