Topic models and neural networks can discover meaningful low-dimensional latent representations of text corpora; as such. they have become a key technology of document representation. However. such models presume all documents are non-discriminatory. https://parisnaturalfoodes.shop/product-category/platinum-100-glutamine/