Udacity - Global AI Challenger

Artificial Intelligence has hit fever pitch in China. From cutting-edge tech powerhouses to established large corporations, everyone is setting up new AI Labs, hiring new talent, and announcing new platforms. “AI First” is the order of the day in China, and the excitement around this transformational technology is incredible. Udacity is thrilled to be playing a key role, through important partnerships with companies such as Tencent, Didi Chuxing, and Alibaba.

Competitions have emerged as an especially compelling way to engage with a global community of AI enthusiasts, and recent challenge events include the Self-Driving Car Challenge with Didi Chuxing, and the Customer Flow Forecasts competition with Alibaba. These competitions feature large cash prizes, opportunities for career advancement, and for our students, a chance to practice and demonstrate newly-mastered skills with real-world data and projects.

Today, we’re thrilled to share details of a new AI competition and open data initiative featuring three amazing companies:

  • Sinovation Ventures, which is led by Kai-fu Lee, a leading tech venture firm in China
  • Sogou, a major Chinese search engine and AI company
  • Toutiao (ByteDance), a leading AI powered information distributor

Together, they are launching a joint global AI competition, called the AI Challenger, which offers a grand prize of US $300,000!

This competition is the result of a unique cross-domain, private and public partnership bound by a collective mission to build a more open, collaborative, and international AI ecosystem that will “see” and “read” our world more clearly, more effectively, and more productively.

There will be five tracks in the competition, built around three large datasets, including Human Skeletal System Keypoints, which features more than 300,000 images, each annotated with 14 critical skeletal points across over 700,000 people. The total number of people annotated, as well as the complexity of the scenes, human movements, and physical occlusions of this dataset are much higher than many other datasets available today. Research results based on this dataset can be applied to:

  • Autonomous Driving (pedestrian movements)
  • Safety & Security (abnormal intrusions)
  • Human-Computer Interaction (Kinect games, motion analysis)
  • and more …

The 2 tracks of competition for this dataset will be a Scene Classification Challenge and a Human Skeletal System Keypoints Detection Challenge. The former asks students to develop new algorithms to improve the state of the arts in visual scene understanding, and the latter asks students to simultaneously detect people, and localize their keypoints in challenging, uncontrolled conditions.

As our international partnerships continue to expand, we look forward to exploring more competition opportunities, including events such as the Path Planning Challenge that we recently launched with Bosch, which offers competitors the chance to compete for an interview with Bosch’s Automated Driving group.

We’re excited to see where these global challenges take our students in their journeys towards the jobs of tomorrow, and we’re excited for students to enter the AI Challenger today!