MVFlow leverages compressed video motion vectors to improve optical flow estimation, reducing AEPE by 1.09 and saving 52% computation time.

Shili Zhou, Xuhao Jiang, Weimin Tan et al.

OpenFlamingo is an open-source framework for training autoregressive vision-language models with 3B to 9B parameters, achieving 80-89% of Flamingo's performance.

Anas Awadalla, Irena Gao, Josh Gardner et al.

Proposes a certified multi-fidelity zeroth-order optimization algorithm MFDOO variant with near-optimal cost bounds based on the Sβ,L(f,ε) metric.

Étienne de Montbrun, Sébastien Gerchinovitz

Transforming overdetermined linear systems via sample abundance guarantees efficient one-bit sensing recovery with randomized Kaczmarz algorithms.

Arian Eamaz, Farhang Yeganegi, Deanna Needell et al.

LISA integrates large language models for reasoning segmentation, achieving over 56% gIoU on ReasonSeg with minimal fine-tuning.

Xin Lai, Zhuotao Tian, Yukang Chen et al.

Proposes GenQR and GenPRF frameworks using pretrained models to enhance retrieval effectiveness.

Xiao Wang, Sean MacAvaney, Craig Macdonald et al.

DriveAdapter employs feature alignment with adapters to decouple perception and planning, boosting autonomous driving performance.

Xiaosong Jia, Yulu Gao, Li Chen et al.

MetaGPT employs SOP-based structured prompts and role specialization, achieving 85.9% Pass@1 in code generation benchmarks with multi-agent collaboration.

Sirui Hong, Mingchen Zhuge, Jiaqi Chen et al.

Proposed a hybrid optimization algorithm with LSTM to enhance DER control resilience, validated on IEEE 37-node.

Mohammad Panahazari, Matthew Koscak, Jianhua Zhang et al.

Systematic analysis of 11 optical features reveals potential speed and energy-efficiency advantages over electronics.

Peter L. McMahon

This paper introduces Dynalang, a multimodal world model that predicts future text and images using diverse language cues, outperforming traditional language-conditioned policies.

Jessy Lin, Yuqing Du, Olivia Watkins et al.

Evaluates instruction-following models for QA correctness and faithfulness, proposing new metrics.

Vaibhav Adlakha, Parishad BehnamGhader, Xing Han Lu et al.

ToolLLM constructs ToolBench with 16,464 real APIs, employing DFSDT for multi-path reasoning, achieving near ChatGPT performance in tool use.

Yujia Qin, Shihao Liang, Yining Ye et al.

SpatialNet combines narrow-band self-attention and frequency convolution modules to leverage spatial cues for multichannel speech separation, denoising, and dereverberation, achieving SOTA results.

Changsheng Quan, Xiaofei Li

Proposed PACC and PACC-PE models significantly mitigate position bias in eCommerce search.

Yibo Wang, Yanbing Xue, Bo Liu et al.

OPTIMIS method improves efficiency by 3.5x and accuracy by 3x in high-dimensional SRAM evaluation.

Yanfang Liu, Guohao Dai, Wei W. Xing

Proposes ADTrans framework using semantic prototypes to address bias in PSG, improving R@100 by 3-4% and balancing long-tail relations.

Li Li, Wei Ji, Yiming Wu et al.

This survey formalizes RLHF’s three-stage pipeline and shows why human feedback is not a sufficient safety guarantee.

Stephen Casper, Xander Davies, Claudia Shi et al.

PointOdyssey constructs a large-scale synthetic dataset using real motion capture and structure-from-motion trajectories, improving long-term point tracking accuracy.

Yang Zheng, Adam W. Harley, Bokui Shen et al.

This paper uncovers four widespread flaws in offline recommender system evaluation, demonstrating their impact with empirical evidence.

Balázs Hidasi, Ádám Tibor Czapp