Applying Causal Inference in Industry 4.0: A Case Study from Glasswool Production
summary
Case study from Saint-Gobain with Inria claims a measurable cut in raw material use reduced raw material usage without compromising product quality and shows the causal chain: expert-built DAG, identified confounders, continuous treatments and double machine learning. Implementation uses DoWhy, EconML, DAGitty. Reported estimates include a specific effect — spinner used twice as often → quality drop ≈ 1.66 (product-specific effects up to ~2.1 or ~0.88) — and recommendations are ready but plant trials are pending.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:07Glass wool production
- 6:09Machine learning limits
- 7:33Causal inference basics
- 10:29Double ML for estimation
- 11:54Causal workflow in practice
- 13:58Building causal graph
- 17:29Graph validation steps
- 18:53Estimation and sensitivity
- 22:44Business impact results
- 24:11When to use causal
- 26:22Q&A: Missing confounders
- 28:05Q&A: CATE learners
- 29:28Q&A: Business adoption
- 30:51Q&A: Treatment effect heterogeneity
- 32:40Q&A: Validating in production
related talks
34mInterrogating your twin: causal reasoning in manufacturing systems
24mAn introduction to Causal Inference with Python – making accurate estimates of cause and effect from
41mOn the Path to Causal Inference
29m🧠🧹 Causality - Mental Hygiene for Data Science
11mCausal Inference For Socio-Economic & Engineering Systems
31mPython and the Holy Grail of Causal Inference
35mIs it a predictive model? Is it causal inference? Well... It is running a greenhouse.
28mTurning a Wasting into a Learning Culture