AI Journal - Classic and Prominent Paper Awards 2019

The AI Journal is pleased to Announce Two Awards for papers published in the journal:

- The AIJ Prominent Paper Award
recognizes outstanding papers published not more than seven years ago in the AI Journal that are exceptional in their significance and impact.

Nominations are now solicited for papers published in AIJ between 2012 and 2018 for the 2019 Prominent Paper Award.

Here is the complete list of this year's Eligible Prominent Papers.


The AIJ Classic Paper Award 
recognizes outstanding papers published at least 15 calendar years ago in the AI Journal that are exceptional in their significance and impact.

So for this year, eligible papers for the Classic Paper award are those published in AIJ between 1970 and 2002.

Here is the complete list of this year's Eligible Classic Papers.

The window for this round of nominations is closed. Information about the format of nominations and the evaluation criteria for these two awards can be found here:

AIJ Classic Paper Award.
AIJ Prominent Paper Award.

The 2019 Classic and Prominent Paper Awards were given to the following papers:

2019 AIJ Classic Paper Award

Between MDPs and semi-MDPs: A framework for temporal abstraction in reinforcement learning
RS Sutton, D Precup, S Singh - Artificial intelligence, 1999 - Elsevier

URL: https://www.sciencedirect.com/science/article/pii/S0004370299000521

This is the seminal paper on options, a widely used framework in reinforcement learning (RL) for representing actions
at different temporal scales and levels of abstraction. The option framework has  shaped subsequent work in the area,
and still defines  the terms in which the problem of representation learning of hierarchical  actions and knowledge is formulated and addressed in RL.

2019 AIJ Prominent Paper Award

The dropout learning algorithm
P Baldi, P Sadowski - Artificial intelligence, 2014 - Elsevier

URL: https://www.sciencedirect.com/science/article/pii/S0004370214000216

Dropout has become a key machine learning technique used mainly to avoid  overfitting when training large neural networks. The mathematical analysis in the paper uncovers the properties that have made dropout a cornerstone in deep learning. The article also provides connections to other areas of statistics, and to the theory of ensemble learning and spiking neurons.

Here is the list of All Previous Award Winners.