Update 7-inverse-design.ipynb

docs/source/tutorials/7-inverse-design.ipynb

@@ -11,7 +11,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "This tutorial provides step by step guidance to all the essential components for running iQSPR, an inverse molecular design algorithm based on machine learning. We provide a set of in-house data and pre-trained models for demonstration purpose. We reserve all rights for using these resources outside of this tutorial. We recommend readers to have prior knowledge about python programming, use of numpy and pandas packages, building models with scikit-learn, and understanding on the fundamental functions of XenonPy (refer to the tutorial on the descriptor calculation and model building with XenonPy). For any question, please contact the developer of XenonPy. (v.0.4.1)\n",
+    "This tutorial provides step by step guidance to all the essential components for running iQSPR, an inverse molecular design algorithm based on machine learning. We provide a set of in-house data and pre-trained models for demonstration purpose. We reserve all rights for using these resources outside of this tutorial. We recommend readers to have prior knowledge about python programming, use of numpy and pandas packages, building models with scikit-learn, and understanding on the fundamental functions of XenonPy (refer to the tutorial on the descriptor calculation and model building with XenonPy). For any question, please contact the developer of XenonPy.\n",
     "\n",
     "Overview - We are interested in finding molecular structures 𝑆 such that their properties 𝑌 have a high probability of falling into a target region 𝑈, i.e., we want to sample from the posterior probability 𝑝(𝑆|𝑌∈𝑈) that is proportional to 𝑝(𝑌∈𝑈|𝑆)𝑝(𝑆) by the Bayes’ theorem. 𝑝(𝑌∈𝑈|𝑆) is the likelihood function that can be derived from any machine learning models predicting 𝑌 for a given 𝑆. 𝑝(𝑆) is the prior that represents all possible candidates of S to be considered. iQSPR is a numerical implementation for this Bayesian formulation based on sequential Monte Carlo sampling, which requires a likelihood model and a prior model, to begin with.\n",
     "\n",