Update README.md

README.md

@@ -24,12 +24,13 @@ One notable feature is that the architecture (trained or not) admits a *continuo
 FAQ (as the author imagines):
 
 - Q: Who needs another ConvNet, when the SOTA for ImageNet-1k is now in the low 80s with models of comparable size?
-- A: Aside from shortage of resources to perform extensive experiments, the real answer is that the new symmetry has the potential to be exploited, in different ways. The non-elementwise nonlinearity does have more "natural"-ness (coordinate independence) that is inherent in equations in mathematics and physics.
+- A: Aside from shortage of resources to perform extensive experiments, the real answer is that the new symmetry has the potential to be exploited, in different ways (e.g., in optimization). The non-elementwise nonlinearity does have more "natural"-ness (coordinate independence) that is inherent in equations in mathematics and physics.
 - Q: Multiplication is too simple, someone must have tried it?
 - A: Perhaps. My bet is whoever tried it soon found the model fail to train with standard ReLU. Without the belief in the underlying PDE perspective, maybe it wasn't pushed to its limit.
 - Q: Is it not similar to attention in Transformer?
 - A: It is, indeed. It's natural to wonder if the activation functions in Transformer could be removed (or reduced) while still achieve comparable performance.
-
+- Q: If the weights has a symmetry (other than permutations), perhaps there's redundancy in the weights.
+- A: The transformation in the demo indeed can be used to reduce the weights from the get-go. However, there are variants that admit an even large symmetry that would be hard to remove. Also it is related to the phenomenon of "flat minima" found empirically in conventional Deep neural network models.
 
 *This modelcard aims to be a base template for new models. It has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/modelcard_template.md?plain=1).*