Proposed ClippedGossip algorithm achieves neighborhood convergence for non-convex decentralized training under Byzantine attacks, leveraging spectral gap γ and attack ratio δ.

Lie He, Sai Praneeth Karimireddy, Martin Jaggi

Proposes Nash-MTL, modeling gradient aggregation as a bargaining game, achieving proportional fairness and improving multi-task learning performance.

Aviv Navon, Aviv Shamsian, Idan Achituve et al.

Locally typical sampling enforces per-word information content constraints, reducing repetition and improving text quality in language generation.

Clara Meister, Tiago Pimentel, Gian Wiher et al.

Demonstrates that non-Gaussian higher-order input statistics enable fully-connected networks to autonomously develop convolutional, localized receptive fields, forming space-tiling patterns.

Alessandro Ingrosso, Sebastian Goldt

StyleGAN-XL uses Projected GAN strategy to successfully train on ImageNet, generating 1024² resolution images.

Axel Sauer, Katja Schwarz, Andreas Geiger

Formalizes causal sufficient, counterfactual, and actual causation definitions to enhance action-guiding XAI, improving fairness and interpretability.

Sander Beckers

Proposes functa framework, representing data points as neural implicit functions for multi-modal deep learning.

Emilien Dupont, Hyunjik Kim, S. M. Ali Eslami et al.

Proposed GPFQ, a greedy path-following post-training quantization method with provable error bounds across diverse input distributions and architectures.

Jinjie Zhang, Yixuan Zhou, Rayan Saab

ShapeFormer uses sparse representation with Transformer to generate high-quality 3D shape completions, outperforming existing methods.

Xingguang Yan, Liqiang Lin, Niloy J. Mitra et al.

Proposes a multi-scale attention-based semantic-guided VPR algorithm, achieving 92.2% Recall@1 on Oxford RobotCar, outperforming SOTA.

Valerio Paolicelli, Antonio Tavera, Carlo Masone et al.

Introducing UOT-Pooling, a unified, learnable global pooling layer based on unbalanced optimal transport, improving performance across multiple tasks.

Minjie Cheng, Hongteng Xu

Proposes a diffractive all-optical network for quantitative phase imaging, converting phase to intensity via deep learning-optimized passive layers, enabling compact, high-speed, low-power systems.

Deniz Mengu, Aydogan Ozcan

Proposes goal-conditioned RL algorithms with goal relabeling, reward shaping, and model-based planning, achieving 20-30% efficiency gains.

Minghuan Liu, Menghui Zhu, Weinan Zhang

Proposes Co-12 properties for systematic evaluation of 300+ XAI papers, emphasizing quantitative metrics.

Meike Nauta, Jan Trienes, Shreyasi Pathak et al.

Using Transformer models for novel protein design, enhancing function prediction accuracy.

Noelia Ferruz, Birte Höcker

Introduces CoAuthor dataset, analyzing GPT-3's language, ideation, and collaboration in interactive writing with 1445 sessions.

Mina Lee, Percy Liang, Qian Yang

COLD dataset with 37,480 Chinese offensive sentences; COLDETECTOR fine-tuned on it achieves 81% accuracy in offensive language detection.

Jiawen Deng, Jingyan Zhou, Hao Sun et al.

Proposes multiresolution hash encoding, enabling seconds-level training and tens-of-milliseconds rendering for neural graphics primitives with high quality.

Thomas Müller, Alex Evans, Christoph Schied et al.

DeepSpeed-MoE advances MoE model efficiency with novel architecture and compression, achieving 4.5x faster and 9x cheaper inference.

Samyam Rajbhandari, Conglong Li, Zhewei Yao et al.

Proposes physics-constrained adversarial attack boosting trajectory prediction error by 150%, highlighting safety risks in autonomous driving.

Qingzhao Zhang, Shengtuo Hu, Jiachen Sun et al.