Mario Baroni
Biographic Data
| ID | 157513 |
|---|---|
| NAME | Mario Baroni |
| GIVEN NAMES | Mario |
| FAMILY NAME | Baroni |
| SIGNATURE | BARONI M |
| AFFILIATIONS | University of Trento |
| ORCID | 0000-0001-5066-3580 |
| VERIFIED | Yes |
| TOTAL WORKS | 21 |
| TOTAL CITATIONS | 36 |
| WORKS AS AUTHOR | 21 |
| WORKS AS EDITOR | 0 |
| FIRST PUBLICATION YEAR | 1982 |
| LATEST PUBLICATION YEAR | 2021 |
| H-INDEX | 2 |
Mechanisms for handling nested dependencies in neural-network language models and humans
Syntactic Structure from Deep Learning
Modern deep neural networks achieve impressive performance in engineering applications that require extensive linguistic skills, such as machine translation. This success has sparked interest in probing whether these models are inducing human-like grammatical knowledge from the raw data they are exposed to and, consequently, whether they can shed new light on long-standing debates concerning the innate structure necessary for language acquisition…
When the Whole Is Less Than the Sum of Its Parts
Distributional semantic models, deriving vector-based word representations from patterns of word usage in corpora, have many useful applications (Turney and Pantel 2010 ). Recently, there has been interest in compositional distributional models, which derive vectors for phrases from representations of their constituent words (Mitchell and Lapata 2010 ). Often, the values of distributional vectors are pointwise mutual information (PMI) scores obta…
There Is No Logical Negation Here, But There Are Alternatives
Logical negation is a challenge for distributional semantics, because predicates and their negations tend to occur in very similar contexts, and consequently their distributional vectors are very similar. Indeed, it is not even clear what properties a “negated” distributional vector should possess. However, when linguistic negation is considered in its actual discourse usage, it often performs a role that is quite different from straightforward l…
Grounding Distributional Semantics in the Visual World
Distributional semantic models build vector-based word meaning representations on top of contextual information extracted from large collections of text. Object recognition methods from computer vision derive vector-based representations of visual content from natural images. This article reviews how methods from computer vision are exploited to tackle the fundamental problem ofgroundingdistributional semantic models, bringing them closer to prov…
When the Whole Is Not Greater Than the Combination of Its Parts
Distributional semantics has been extended to phrases and sentences by means of composition operations. We look at how these operations affect similarity measurements, showing that similarity equations of an important class of composition methods can be decomposed into operations performed on the subparts of the input phrases. This establishes a strong link between these models and convolution kernels
Composition in Distributional Semantics
Distributional Semantic Models, which automatically induce word meaning representations from naturally occurring textual data, are a success story of computational linguistics. Recently, there has been much interest in whether such models, endowed with a compositional component, can also successfully approximate the meaning of phrases and sentences. In this article, mostly addressed to theoretical linguists curious about distributional semantics,…
GTTM and post-tonal music
A musical grammar, according to Fred Lerdahl's theory (Lerdahl 1988), is not only formed by a system of rules able to produce music, but also by structural aspects allowing listeners to implicitly infer a description of a piece and to build a mental representation of it: a compositional grammar also includes a listening grammar. In much avant-garde music (starting from Schönberg's serial system) a coherence between compositional and listening pro…
Distributional Memory
Research into corpus-based semantics has focused on the development of ad hoc models that treat single tasks, or sets of closely related tasks, as unrelated challenges to be tackled by extracting different kinds of distributional information from the corpus. As an alternative to this "one task, one model" approach, the Distributional Memory framework extracts distributional information once and for all from the corpus, in the form of a set of wei…
The WaCky wide web
A different kind of similarity
Recognising a musical style implies managing a number of similarities. However, the nature of such similarities and their organisation within a piece of music are radically different from those discussed in Forum A4, where themes, patterns or melodies were compared. The perception of a style involves the totality of the structural features present in a given piece or set of pieces. According to the theory put forward in the present article, the a…
Music, musicality, “musilanguage”
The aims of this article are to select, to summarise, and to critically discuss the principal topics presented in the papers of this special issue. Two main themes have been developed: the first makes reference to the origins of music, with a comparison of anthropological conceptions (the origins of mankind) with psychological conceptions (the musical development of infants). Particular attention is given to the “musilanguage” theory proposed by …
Yizhak Sadaï, Traité de sujets musicaux. Vers une épistémologie musicale. Paris
The Macroform in Post-Tonalmusic. Listening and Analysis
Form in classical music is fundamentally a question of organising musical time in order to facilitate a listening corresponding to author's expectations. In the past this was obtained by coordinating different parameters towards a single goal. In twentieth century music, the more detached relationship between the composer and the listener meant that less importance was given to the idea of “correct” listening and to the coordination of parameters…
In praise of Musikwissenschaft
Musical Grammar and the Study of Cognitive Processes of Composition
