TESTING
a powerful screening tool, significantly reducing the need for potentially redundant and labour- intensive experiments. The specialised computational techniques
applied for this model were vital, but equally important was the specific guidance taken by the domain experts. The IMPACT team and formulation scientists blend domain expertise with data science skills, reviewing the available data to determine the feasibility of data processing techniques. This interdisciplinary approach ensured the model was not only technically sound but also aligned with practical insights and expertise, driving towards the desired outcome. Building a library of historical data through high-quality experiments gradually creates a unique resource that can be used by machine learning models to predict the behaviour and performance of materials under various operating conditions and scenarios. By using computational modeling
techniques in combination with analysis and evaluation from Lucideon’s in-house formulation team, we create more accurate and robust simulations of materials or systems that can account for uncertainties, nonlinearities, or interactions among multiple factors. Creating a system with as much room for
complexity and nuance as the final model risks increasing the difficulty of effectively using the model for practical purposes. To avoid this, an interactive and user-friendly interface was developed and integrated into the tool to maximise the usability and impact of the predictive model for the end user. This interface served as a bridge between
complex data science algorithms and practical, real-world applications, making it accessible to users without requiring either deep technical expertise or extensive training. The interface was designed with a handful of features to
ensure smooth interaction and visualisation of both data and model predictions. Key features included:
1. Workflow Integration The interface was carefully built to mimic the workflow of a sensory panel. Users can utilise the model to rank how similar one toothpaste might be to a reference paste if it were to undergo sensory panel testing. This feature simplifies the process of
evaluating new formulations, making it intuitive and efficient.
2. Data Health-Check A dedicated function was included to routinely monitor the health of the foam characteristics data imported from the foam analyser. By comparing new data with previously evaluated toothpaste data, the system can identify discrepancies through specialised interactive plots.
This visual and interactive approach helps
users to quickly spot and understand any variations that may occur in the foam analyser data.
3. Variation Indication While not a robust health check for the model itself, this feature provides valuable insights into potential inconsistencies in foam analyser data. If data is taken in a batch and variations are detected, the system flags this to the user, indicating the foam characteristic data might not have been recorded in accordance with previous measurements and is therefore a candidate to be re-evaluated. The development of this interface was guided
by a commitment to making the predictive model as user-friendly and practical as possible. The intended outcome for the project was for the partner company to be capable of operating
47
a final model without requiring oversight or assistance from the Lucideon team. By streamlining the workflow and providing
intuitive visualisation tools, end-users have been empowered to leverage advanced data science techniques without needing extensive technical training. This ensures the predictive model can be effectively used in real-world applications, driving innovation and efficiency in the healthcare materials science industry.
Summary and conclusion By applying machine learning methodologies, a model has been created that can link the physical and analytical properties of prototype toothpaste formulations with the corresponding sensory panel data. On top of this model, a simple but impactful user interface was built that allows our partner company’s development teams to interact with and use the model with precision and without significant training requirements. The project output is therefore a packaged and sensory panel screening tool that can potentially reduce the number of required sensory panel tests and accelerate the product development cycle. By prioritising promising formulations for detailed examination and helping to focus efforts where they are most likely to yield valuable insights, the product development process is streamlined in a way that not only saves time and organisational capacity but also substantially cuts costs overall. By harnessing the power of Lucideon’s
IMPACT team, our partner company is now equipped with a powerful tool to accelerate and reduce the costs of toothpaste development. This innovation benefits both the company and consumers, ensuring only the most promising formulations reach the final stages of testing and allowing for developmental focus to be reallocated as needed.
PC
www.personalcaremagazine.com
November 2024 PERSONAL CARE
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 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100
