Combining experiments and simulations, this study investigates how aging affects nonlinear elasticity and memory in disordered networks, introducing two models: strength weakening and geometric distortion.

Daniel Hexner, Nidhi Pashine, Andrea J. Liu et al.

Using DNN-generated response patterns (RP) to fit IRT models, revealing task difficulty with high correlation to human data.

John P. Lalor, Hao Wu, Hong Yu

Introduces SPair-71k, a large-scale dataset with 70,958 image pairs across diverse viewpoints and scales, advancing semantic correspondence research.

Juhong Min, Jongmin Lee, Jean Ponce et al.

Proposes Once-for-All (OFA) network supporting diverse architectures, achieving 80.0% ImageNet top-1 accuracy with low complexity.

Han Cai, Chuang Gan, Tianzhe Wang et al.

Wasserstein-based distributionally robust optimization enhances model robustness against high-dimensional distributional uncertainty.

Daniel Kuhn, Peyman Mohajerin Esfahani, Viet Anh Nguyen et al.

The PROX method uses 3D scene constraints to significantly reduce 3D human pose estimation errors.

Mohamed Hassan, Vasileios Choutas, Dimitrios Tzionas et al.

LXMERT employs a multi-encoder Transformer architecture with multi-task pretraining, achieving state-of-the-art results on VQA, GQA, and NLVR2 datasets.

Hao Tan, Mohit Bansal

TabNet uses sequential attention to select features, enhancing tabular data learning performance and interpretability.

Sercan O. Arik, Tomas Pfister

Proposed a hierarchical RL framework using subtask controllers and a meta controller to synthesize human-chair interactions, achieving a 31.61% success rate.

Yu-Wei Chao, Jimei Yang, Weifeng Chen et al.

Constructed PubLayNet with 360k images via XML-PDF matching, enabling accurate document layout recognition and transfer learning.

Xu Zhong, Jianbin Tang, Antonio Jimeno Yepes

Proposes KBRD, an end-to-end knowledge-enhanced recommendation dialogue system, improving recommendation accuracy by 0.7 and dialogue diversity using R-GCN and Transformer.

Qibin Chen, Junyang Lin, Yichang Zhang et al.

Hierarchical sim2real framework enables zero-shot transfer of multi-agent manipulation policies trained in simulation, achieving over 80% success in real-world tasks.

Ofir Nachum, Michael Ahn, Hugo Ponte et al.

Proposed Unlikelihood Training significantly reduces repetition and dullness in neural text generation, outperforming standard methods.

Sean Welleck, Ilia Kulikov, Stephen Roller et al.

This paper proves attention weights in Transformers are non-identifiable for sequences longer than head dimension, introduces effective attention, and shows token identity is mainly encoded in embedding angles.

Gino Brunner, Yang Liu, Damián Pascual et al.

AutoGAN optimizes GAN architecture via NAS, achieving an FID score of 12.42 on CIFAR-10.

Xinyu Gong, Shiyu Chang, Yifan Jiang et al.

DoorGym uses domain randomization to train door-opening policies, achieving up to 95% success.

Yusuke Urakami, Alec Hodgkinson, Casey Carlin et al.

Pixel2Mesh++ integrates multi-view perceptual features with graph convolutional deformation, achieving 80.30 F-score on ShapeNet, outperforming single-view methods.

Chao Wen, Yinda Zhang, Zhuwen Li et al.

Proposes MIND module with deep RL to enhance environment modeling and planning in EmbodiedQA, improving accuracy and generalization.

Juncheng Li, Siliang Tang, Fei Wu et al.

AutoML advances neural architecture search (NAS), achieving 97.4% on CIFAR-10 and 75.2% Top-1 accuracy on ImageNet with efficient algorithms.

Xin He, Kaiyong Zhao, Xiaowen Chu

DIODE uses laser scanning to create high-density indoor-outdoor depth datasets, enabling cross-domain depth estimation.

Igor Vasiljevic, Nick Kolkin, Shanyi Zhang et al.