The concept of “grammar” is at the basis both of listening to and of composing music. Music grammar includes morphology and syntax, the former being defined as the identification of different categories of musical structures, the latter as that of rules connecting morphological unities. In listening to music morphology corresponds to “categorical” perception and syntax to “structural” perception. In composing and in music analysis, the intuitive …
Musical rules and musical signification
Musical signification — according to the main hypothesis of this paper — is based on a syntactic organization of micro-components (minimal units of signification). Each of them contributes to the global meaning of the more extended units of which they are a part. Such a meaning is neither to be intended as a sum of its components, nor as a fixed and well definable entity. On the contrary, it is open to the possibility of several coherent interpre…
Relationships between music and poetry in the arias of Giovanni Legrenzi
This paper presents in detail part of a project set up to analyze the arias of Giovanni Legrenzi. Its aim is to study the relationship between poetry and music. By means of a metric-syntactic tree describing the structure of the poem, it is possible to formulate the rules which govern the form of the arias and their tonal organization. A computer program verifies the completeness and self-consistency of the twenty-eight generative rules used in t…
Round Table VII
The Concept of Musical Grammar
Music analysis and, in particular music criticism, two fundamental tools of historical research and musicology in the broadest sense have become extremely difficult and problematical disciplines in the last few decades. The actual practice of analysis and criticism has by no means come to a halt, but the amount of 'truth' which was once thought obtainable from them or, at least, which was believed to have been won in the actual moment of undertak…
The Signifier and the Signified
Distributional Memory
Research into corpus-based semantics has focused on the development of ad hoc models that treat single tasks, or sets of closely related tasks, as unrelated challenges to be tackled by extracting different kinds of distributional information from the corpus. As an alternative to this "one task, one model" approach, the Distributional Memory framework extracts distributional information once and for all from the corpus, in the form of a set of wei…
The Concept of Musical Grammar
Music analysis and, in particular music criticism, two fundamental tools of historical research and musicology in the broadest sense have become extremely difficult and problematical disciplines in the last few decades. The actual practice of analysis and criticism has by no means come to a halt, but the amount of 'truth' which was once thought obtainable from them or, at least, which was believed to have been won in the actual moment of undertak…
Syntactic Structure from Deep Learning
Modern deep neural networks achieve impressive performance in engineering applications that require extensive linguistic skills, such as machine translation. This success has sparked interest in probing whether these models are inducing human-like grammatical knowledge from the raw data they are exposed to and, consequently, whether they can shed new light on long-standing debates concerning the innate structure necessary for language acquisition…
When the Whole Is Less Than the Sum of Its Parts
Distributional semantic models, deriving vector-based word representations from patterns of word usage in corpora, have many useful applications (Turney and Pantel 2010 ). Recently, there has been interest in compositional distributional models, which derive vectors for phrases from representations of their constituent words (Mitchell and Lapata 2010 ). Often, the values of distributional vectors are pointwise mutual information (PMI) scores obta…
Grounding Distributional Semantics in the Visual World
Distributional semantic models build vector-based word meaning representations on top of contextual information extracted from large collections of text. Object recognition methods from computer vision derive vector-based representations of visual content from natural images. This article reviews how methods from computer vision are exploited to tackle the fundamental problem ofgroundingdistributional semantic models, bringing them closer to prov…
When the Whole Is Not Greater Than the Combination of Its Parts
Distributional semantics has been extended to phrases and sentences by means of composition operations. We look at how these operations affect similarity measurements, showing that similarity equations of an important class of composition methods can be decomposed into operations performed on the subparts of the input phrases. This establishes a strong link between these models and convolution kernels
The Signifier and the Signified
The Concept of Musical Grammar
Music analysis and, in particular music criticism, two fundamental tools of historical research and musicology in the broadest sense have become extremely difficult and problematical disciplines in the last few decades. The actual practice of analysis and criticism has by no means come to a halt, but the amount of 'truth' which was once thought obtainable from them or, at least, which was believed to have been won in the actual moment of undertak…
Round Table VII
Relationships between music and poetry in the arias of Giovanni Legrenzi
This paper presents in detail part of a project set up to analyze the arias of Giovanni Legrenzi. Its aim is to study the relationship between poetry and music. By means of a metric-syntactic tree describing the structure of the poem, it is possible to formulate the rules which govern the form of the arias and their tonal organization. A computer program verifies the completeness and self-consistency of the twenty-eight generative rules used in t…
Musical rules and musical signification
Musical signification — according to the main hypothesis of this paper — is based on a syntactic organization of micro-components (minimal units of signification). Each of them contributes to the global meaning of the more extended units of which they are a part. Such a meaning is neither to be intended as a sum of its components, nor as a fixed and well definable entity. On the contrary, it is open to the possibility of several coherent interpre…
Musical Grammar and the Study of Cognitive Processes of Composition
