ImageNet Classification with Deep Convolutional Neural Networks
2012Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton
A landmark paper that introduced AlexNet, demonstrating the power of deep convolutional neural networks on the ImageNet dataset and sparking the deep learning revolution.
Read Paper Attention is All You Need
2017Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, et al.
Introduced the Transformer architecture, which relies entirely on attention mechanisms, revolutionizing NLP and enabling models like BERT and GPT.
Read Paper Playing Atari with Deep Reinforcement Learning
2013Volodymyr Mnih, Koray Kavukcuoglu, David Silver, et al.
Demonstrated the use of deep Q-networks (DQN) to play Atari games directly from pixels, a milestone in deep reinforcement learning.
Read Paper Generative Adversarial Nets
2014Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, et al.
Introduced GANs, a framework for training generative models via adversarial processes, leading to breakthroughs in image synthesis.
Read Paper BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
2018Jacob Devlin, Ming-Wei Chang, Kenton Lee, Kristina Toutanova
Presented BERT, a transformer-based model pre-trained on large corpora, achieving state-of-the-art results on a wide range of NLP tasks.
Read Paper Mastering the Game of Go with Deep Neural Networks and Tree Search
2016David Silver, Aja Huang, Chris J Maddison, et al.
Describes AlphaGo, the first computer program to defeat a world champion in Go, combining deep neural networks and Monte Carlo tree search.
Read Paper DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
2019Victor Sanh, Lysandre Debut, Julien Chaumond, Thomas Wolf
Introduced DistilBERT, a smaller and faster version of BERT, making transformer models more accessible for practical applications.
Read Paper YOLOv1: You Only Look Once: Unified, Real-Time Object Detection
2016Joseph Redmon, Santosh Divvala, Ross Girshick, Ali Farhadi
Proposed YOLO, a real-time object detection system that reframes detection as a single regression problem, enabling fast and accurate detection.
Read Paper Tips for Reading Research Papers
Read this insightful guide: How to Read a Paper (Stanford)
This article is highly recommended for students new to reading research papers. It provides a practical, step-by-step approach to understanding and analyzing academic work.
- Start by reading the abstract and conclusion to get the main idea.
- Skim the introduction and figures to understand the motivation and results.
- Focus on the methods and experiments sections for technical details.
- Don't get stuck on every equation—try to grasp the intuition first.
- Take notes and summarize each section in your own words.
- Look up unfamiliar terms or references as you go.
- Discuss the paper with peers or mentors to deepen your understanding.