A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Vikram S Chundawat, Ayush K Tarun, Murari Mandal, et al.
Introduces zero-shot machine unlearning, enabling data deletion from ML models without any original training samples, using error minimizing-maximizing noise and gated knowledge transfer.
Melvin Johnson, Mike Schuster, Quoc V. Le, et al.
Proposes a single multilingual NMT model using a target-language token for zero-shot translation across multiple language pairs.
Supriya V. Mahadevkar, Bharti Khemani, Shruti Patil, et al.
A systematic review of machine learning styles in computer vision, covering supervised, unsupervised, reinforcement, and emerging techniques like zero-shot and self-supervised learning.
Caleb Ziems, William A. Held, Omar Ahmed Shaikh, et al.
This paper provides a roadmap for using zero-shot LLMs as tools in computational social science, showing they can augment human annotation and bootstrapping creative generation tasks.
Unknown
Introduces a novel image segmentation task, model, and dataset for prompt-able, zero-shot transfer learning.
Unknown
CLIP learns visual representations from 400M image-text pairs via contrastive pre-training, enabling zero-shot transfer to diverse tasks.
Unknown
InstructBLIP introduces instruction-aware Query Transformer for vision-language instruction tuning, achieving state-of-the-art zero-shot generalization.
Unknown
ModernBERT-Large-Instruct uses its MLM head for generative classification, achieving zero-shot performance rivaling larger LLMs.
Unknown
Multitask prompted fine-tuning on English tasks enables zero-shot cross-lingual generalization in large multilingual models.
Jason Wei, Maarten Bosma, Vincent Y. Zhao, et al.
Instruction tuning on 62 NLP datasets with natural language templates enables a 137B parameter model to achieve strong zero-shot performance on unseen tasks.
Kexin Huang, Payal Chandak, Qianwen Wang, et al.
TxGNN is a graph foundation model for zero-shot drug repurposing that identifies therapeutic candidates across 17,080 diseases, improving prediction accuracy by 49.2% for indications and 35.1% for contraindications.
Pengfei Liu, Weizhe Yuan, Jinlan Fu, et al.
This paper systematically surveys prompt-based learning in NLP, a paradigm where pre-trained language models are adapted to tasks via textual prompts for few-shot or zero-shot learning.