Artificial Intelligence Machine Learning Contest
2018

Overview 

New: We will use DN-2 in AIML Contest 2018!

The terms artificial intelligence, machine learning, robotics, signal processing, control, dynamic systems, data mining, big data, and brain projects often have different emphases, but the related disciplines are converging.   The Artificial Intelligence Machine Learning (AIML) Contest serves as a converging platform for these highly related disciplines and beyond.  It is open to, but not limited to, all researchers, practitioners, students and investors.   The main goal of the contest is to promote understanding of both natural intelligence and artificial intelligence, beyond the currently popular pattern classification.   The AIML Contest aims to address major learning mechanisms in natural and artificial intelligence, including perception, cognition, behavior and motivation that occur in cluttered real-world environments.   Attention, segmentation, emergence of spatiotemporal representations, and incremental scaffolding are part of each life-long learning stream.

The major characteristics of this contest include:

1. Use inspirations from learning by natural brains, such as grounding, emerging, natural inputs, incremental learning, real-time and online, attention, motivation, and abstraction from raw sensorimotor data.
   
2. General purpose learning engines that are task-independant.   Task-independant means that the learning engine is capable of being trained to generate a machine "brain" to learn and do any collection of body-capable and open-ended tasks.   Base engines will be available to participants and open for enhancements.  The providers of base engines are free to provide assistants to participants, such as courses, tutorials, and workshops.
   
3. Modality-independant engines.   Modalities that are the well-recognized bottlenecks of AI will be tested on the same machine learning engine from each contest entry, including vision, audition, language understanding, and autonomous thinking.
   
4. Training-and-testing sensorimotor streams will be provided to the participants.  Each frame of the stream contains a sensory vector and a motoric vector.  Training and testing are mixed in the streams, so that learning systems can perform scaffolding: early learned simpler skills are automatically selected and used for learning later more complex skills.
   

Organizers 

General Chair: Juyang Weng  

Program Chair: Juan Castro-Garcia

Vision Chair: Zejia Zheng

Audition Chair: Xiang Wu

Language Chair: Juan Castro-Garcia