The concept of “grammar” is at the basis both of listening to and of composing music. Music grammar includes morphology and syntax, the former being defined as the identification of different categories of musical structures, the latter as that of rules connecting morphological unities. In listening to music morphology corresponds to “categorical” perception and syntax to “structural” perception. In composing and in music analysis, the intuitive …
In praise of Musikwissenschaft
The Macroform in Post-Tonalmusic. Listening and Analysis
Form in classical music is fundamentally a question of organising musical time in order to facilitate a listening corresponding to author's expectations. In the past this was obtained by coordinating different parameters towards a single goal. In twentieth century music, the more detached relationship between the composer and the listener meant that less importance was given to the idea of “correct” listening and to the coordination of parameters…
Yizhak Sadaï, Traité de sujets musicaux. Vers une épistémologie musicale. Paris
Music, musicality, “musilanguage”
The aims of this article are to select, to summarise, and to critically discuss the principal topics presented in the papers of this special issue. Two main themes have been developed: the first makes reference to the origins of music, with a comparison of anthropological conceptions (the origins of mankind) with psychological conceptions (the musical development of infants). Particular attention is given to the “musilanguage” theory proposed by …
The WaCky wide web
A different kind of similarity
Recognising a musical style implies managing a number of similarities. However, the nature of such similarities and their organisation within a piece of music are radically different from those discussed in Forum A4, where themes, patterns or melodies were compared. The perception of a style involves the totality of the structural features present in a given piece or set of pieces. According to the theory put forward in the present article, the a…
GTTM and post-tonal music
A musical grammar, according to Fred Lerdahl's theory (Lerdahl 1988), is not only formed by a system of rules able to produce music, but also by structural aspects allowing listeners to implicitly infer a description of a piece and to build a mental representation of it: a compositional grammar also includes a listening grammar. In much avant-garde music (starting from Schönberg's serial system) a coherence between compositional and listening pro…
Distributional Memory
Research into corpus-based semantics has focused on the development of ad hoc models that treat single tasks, or sets of closely related tasks, as unrelated challenges to be tackled by extracting different kinds of distributional information from the corpus. As an alternative to this "one task, one model" approach, the Distributional Memory framework extracts distributional information once and for all from the corpus, in the form of a set of wei…
Composition in Distributional Semantics
Distributional Semantic Models, which automatically induce word meaning representations from naturally occurring textual data, are a success story of computational linguistics. Recently, there has been much interest in whether such models, endowed with a compositional component, can also successfully approximate the meaning of phrases and sentences. In this article, mostly addressed to theoretical linguists curious about distributional semantics,…
When the Whole Is Not Greater Than the Combination of Its Parts
Distributional semantics has been extended to phrases and sentences by means of composition operations. We look at how these operations affect similarity measurements, showing that similarity equations of an important class of composition methods can be decomposed into operations performed on the subparts of the input phrases. This establishes a strong link between these models and convolution kernels
When the Whole Is Less Than the Sum of Its Parts
Distributional semantic models, deriving vector-based word representations from patterns of word usage in corpora, have many useful applications (Turney and Pantel 2010 ). Recently, there has been interest in compositional distributional models, which derive vectors for phrases from representations of their constituent words (Mitchell and Lapata 2010 ). Often, the values of distributional vectors are pointwise mutual information (PMI) scores obta…
There Is No Logical Negation Here, But There Are Alternatives
Logical negation is a challenge for distributional semantics, because predicates and their negations tend to occur in very similar contexts, and consequently their distributional vectors are very similar. Indeed, it is not even clear what properties a “negated” distributional vector should possess. However, when linguistic negation is considered in its actual discourse usage, it often performs a role that is quite different from straightforward l…
Grounding Distributional Semantics in the Visual World
Distributional semantic models build vector-based word meaning representations on top of contextual information extracted from large collections of text. Object recognition methods from computer vision derive vector-based representations of visual content from natural images. This article reviews how methods from computer vision are exploited to tackle the fundamental problem ofgroundingdistributional semantic models, bringing them closer to prov…
Syntactic Structure from Deep Learning
Modern deep neural networks achieve impressive performance in engineering applications that require extensive linguistic skills, such as machine translation. This success has sparked interest in probing whether these models are inducing human-like grammatical knowledge from the raw data they are exposed to and, consequently, whether they can shed new light on long-standing debates concerning the innate structure necessary for language acquisition…
Mechanisms for handling nested dependencies in neural-network language models and humans
Computer Science (15 works) · Linguistics (15 works) · Art (12 works) · Philosophy (11 works) · Natural Language Processing Techniques (9 works) · Artificial Intelligence (8 works) · Musical (8 works) · Psychology (8 works) · Literature (7 works) · Natural language processing (7 works)