Since 1975
Highly productive single-pass inkjet technology, customizable, printing personalized short and medium-sized jobs
Unexpected Inkjet Solutions
Neos Bombadier B-Belt
info@tdwinc.com
www.magneticcylinders.com
www.neoslab.it -
sales@neoslab.it
resins by suggesting starting point formulations faster. To date, we have seen on average a 20% reduction in the amount of experimental lab work needed to complete a project using these models. The percentage reduction in experimental resources and the percentage of time devoted to data capture/retrieval are derived from internal Eastman AI/ML projects.
Taking advantage of AI/ML models
While our internal models are created and maintained by Eastman’s R&D and computer scientists, we also actively collaborate with external partners to maintain our competitive advantage.
Above: Accelerating innovation with AI in Eastman’s labs
For example, in collaboration with the University of Tennessee at Knoxville we have developed a multitask ML architecture that relies on polymeric features and graph neural networks to predict polyester properties. Additionally, we also work with Citrine Informatics to utilise their world- class ML models for materials science innovation. By leveraging the expertise and knowledge gained through these partnerships, Eastman is continuously improving our ML models to better analyse and understand the resin and formulation landscape. This enables the identification of innovative solutions in a more efficient and timely manner. We have found that the biggest
bottleneck in taking advantage of AI/ML is not in building these models, but rather in locating, reformatting and validating data itself. One of the key challenges faced by research laboratories in every discipline is the collection, storage and retrieval of data. With the introduction of new test methods, materials and processes, managing and organising data has become increasingly complex. Moreover, ensuring the level of detail required for comprehensive analysis adds
39
metalpackager.com
to the challenge. It is further compounded by the need to retrieve and access old data from files. Altogether, these challenges constitute
on average about 80% of the time and resources needed to successfully build a new AI/ML model at Eastman, regardless of whether we build that model internally or make use of an external platform. To address these challenges, automated systems are being developed to capture resin and formulation characterisation data into a standardised relational database, streamlining data management processes. Building such a database allows us to take advantage of AI/ML models and build data visualisation tools to aid in seeing what has been done and where any gaps may still be. In conclusion, innovation is an
evolving process that takes a lot of time and effort. Hiring a team of scientists is only a part of the story. If you want to do it effectively, your AI strategy must be well aligned with your overall growth strategy, and you must give your team the tools necessary to guide you through uncharted territory. Doing so can help your bottom line and help change the world in a material way.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52
