A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Shubham Vatsal, Harsh Dubey
A survey of 44 papers on 39 prompt engineering methods across 29 NLP tasks, showing how structured prompts improve LLM performance without retraining.
Yuqing Yang, Yan Ma, Pengfei Liu
Proposes a progressive weak-to-strong reasoning framework where a strong model refines its own training data without human or advanced model input, significantly improving reasoning on GSM8K and MATH.
Keyon Vafa, Peter G. Chang, Ashesh Rambachan, et al.
Proposes an inductive bias probe to evaluate whether foundation models capture deeper domain structure, finding they often fail to generalize beyond training tasks.
Chengshuai Zhao, Zhen Tan, Pingchuan Ma, et al.
Proposes a data distribution lens to understand when and why Chain-of-Thought reasoning succeeds or fails, revealing it as a brittle mirage beyond training distributions.
Rachit Bansal, Aston Zhang, Rishabh Tiwari, et al.
Proposes test-time training via targeted gradient updates on context to overcome static self-attention limitations in long-context LLMs, outperforming inference-time scaling.
Yaoming Li, Guangxiang Zhao, Qilong Shi, et al.
First primer synthesizing over 150 studies on post-training reasoning data, organizing the field around four key questions.
Gokul Yenduri, M. Ramalingam, G. Chemmalar Selvi, et al.
A comprehensive review of GPT covering architecture, training, enabling technologies, applications, challenges, and future directions.
Tianyu Gu, Kang Liu, Brendan Dolan-Gavitt, et al.
This paper demonstrates that outsourced training of deep neural networks introduces security risks where adversaries can create backdoored networks that perform well on normal inputs but fail on attacker-chosen inputs.
Takeru Miyato, Shin‐ichi Maeda, Masanori Koyama, et al.
Virtual Adversarial Training (VAT) regularizes models by enforcing local smoothness of the label distribution via a computationally efficient, label-free adversarial direction, achieving state-of-the-art semi-supervised learning on SVHN and CIFAR-10.
Matteo Stefanini, Marcella Cornia, Lorenzo Baraldi, et al.
A comprehensive survey of deep learning-based image captioning, covering visual encoding, text generation, training strategies, datasets, and evaluation metrics.
Ilia Shumailov, Zakhar Shumaylov, Yiren Zhao, et al.
This paper demonstrates that training generative AI models on recursively generated data leads to irreversible model collapse, where tails of the original distribution disappear.
Antonio Emanuele Ciná, Kathrin Grosse, Ambra Demontis, et al.
A comprehensive survey systematizing 15 years of poisoning attacks and defenses in machine learning, with a focus on computer vision.