Explorations in the distributional and semantic similarity of words

A straightforward computation of the list of the words (the `tail words' of the list) that are distributionally most similar to a given word (the `head word' of the list) leads to the question: How semantically similar to the head word are the tail words; that is: how similar are their meanings to its meaning? And can we do better? The experiment was done on nearly 18,000 most frequent nouns in a Finnish newsgroup corpus. These nouns are considered to be distributionally similar to the extent that they occur in the same direct dependency relations with the same nouns, adjectives and verbs. The extent of the similarity of their computational representations is quantified with the information radius. The semantic classification of head-tail pairs is intuitive; some tail words seem to be semantically similar to the head word, some do not. Each such pair is also associated with a number of further distributional variables. Individually, their overlap for the semantic classes is large, but the trained classification-tree models have some success in using combinations to predict the semantic class. The training data consists of a random sample of 400 head-tail pairs with the tail word ranked among the 20 distributionally most similar to the head word, excluding names. The models are then tested on a random sample of another 100 such pairs. The best success rates range from 70% to 92% of the test pairs, where a success means that the model predicted my intuitive semantic class of the pair. This seems somewhat promising when distributional similarity is used to capture semantically similar words. This analysis also includes a general discussion of several different similarity formulas, arranged in three groups: those that apply to sets with graded membership, those that apply to the members of a vector space, and those that apply to probability mass functions.

An exact similarity solution in radiation-gas-dynamics

We have found an exact similarity solution of the point explosion problem in the case when the total energy of the shock wave that is produced is not constant but decreases with time and when the loss due to radiation escape is significant. We have compared the results of our exact solution with those of exact numerical solutions of Elliot and Wang and have explained the cause why our solution differs from theirs in certain aspects.

Extended self-similarity works for the Burgers equation and why

Extended self-similarity (ESS), a procedure that remarkably extends the range of scaling for structure functions in Navier-Stokes turbulence and thus allows improved determination of intermittency exponents, has never been fully explained. We show that ESS applies to Burgers turbulence at high Reynolds numbers and we give the theoretical explanation of the numerically observed improved scaling at both the IR and UV end, in total a gain of about three quarters of a decade: there is a reduction of subdominant contributions to scaling when going from the standard structure function representation to the ESS representation. We conjecture that a similar situation holds for three-dimensional incompressible turbulence and suggest ways of capturing subdominant contributions to scaling.

Traditional patterns and textures as values for meaningful automatization in music

This contribution focuses on the accelerated loss of traditional sound patterning in music, parallel to the exponential loss of linguistic and cultural variety in a world increasingly 'globalized' by market policies and economic liberalization, in which scientific or technical justification plays a crucial role. As a suggestion to an alternative trend, composers and music theorists are invited to explore the world of design and patterning by grammar rules from non-dominant cultures, and to make an effort to understand their contextual usage and its transformation, in order to appreciate their symbolism and aesthetic depth. Practical examples are provided.

The Role of Abduction in Self-Similarity

Based on the Aristotelian criterion referred to as 'abductio', Peirce suggests a method of hypothetical inference, which operates in a different way than the deductive and inductive methods. “Abduction is nothing but guessing” (Peirce, 7.219). This principle is of extreme value for the study of our understanding of mathematical self-similarity in both of its typical presentations: relative or absolute. For the first case, abduction incarnates the quantitative/qualitative relationships of a self-similar object or process; for the second case, abduction makes understandable the statistical treatment of self-similarity, 'guessing' the continuity of geometric features to the infinity through the use of a systematic stereotype (for instance, the assumption that the general shape of the Sierpiński triangle continuates identically into its particular shapes). The metaphor coined by Peirce, of an exact map containig itself the same exact map (a map of itself), is not only the most important precedent of Mandelbrot’s problem of measuring the boundaries of a continuous irregular surface with a logarithmic ruler, but also still being a useful abstraction for the conceptualisation of relative and absolute self-similarity, and its mechanisms of implementation. It is useful, also, for explaining some of the most basic geometric ontologies as mental constructions: in the notion of infinite convergence of points in the corners of a triangle, or the intuition for defining two parallel straight lines as two lines in a plane that 'never' intersect.

Music in the recovering brain

Listening to music involves a widely distributed bilateral network of brain regions that controls many auditory perceptual, cognitive, emotional, and motor functions. Exposure to music can also temporarily improve mood, reduce stress, and enhance cognitive performance as well as promote neural plasticity. However, very little is currently known about the relationship between music perception and auditory and cognitive processes or about the potential therapeutic effects of listening to music after neural damage. This thesis explores the interplay of auditory, cognitive, and emotional factors related to music processing after a middle cerebral artery (MCA) stroke. In the acute recovery phase, 60 MCA stroke patients were randomly assigned to a music listening group, an audio book listening group, or a control group. All patients underwent neuropsychological assessments, magnetoencephalography (MEG) measurements, and magnetic resonance imaging (MRI) scans repeatedly during a six-month post-stroke period. The results revealed that amusia, a deficit of music perception, is a common and persistent deficit after a stroke, especially if the stroke affects the frontal and temporal brain areas in the right hemisphere. Amusia is clearly associated with deficits in both auditory encoding, as indicated by the magnetic mismatch negativity (MMNm) response, and domain-general cognitive processes, such as attention, working memory, and executive functions. Furthermore, both music and audio book listening increased the MMNm, whereas only music listening improved the recovery of verbal memory and focused attention as well as prevented a depressed and confused mood during the first post-stroke months. These findings indicate a close link between musical, auditory, and cognitive processes in the brain. Importantly, they also encourage the use of listening to music as a rehabilitative leisure activity after a stroke and suggest that the auditory environment can induce long-term plastic changes in the recovering brain.

The Ecologic Foundations of Stylistics in Music and in Language

This contribution suggests that it is possible to describe the transformations of musical style in an analogous way to the transformations of style in language, and also that it can be explained how the ‘musics in contact’ behave in an analogous way to the ‘languages in contact’. According to this idea, the ‘evolution’ of styles in music and in language can be identified and studied as dynamic exchanges in ecological niches. It is suggested, also, that the idiolectic-ecolectic, and acrolectic-basilectic relationships in music and language are functions of cycles in several ‘layers’ and ‘rhythms’. The presence of stylistic varieties and influences in music and in language may imply that they are part of major sign systems within a more complex ecological relationship.

A speech-music discriminator using HILN model based features

We propose a simple speech music discriminator that uses features based on HILN(Harmonics, Individual Lines and Noise) model. We have been able to test the strength of the feature set on a standard database of 66 files and get an accuracy of around 97%. We also have tested on sung queries and polyphonic music and have got very good results. The current algorithm is being used to discriminate between sung queries and played (using an instrument like flute) queries for a Query by Humming(QBH) system currently under development in the lab.