Speaker
Description
Industrial microbial production systems generate rich process data, yet translating these data into actionable parameter recommendations remains challenging. In this talk, we present a model based framework for generating and interpreting recommendations to optimize microbial production parameters using supervised machine learning.
Using two distinct industrial probiotic strains (GG and NCFM), we train predictive models to learn relationships between controllable process parameters and production outcomes. We then derive model based recommendations by systematically perturbing input parameters and evaluating predicted responses, enabling comparison of locally optimal parameter settings across organisms and production contexts. Emphasis is placed on model interpretability and production actionability, including the use of explainability techniques to assess feature importance and any predictor’s relationship with the outcome variable.
By contrasting model behavior and recommendations between GG and NCFM, we highlight how biological recommendation vary between similar probiotic strains and operational considerations. This comparison illustrates both the strengths and limitations of predictive modeling for decision support in industrial microbiology.
The case study demonstrates how machine learning models can be operationalized beyond prediction to support parameter selection, while maintaining transparency and biological plausibility. The presented word in intended as a instigator of thought and invitation to consider how a machine learning could be applied in real world production environments.
If you used AI tools or services to support the preparation of this submission, please state the name and reason for using each of them.
MS Copilot (Engine: OpenAI’s GPT "5, chat-optimized”):
- Abstract, for a first draft, before carefully editting
- Keyword sounding board and wordsmithing
- Title brainstorming for better names than dry literal description
- Material Licenses, to better understand CC-BY 4.0, it's neightbors, and how Fortune 500 companies typically view confrence presentations
| Keywords: Please list up to 5 keywords to help us find the right session for your contribution. | Machine learning, Process optimization, Explainable AI, Industrial microbiology, Applied data science |
|---|---|
| Virtual Option | This submission is for onsite presentation primarily, but I would also like it to be considered for pre-recorded virtual presentation if I don't get an onsite slot |
| Material License | CC‑BY‑NC 4.0 |
| Video Recording | Please don't share recordings of my talk |
| The author(s) agree(s) to take responsibility and be accountable for the contents of the submission and is/are authorized to present it. | Confirm |
| Interested in serving as reviewer? | nicholas.spyrison@iff.com |