Semantic Spaces
Any natural language can be considered as a tool for producing large databases (consisting of texts, written, or discursive). This tool for its description in turn requires other large databases (dictionaries, grammars etc.). Nowadays, the notion of database is associated with computer processing and computer memory. However, a natural language resides also in human brains and functions in human communication, from interpersonal to intergenerational one. We discuss in this survey/research paper mathematical, in particular geometric, constructions, which help to bridge these two worlds. In particular, in this paper we consider the Vector Space Model of semantics based on frequency matrices, as used in Natural Language Processing. We investigate underlying geometries, formulated in terms of Grassmannians, projective spaces, and flag varieties. We formulate the relation between vector space models and semantic spaces based on semic axes in terms of projectability of subvarieties in Grassmannians and projective spaces. We interpret Latent Semantics as a geometric flow on Grassmannians. We also discuss how to formulate G\"ardenfors' notion of "meeting of minds" in our geometric setting.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Cross-lingual Word Analogies using Linear Transformations between Semantic Spaces
We generalize the word analogy task across languages, to provide a new intrinsic evaluation method for cross-lingual semantic spaces. We experiment with six languages within different language families, including English…
Searching for Programmatic Policies in Semantic Spaces
Syntax-guided synthesis is commonly used to generate programs encoding policies. In this approach, the set of programs, that can be written in a domain-specific language defines the search space, and an algorithm searche…
Semantics of higher-order probabilistic programs with conditioning
We present a denotational semantics for higher-order probabilistic programs in terms of linear operators between Banach spaces. Our semantics is rooted in the classical theory of Banach spaces and their tensor products, …
Modelling Salient Features as Directions in Fine-Tuned Semantic Spaces
In this paper we consider semantic spaces consisting of objects from some particular domain (e.g. IMDB movie reviews). Various authors have observed that such semantic spaces often model salient features (e.g. how scary …
Knowledge Graph EmbeddingsRecommendation SystemsWord EmbeddingsOperational Latent Spaces
We investigate the construction of latent spaces through self-supervised learning to support semantically meaningful operations. Analogous to operational amplifiers, these "operational latent spaces" (OpLaS) not only dem…
ClusteringMetric LearningSelf-Supervised Learning