Miguel Molina-Solana

Inmamusys: Intelligent multiagent music system

Music generation is a complex task even for human beings. This paper describes a two level competitive/collaborative multiagent approach for autonomous, non-deterministic, computer music composition. Our aim is to build a high modular system that composes music on its own by using Experts Systems technology and rule-based systems principles. To do that, rules issued from musical knowledge are used and emotional inputs from the users are introduced. In fact, users are not allowed to directly control the composition process. Two main goals are sought after: investigating relationships between computers and emotions and how the latter can be represented into the former, and developing a framework for music composition that can be useful for future experiments. The system has been successfully tested by asking several people to match compositions with suggested emotions.

Visualizing Dynamic Bitcoin Transaction Patterns.

Abstract This work presents a systemic top-down visualization of Bitcoin transaction activity to explore dynamically generated patterns of algorithmic behavior. Bitcoin dominates the cryptocurrency markets and presents researchers with a rich source of real-time transactional data. The pseudonymous yet public nature of the data presents opportunities for the discovery of human and algorithmic behavioral patterns of interest to many parties such as financial regulators, protocol designers, and security analysts. However, retaining visual fidelity to the underlying data to retain a fuller understanding of activity within the network remains challenging, particularly in real time. We expose an effective force-directed graph visualization employed in our large-scale data observation facility to accelerate this data exploration and derive useful insight among domain experts and the general public alike. The high-fidelity visualizations demonstrated in this article allowed for collaborative discovery of unexpected high frequency transaction patterns, including automated laundering operations, and the evolution of multiple distinct algorithmic denial of service attacks on the Bitcoin network.

A state of the art on computational music performance

Musical expressivity can be defined as the deviation from a musical standard when a score is performed by a musician. This deviation is made in terms of intrinsic note attributes like pitch, timbre, timing and dynamics. The advances in computational power capabilities and digital sound synthesis have allowed real-time control of synthesized sounds. Expressive control becomes then an area of great interest in the sound and music computing field. Musical expressivity can be approached from different perspectives. One approach is the musicological analysis of music and the study of the different stylistic schools. This approach provides a valuable understanding about musical expressivity. Another perspective is the computational modelling of music performance by means of automatic analysis of recordings. It is known that music performance is a complex activity that involves complementary aspects from other disciplines such as psychology and acoustics. It requires creativity and eventually, some manual abilities, being a hard task even for humans. Therefore, using machines appears as a very interesting and fascinating issue. In this paper, we present an overall view of the works many researchers have done so far in the field of expressive music performance, with special attention to the computational approach.

Mining transposed motifs in music

The discovery of frequent musical patterns (motifs) is a relevant problem in musicology. This paper introduces an unsupervised algorithm to address this problem in symbolically-represented musical melodies. Our algorithm is able to identify transposed patterns including exact matchings, i.e., null transpositions. We have tested our algorithm on a corpus of songs and the results suggest that our approach is promising, specially when dealing with songs that include non-exact repetitions.

Identifying violin performers by their expressive trends

Understanding the way performers use expressive resources of a given instrument to communicate with the audience is a challenging problem in the sound and music computing field. Working directly with commercial recordings is a good opportunity for tackling this implicit knowledge and studying well-known performers. The huge amount of information to be analyzed suggests the use of automatic techniques, which have to deal with imprecise analysis and manage the information in a broader perspective. This work presents a new approach, Trend-based modeling, for identifying professional performers in commercial recordings. Concretely, starting from automatically extracted descriptors provided by state-of-the-art tools, our approach performs a qualitative analysis of the detected trends for a given set of melodic patterns. The feasibility of our approach is shown for a dataset of monophonic violin recordings from 23 well-known performers.

Improving data exploration in graphs with fuzzy logic and large-scale visualisation

Graphical abstractDisplay Omitted HighlightsBetter graph-sensemaking through fuzzy queries and large-scale visualisation.Fuzzy logic to identify interesting nodes.Illustrative examples of such data exploration provided. This work presents three case-studies of how fuzzy logic can be combined with large-scale immersive visualisation to enhance the process of graph sensemaking, enabling interactive fuzzy filtering of large global views of graphs. The aim is to provide users a mechanism to quickly identify interesting nodes for further analysis. Fuzzy logic allows a flexible framework to ask human-like curiosity-driven questions over the data, and visualisation allows its communication and understanding. Together, these two technologies successfully empower novices and experts to a faster and deeper understanding of the underlying patterns in big datasets compared to traditional means in a desktop screen with crisp queries. Among other examples, we provide evidence of how these two technologies successfully enable the identification of relevant transaction patterns in the Bitcoin network.

