A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Unknown
Viewformer is a NeRF-free neural rendering method that uses transformers to synthesize novel views from few sparsely covering context images.
Unknown
Q-dit introduces a post-training quantization method for diffusion transformers that reduces memory and latency while maintaining generation quality.
Unknown
Lazydit introduces a lazy learning approach to accelerate Diffusion Transformers by selectively skipping redundant computations during inference.
Unknown
The paper argues that semantically richer latent representations can substantially improve diffusion Transformer optimization efficiency and expand the reconstruction–generation Pareto frontier.
Unknown
A statistical and approximation theory framework predicting neural scaling laws for transformers on intrinsically low-dimensional data, quantifying data and model scaling exponents.
Unknown
IRIS introduces a Transformer-based world model that learns from pixels and achieves human-level performance on Atari 100k with high sample efficiency.
Jos'ephine Raugel, Marc Szafraniec, Huy V. Vo, et al.
This paper systematically varies model size, training amount, and image type in DINOv3 vision transformers to disentangle how these factors independently and interactively drive brain-like representations, revealing a developmental chronology aligned
Yin Dai, Yifan Gao, Fayu Liu
TransMed combines CNN and transformer architectures for multi-modal medical image classification, achieving significant accuracy improvements on parotid gland and knee injury datasets.
Haris Riaz, Hyungji Kim, Mihai Surdeanu
SiPE learns a lightweight syntactic prior from dependency parses during pretraining and injects it into positional embeddings, improving syntax and language understanding benchmarks.
Zitian Gao, Lynx Chen, Yihao Xiao, et al.
The paper proposes the Universal Reasoning Model (URM), which enhances universal transformers with short convolution and truncated backpropagation, achieving state-of-the-art results on ARC-AGI benchmarks.
Unknown
Introduces Structured 3D Latents (SLAT) with rectified flow transformers for scalable, versatile, and high-quality 3D asset generation.
Unknown
Quantsparse is a unified framework that integrates model quantization with attention sparsification to comprehensively compress video diffusion transformers.