Spotlight Others

Text-DiFuse: A novel interactive multi-modal image fusion framework leverages text-modulated diffusion models for superior performance in complex scenarios.

TeT-Splatting: a novel 3D representation enabling fast convergence, real-time rendering, and precise mesh extraction for high-fidelity 3D generation.

StoryDiffusion enhances long-range image & video generation by introducing a simple yet effective self-attention mechanism and a semantic motion predictor, achieving high content consistency without t…

S-DANE & ACC-S-DANE achieve best-known communication complexity for federated learning, improving local computation efficiency via stabilized proximal-point methods.

SpikeReveal: Self-supervised learning unlocks sharp video sequences from blurry, real-world spike camera data, overcoming limitations of prior supervised approaches.

SocraticLM achieves a Socratic teaching paradigm, surpassing GPT-4 by 12%, through a novel multi-agent training pipeline and a comprehensive evaluation system.

Slightly corrupting pre-training data significantly improves diffusion models’ image generation quality, diversity, and fidelity.

MeshRet: A novel retargeting framework that uses dense geometric interaction modeling for realistic, artifact-free skinned character animation.

This study proves that combining labeled and unlabeled data significantly improves high-dimensional sparse Gaussian classification, offering a polynomial-time SSL algorithm that outperforms supervised…

S2ML2-BBAM: A new semi-supervised multi-label learning method that balances feature angle distributions to improve accuracy and fairness.

Curated synthetic data provably optimizes human preferences in iterative generative model training, maximizing expected reward while mitigating variance.

SegVol: A universal, interactive 3D medical image segmentation model achieving state-of-the-art performance across diverse anatomical categories.

SOFO: a novel second-order optimizer enables efficient and memory-friendly RNN training for neuroscience tasks, surpassing Adam’s performance, especially on long time horizons.

Accelerate unpaired data translation with Schrödinger Bridge Flow, a novel algorithm solving optimal transport problems efficiently without repeatedly training models!

DisCo: a novel framework for generalizable complex image generation using scene graph disentanglement and composition, achieving superior performance over existing methods.

Heterogeneous Pre-trained Transformers (HPT) enables robots to learn generalizable policies from diverse data, drastically improving performance on unseen tasks.

Researchers scaled continuous latent variable models by building DAG-shaped probabilistic integral circuits (PICs) and training them efficiently using tensorized architectures and neural functional sh…

SARDet-100K: A new benchmark dataset and open-source toolkit revolutionizes large-scale SAR object detection.

Boosting AI’s continual learning via saliency-driven experience replay, achieving up to 20% accuracy improvement.

SA3DIP boosts 3D instance segmentation accuracy by cleverly using 3D spatial and textural cues alongside 2D multi-view masks, overcoming limitations of previous methods.