Consider a toy word co$-$occurrence matrix in Figure $2.$ This vector$-$space

model provides distributional representations for tokens $$stone$,$jazz$$ and

$$rock$.$ Identify in what way representations of these tokens can be seen as

problematic and discuss what language phenomenon could have led to this

issue. We often use distributed representations in downstream tasks, please

give an example of a task which would be negatively affected and explain

why. Discuss in $1$ or $2$ sentences how this limitation of the traditional vectorspace models can be addressed.