-
Workshop: Artificial Intelligence for Materials Science (AIMS)
August 1-2, 2019
-
The Materials Genome Initiative (MGI) promises to expedite materials discovery through high-through computation and high-throughput experiments. While the MGI effort has been successful to screen interesting materials among thousands of materials, the possible materials can span up to 10100 limiting the current MGI philosophy.
One of the possible approaches to deal with this problem is using artificial-intelligence (AI) tools such as machine-learning, deep-learning and various optimization techniques to efficiently evaluate materials performance. Although AI has been very successful in fields such as voice-recognition, self-driving cars, language translation etc., its applicability to materials design is still in its developing phase. Two key challenges in employing AI techniques to materials are: choosing effective descriptors for materials and choosing algorithm/work-flow during AI design. The idea of including physics-based models in the AI framework is also fascinating. Lastly, uncertainty quantification in AI based predictions for material properties and issues related to building infrastructure for disseminating AI knowledge are of immense importance for making AI based investigation of materials successful. This workshop is intended to cover all the above-mentioned challenges. To make the workshop as effective as possible we plan to mainly focus on inorganic solid-state materials, but are not limited by it.
Information for participants:Please note that the seats are limited and priority would be given based on the date of your registration and the registration closes on July 25. We are expecting around 150 participants, so please arrive at the NIST gate between 7:30-8:00 AM. The workshop won't cover travel and living expenses of the attendees. We recommend 46 inch x 46 inch size for the posters. Participants should bring their laptops for the hands-on session. We will be using jarvis-tools and Google-colab for the hands-on session, but you may also install the jarvis-tools package on your laptop from https://github.com/usnistgov/jarvis .
NIST employees/associates do not need to register, just walk in. There will be a VTC connection to NIST-Boulder. We have reserved room 1-4072 (7:00 AM-2:30 PM) there both days.
Workshop topics
-
bookmark_borderApplication of classification/regression techniques
-
bookmark_borderApplication of physics-based constraints
-
bookmark_borderSelection and importance of features/descriptors
-
bookmark_borderComparison metrics of AI techniques
-
bookmark_borderChallenges applying AI to materials
-
bookmark_borderDataset and tools for employing AI.
-
bookmark_borderIntegrating experiments with AI techniques
-
bookmark_borderUsing AI to develop classical force-fields
Invited speakers














