Introduced ebc-exstream, leveraging model explanations and human feedback to detect spurious correlations in concept drift, reducing annotation costs.

Cristiana Lalletti, Stefano Teso

OriGen employs code-to-code augmentation and self-reflection to improve RTL code generation, surpassing open-source and some commercial models with 12.8% higher pass@1.

Fan Cui, Chenyang Yin, Kexing Zhou et al.

Proposes CMR, combining neural rule memory with symbolic evaluation for interpretable, verifiable AI, achieving high accuracy and logical rule discovery.

David Debot, Pietro Barbiero, Francesco Giannini et al.

SNIGL model leverages PNS theory to extract necessary and sufficient invariant subgraphs, boosting graph OOD generalization.

Xuexin Chen, Ruichu Cai, Kaitao Zheng et al.

AutoVCoder integrates high-quality data, two-round fine-tuning, and RAG to improve Verilog code generation accuracy.

Mingzhe Gao, Jieru Zhao, Zhe Lin et al.

Open-source Latent Diffusion model achieves high-quality 44.1kHz stereo audio synthesis, comparable to SOTA, trained solely on CC-licensed data.

Zach Evans, Julian D. Parker, CJ Carr et al.

Proposes noise selection and optimization based on inversion stability, improving diffusion model outputs by up to 72.5%.

Zipeng Qi, Lichen Bai, Haoyi Xiong et al.

Reinforcement learning (e.g., PPO, path consistency) fine-tunes diffusion models to maximize downstream rewards, applicable in bio-sequences and image generation.

Masatoshi Uehara, Yulai Zhao, Tommaso Biancalani et al.

Proposes three methods—IsoFLOPs analysis, derivative estimation, and loss fitting—to predict optimal vocabulary size for LLMs, finding larger models need bigger vocabularies.

Chaofan Tao, Qian Liu, Longxu Dou et al.

MEDIC system employs Disentangled Inversion Control to enable zero-shot music editing, significantly improving content preservation and editing fidelity.

Huadai Liu, Jialei Wang, Xiangtai Li et al.

MLSA4Rec combines Mamba with low-rank self-attention to enhance recommendation accuracy.

Jinzhao Su, Zhenhua Huang

SharpBalance preserves ensemble diversity while reducing sharpness, reaching 76.12% on CIFAR-10 versus 75.45% for SAM.

Haiquan Lu, Xiaotian Liu, Yefan Zhou et al.

R+X leverages vision-language models for retrieval and in-context imitation learning, enabling robots to learn from unlabelled human videos without training.

Georgios Papagiannis, Norman Di Palo, Pietro Vitiello et al.

VD3D introduces Plücker coordinate-based spatiotemporal camera embeddings for controlling large video transformers, achieving state-of-the-art accuracy.

Sherwin Bahmani, Ivan Skorokhodov, Aliaksandr Siarohin et al.

Introduces LMMS-EVAL, LMMS-EVAL LITE, and LIVEBENCH, covering 50+ tasks, to address the evaluation trilemma of large multimodal models.

Kaichen Zhang, Bo Li, Peiyuan Zhang et al.

Proposes OPZO, a single-pass online training method with noise and momentum feedback, achieving performance comparable to spatial BP with lower costs.

Mingqing Xiao, Qingyan Meng, Zongpeng Zhang et al.

Case2Code task enhances code generation via large-scale synthetic data.

Yunfan Shao, Linyang Li, Yichuan Ma et al.

STAD employs Bayesian filtering within state-space models to adapt to gradually evolving temporal distribution shifts, outperforming traditional TTA methods.

Mona Schirmer, Dan Zhang, Eric Nalisnick

ProtoFlow combines concept networks and generative classifiers, achieving 99.36% accuracy with interpretability and invertibility.

Zachariah Carmichael, Timothy Redgrave, Daniel Gonzalez Cedre et al.

NeedleBench evaluates LLMs' retrieval and reasoning across sparse and dense contexts, revealing weaknesses in dense reasoning tasks.

Mo Li, Songyang Zhang, Taolin Zhang et al.