Meta-association rules for mining interesting associations in multiple datasets

Graphical abstractProcess flow: from original datasets to final meta-association rules. The process starts from a set of databases {D1, , Dk} which share some of their content, i.e. they have attributes in common. After applying a rule extraction procedure, we obtain k sets of association rules represented by Ri. We are interested in searching associations between the already extracted rules in the sets Ri. For achieving this, we create a meta-database D collecting the information. We propose two different ways, by considering only presence-absence of rules (crisp meta-database, D) or taking into account their reliability represented by a degree in the unit interval (fuzzy meta-database, D). We can also introduce additional information into the process by adding new attributes about the original datasets Di. After this the so-called meta-association rules are mined. Examples of meta-association rules are depicted in the right part of the figure relating the primary rules and the added attributes. The paper explains and compares both proposals (crisp and fuzzy), proposes a level-based mining algorithm using the RL-theory for the representation of fuzziness and makes some experimentation with synthetic and real data. Display Omitted HighlightsMeta-association rules are extracted from regular rules and contextual information.Meta-rules discover which associations are more frequent in multiple datasets.Fuzzy and non-fuzzy approaches are described.The algorithm proposed avoids information loss and supports data imprecision.Fuzzy algorithm allows parameters fine-tuning with acceptable execution time. Association rules have been widely used in many application areas to extract new and useful information expressed in a comprehensive way for decision makers from raw data. However, raw data may not always be available, it can be distributed in multiple datasets and therefore there resulting number of association rules to be inspected is overwhelming. In the light of these observations, we propose meta-association rules, a new framework for mining association rules over previously discovered rules in multiple databases. Meta-association rules are a new tool that convey new information from the patterns extracted from multiple datasets and give a summarized representation about most frequent patterns. We propose and compare two different algorithms based respectively on crisp rules and fuzzy rules, concluding that fuzzy meta-association rules are suitable to incorporate to the meta-mining procedure the obtained quality assessment provided by the rules in the first step of the process, although it consumes more time than the crisp approach. In addition, fuzzy meta-rules give a more manageable set of rules for its posterior analysis and they allow the use of fuzzy items to express additional knowledge about the original databases. The proposed framework is illustrated with real-life data about crime incidents in the city of Chicago. Issues such as the difference with traditional approaches are discussed using synthetic data.

Representation model and learning algorithm for uncertain and imprecise multivariate behaviors, based on correlated trends

Graphical abstractDisplay Omitted HighlightsRepresentation model for uncertain and imprecise multivariate dataseries.Basic idea: finding repeating frequent correlated patterns among the different dimensions of the dataseries.Deal with data imperfection directly, not transforming the data and pretending it has no imperfection.Applicable to several problems that can be represented by series of observations.Provide a fix size representation, regardless of the length of the dataseries. The computational representation and classification of behaviors is a task of growing interest in the field of Behavior Informatics, being series of data a common way of describing those behaviors. However, as these data are often imperfect, new representation models are required in order to effectively handle imperfection in this context. This work presents a new approach, Frequent Correlated Trends, for representing uncertain and imprecise multivariate data series. Such a model can be applied to any domain where behaviors recur in similar-but not identical-shape. In particular, we have already used them to the task of identifying the performers of violin recordings with good results. The present paper describes the abstract model representation and a general learning algorithm, and discusses several potential applications.

Unifying fuzzy controller for indoor environment quality

Optimizing energy consumption while maintaining an appropriate level of comfort is one of the current challenges in the Indoor Environment Quality (IEQ) field. In this paper, we propose an unified fuzzy controller for managing the different aspects involved in IEQ, overcoming the potentially inefficient interactions between several traditional controllers. We also describe the implemented web simulator to test the controller, and the results of applying the controller to a pilot room, measuring the air quality over a period of a month. Even though, further research and experimentation should be done, specially with the experts in order to fine tuning the rules and the controller, initial results are promising enough and the ease of use is quite remarkable by users.

Transcribing Debussy's Syrinx dynamics through Linguistic Description

Advances in computational power have enabled the manipulation of music audio files and the emergence of the Music Information Retrieval field. One of the main research lines in this area is that of Music Transcription, which aims at transforming an audio file into musical notation. So far, most efforts have focused on accurately transcribing the pitch and durations of the notes, and thus neglecting other aspects in the music score. The present work explores a novel line of action in the context of automatic music transcription, focusing on the dynamics, and by means of Linguistic Description. The process described in this paper (called MUDELD: MUsic Dynamics Extraction through Linguistic Description) departs from the data series representing the audio file, and requires the segmentation of the piece in phrases, which is currently done by hand. Initial experiments have been performed on eight recordings of Debussys Syrinx with promising results.