nLab categorical compositional distributional semantics

Contents

Idea

Categorical compositional distributional semantics, also known as DisCoCat for short, uses category theory to combine the benefits of two very different approaches to linguistics: categorial grammar and distributional semantics.

Specifically, it uses the fact that pregroups and the category of finite-dimensional real vector spaces are both examples of rigid monoidal categories. Semantics for pregroup grammars may then be given as a strong monoidal functor, which allows sentence meanings to be computed from word meanings by linear algebra, compositionally on the grammatical structure of the sentence.

DisCoCat was first introduced in Coecke, Sadrzadeh and Clark 2010.

There is a curious (if probably shallow) analogy between DisCoCat and topological quantum field theory, in that both concern strong monoidal functors from a free (rigid monoidal or compact closed) category to a category of vector spaces.

Simple example

Consider a very simple fragment of English consisting of nouns and verbs. Write nn for the type of nouns, and ss for the type of sentences. Let 𝒞\mathcal{C} be the free pregroup generated by nn and ss. In 𝒞\mathcal{C}, a noun will be assigned the type nn and a reflexive verb will be assigned the type n rsn ln^{r} s n^{l}, since it will produce a sentence if provided with nouns on the left and right. Now given nouns such as “Alice” and “Bob” and reflexive verbs such as “loves” and “hates”, the grammaticality of a sentence such as “Alice hates Bob” is witnessed by the fact that

nn rsn lns n n^{r} s n^{l} n \leq s

in the free pregroup.

A strong monoidal functor F:𝒞FVect F : \mathcal{C} \to FVect_{\mathbb{R}} is completely determined by where it sends the generators nn and ss. The vector spaces F(n)F (n) and F(s)F (s) are called the noun space and the sentence space, and their choice is a matter of linguistic modelling. A common approach is to pick a large set nn of basis words, say n=1000n = 1000 (or as many as your sparse matrix implementation can handle), and take F(n)= 1000F (n) = \mathbb{R}^{1000}. The choice of F(s)F (s) is not well understood, but F(s)= 2F (s) = \mathbb{R}^{2} is a common choice, with the standard basis vectors interpreted as “true” and “false”.

Given a grammatical parsing such as nn rsn lnsn n^{r} s n^{l} n \leq s, we automatically get a linear map

F(nn rsn ln):F(n)F(n)F(s)F(n)F(n)F(s) F (n n^{r} s n^{l} n) : F (n) \otimes F (n) \otimes F (s) \otimes F (n) \otimes F (n) \to F (s)

given by appropriate tensor contractions.

The final ingredient we need are word vectors. We need to pick vectors [[Alice]],[[Bob]]F(n)[\![ Alice ]\!], [\![ Bob ]\!] \in F (n) and [[loves]]F(n)F(s)F(n)[\![ loves ]\!] \in F (n) \otimes F (s) \otimes F (n). The choice of these vectors is also a matter for linguistics, but a common idea is to count how often Alice (for example) appears in a corpus (for example the full text of Wikipedia) close to each of the chosen basis words.

Finally, we obtain a sentence-vector

F(nn rsn lns)([[Alice]][[loves]][[Bob]])F(s) F (n n^{r} s n^{l} n \leq s) ([\![ Alice ]\!] \otimes [\![ loves ]\!] \otimes [\![ Bob ]\!]) \in F (s)

Relative pronouns and Frobenius algebras

Free pregroups vs free rigid monoidal categories

All earlier papers on DisCoCat use free pregroups for the grammar category, building on earlier work by Joachim Lambek on pregroup grammars. Unfortunately, in Preller 2014, fact 1 it is proven that any strong monoidal functor from a free pregroup to FVectFVect is necessarily trivial, in the sense that it sends every generator to the monoidal unit. One should instead use a free rigid monoidal category, a kind of categorification. Morphisms in the free autonomous category can be viewed as proof relevant grammatical reductions. They may be encoded as string diagrams, see pregroup grammar.

This allows a slightly more elegant reformulation of our basic example. Let 𝒞\mathcal{C} be the free autonomous category on the objects n,sn, s and the morphisms Alice:1nAlice : 1 \to n, Bob:1nBob : 1 \to n and loves:1n rsn lloves : 1 \to n^{r} s n^{l}. Then a sentence such as “Alice loves Bob”, together with its grammaticality, is witnessed by a morphism f:1sf : 1 \to s. Now the word-vectors [[Alice]]=F(Alice)[\![ Alice ]\!] = F (Alice), [[Bob]]=F(Bob)[\![ Bob ]\!] = F (Bob) and [[loves]]=F(loves)[\![ loves ]\!] = F (loves) become part of the data defining F:𝒞FVectF : \mathcal{C} \to FVect, and the resulting sentence vector is simply F(f)F (f).

The Category of DisCoCat Models

One criticism of DisCoCat is that a single model is unlikely to be complex enough to capture the intricacies of a natural language. A response to this criticism is that DisCoCat models should only model a given fragment of language with some level of detail. These models can then be combined or updated to represent more complex fragments of natural language. Indeed, in Translating and Evolving: Towards a Model of Language Change in DisCoCat the authors construct a category DisCoCatDisCoCat of DisCoCat models where

  • objects are strong monoidal functors

    F:(J,*)(FVect,) F \colon (J,*) \to (FVect,\otimes)

    where (J,*)(J,*) is a category which is freely monoidal on some set of grammatical types. In practice JJ might be freely equipped with an rigid structure or a Frobenius algebra on each object.

  • A morphism from a language model F:JFVectF \colon J \to FVect to a language model F:JFVectF' \colon J' \to FVect is a tuple (j,α)(j , \alpha) where jj is a monoidal functor i:JJi \colon J \to J' and α:FFj\alpha \colon F \Rightarrow F' \circ j is a monoidal natural transformation filling the following triangle

Morphisms in DisCoCatDisCoCat are called translations because they represent ways of changing your semantic interpretations of grammatical types in a way which is coherent with a change in grammar.

The Product Space Representation

In Mathematical foundations for a compositional distributional model of meaning a different description of a DisCoCat model was described. Let JJ be a free autonomous category representing grammar and assume that you have an assignment V aV_a of each grammatical type aa to a finite dimensional vector space. Then grammatical computations of meaning can be performed in the product category J×FVectJ \times FVect. Let a 1a na_1 \cdot \ldots \cdot a_n be an object in JJ and let vV a 1V a nv \in V_{a_1} \otimes \ldots \otimes V_{a_n} be its semantic meaning. Then a grammatical reduction f:abf \colon a \to b built from the units and counits in JJ can be paired with the corresponding structure maps in FVectFVect. This pairing allows you to apply the structure maps corresponding to ff to the vector vv to get a new semantic meaning. In essence, this approach identifies an autonomous subcategory of J×FVectJ \times FVect where meaning computations can occur.

This product space representation can be related to monoidal functor perspective via the category of elements. Using the forgetful functor i:FVectSeti \colon FVect \to Set the product space of a language model F:JFVectF \colon J \to FVect can be defined as the category of elements iF\int i \circ F. The product space representation is useful for equipping a language model with a specific set of words. If WW is a set of words then a map WiFW \to \int i \circ F describes an assignment of each word to a grammatical type and semantic meaning.

Variations

DisCoCat is relatively easy to modify by replacing the semantic category FVectFVect with some other category with the appropriate structure. Examples include

References

The category DisCoCatDisCoCat and it’s relationship to the product space representation is discussed in:

Last revised on December 9, 2020 at 14:17:23. See the history of this page for a list of all contributions to it.