A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Haotian Wang, Han Zhao, Shuaiting Chen, et al.
This paper proposes using outputs from reasoning-intensive LLMs to improve non-reasoning models via supervised fine-tuning, showing consistent benchmark gains.
Yipeng Hu, Marc Modat, Eli Gibson, et al.
This paper introduces a weakly-supervised CNN approach for multimodal image registration that learns voxel-level transformations from anatomical labels, enabling real-time, fully-automated inference without labels.
Yijiang Li, Bingyang Wang, Yijun Liang, et al.
Proposes U-OPSD, an unsupervised on-policy self-distillation method that uses only a model's own generations and majority voting to improve reasoning without external supervision.
Unknown
Portal introduces content-specific tokenization for tabular data, enabling scalable self-supervised learning that outperforms prior methods on diverse benchmarks.
Unknown
This paper introduces TabDPT, a tabular foundation model that scales self-supervised learning on real-world tabular data to address data heterogeneity.
Unknown
This paper addresses the chicken-and-egg problem in causal representation learning by proposing a weakly supervised approach that leverages auxiliary labels to identify causal variables and graph.
Unknown
This paper explores self-supervised learning methods to obtain general-purpose feature embeddings by capturing visual invariances across domains.
Unknown
Data2vec is a general self-supervised learning framework that uses the same learning method for speech, NLP, and computer vision.
Unknown
Applying self-supervised learning to improve robustness and uncertainty of deep learning models beyond purely supervised methods.
Yao-Hung Hubert Tsai, Yue Wu, R. Salakhutdinov, et al.
This paper provides a theoretical framework for understanding self-supervised learning through a multi-view lens, explaining when and why it works.
Unknown
This paper provides a practical cookbook of fundamental self-supervised learning techniques, serving as a foundation for training and exploring SSL methods.
Unknown
The paper proposes a self-supervised learning method that learns pretext-invariant representations by explicitly enforcing invariance to pretext transformations, improving downstream task performance.