Add computational terms to glossary

[matchings]

As a starting point (text taken from SWDB handbook google doc):

Computational terms definition cheat sheet - in plain language that experimentalists can understand 

• [ ] Low rank
• [ ] Pairwise correlation 
• [ ] Dimensionality reduction 
• [ ] PCA
• [ ] Subspace
• [ ] RNN
• [ ] Dynamical system 
• [ ] Network models
• [ ] Biophysically detailed model
• [ ] Data driven model
• [ ] HMM
• [ ] Functional connectivity 
• [ ] Connectivity matrix 
• [ ] Communication subspaces
• [ ] Tensor
• [ ] Attractor network
• [ ] Eigenvalue, eigenvector https://medium.com/fintechexplained/what-are-eigenvalues-and-eigenvectors-a-must-know-concept-for-machine-learning-80d0fd330e47
• [ ] Representational geometry “The representational geometry is defined by all pairwise distances in the representational space among the response patterns corresponding to a set of stimuli. “ Neural tuning and representational geometry | Nature Reviews Neuroscience
• [ ] State space - distributed patterns of activity across many neurons, aka latent variables Mante and Sussilo as key paper. Moving from idea of single cell activity (often averaged across trials) as functional unit to population dynamics (on single trials) as primary unit of computation.
• [ ] Low dimensional sub spaces - projections that extract the core of the computation. Neural state space means looking at activity where each neuron is a dimension and can fluctuate in that dimension. But most of variance in activity is not as high dimensional as the number of neurons. Activity often occupies low dimensional sub spaces.