9,959 research outputs found
Metadata Representations for Queryable ML Model Zoos
No full textMachine learning (ML) practitioners and organizations are building model zoos of pre-trained models, containing metadata describing properties of the ML models and datasets that are useful for reporting, auditing, reproducibility, and interpretability purposes. The metatada is currently not standardised; its expressivity is limited; and there is no interoperable way to store and query it. Consequently, model search, reuse, comparison, and composition are hindered. In this paper, we advocate for standardized ML model metadata representation and management, proposing a toolkit supported to help practitioners manage and query that metadata.Web Information SystemsHuman-Centred Artificial Intelligenc
A Manifesto of Nodalism
No full textThis paper proposes the notion of Nodalism as a means describing contemporary culture and of understanding my own creative practice in electronic music composition. It draws on theories and ideas from Kirby, Bauman, Bourriaud, Deleuze, Guatarri, and Gochenour, to demonstrate how networks of ideas or connectionist neural models of cognitive behaviour can be used to contextualize, understand and become a creative tool for the creation of contemporary electronic music
Optimizing ML Inference Queries Under Constraints
No full textThe proliferation of pre-trained ML models in public Web-based model zoos facilitates the engineering of ML pipelines to address complex inference queries over datasets and streams of unstructured content. Constructing optimal plan for a query is hard, especially when constraints (e.g. accuracy or execution time) must be taken into consideration, and the complexity of the inference query increases. To address this issue, we propose a method for optimizing ML inference queries that selects the most suitable ML models to use, as well as the order in which those models are executed. We formally define the constraint-based ML inference query optimization problem, formulate it as a Mixed Integer Programming (MIP) problem, and develop an optimizer that maximizes accuracy given constraints. This optimizer is capable of navigating a large search space to identify optimal query plans on various model zoos.Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Web Information SystemsHuman-Centred Artificial Intelligenc
Monitoramento e modelagem do processo de interceptação da chuva de uma bacia coberta por floresta ombrófila mista
No full textDissertação (mestrado) - Universidade Federal de Santa Catarina, Centro Tecnológico, Programa de Pós-Graduação em Engenharia Ambiental, Florianópolis, 2015.A interceptação é o processo pelo qual a precipitação que cai sobre a superfície do terreno fica retida, é redistribuída ao solo ou evapora posteriormente. Esse processo é importante especialmente em áreas florestais, pois influencia na reciclagem da umidade do ar e também na quantidade de água que efetivamente chega ao solo. A maioria dos estudos de interceptação nas florestas brasileiras foi feita na região Amazônica e Mata Atlântica, mas poucos foram os estudos realizados na Mata Atlântica do tipo Floresta Ombrófila Mista. O objetivo deste trabalho foi compreender e estimar os processos de interceptação da chuva em uma bacia coberta por Floresta Ombrófila Mista por meio de monitoramento e modelagem. A perda por interceptação (I) foi estimada por meio da medição de chuva externa (P), chuva interna (Tf) e escoamento pelo tronco (Sf). O monitoramento da chuva externa foi feito com três pluviômetros e um pluviógrafo instalados fora da bacia. Nove pluviômetros e uma calha interligada a um pluviógrafo foram instalados para o monitoramento da chuva interna. O monitoramento do escoamento pelo tronco foi realizado em dez árvores interligadas a um recipiente de armazenamento e mais quatro árvores interligadas a um pluviógrafo. Os índices de cobertura do dossel foram estimados a partir de fotografias do dossel. Os dados hidrometeorológicos foram monitorados na estação meteorológica do Rio Feio. A chuva externa nos três anos de monitoramento somou 5.309 mm distribuídos em 321 eventos e houve 489 dias sem ocorrência de chuva. Os modelos de Rutter esparso e de Gash esparso foram utilizados para a modelagem do processo de interceptação. A chuva interna variou de 61 a 83% da chuva externa, o que demonstrou que a chuva interna é heterogênea na bacia coberta por Floresta Ombrófila Mista. O volume de escoamento de tronco variou de 0,1 a 22,9 litros, com média de 2,92 litros e no período monitorado os volumes tenderam a crescer com o aumento do diâmetro do tronco e da área da copa. O índice de cobertura do dossel na bacia variou de 45 a 94%, com média de 80% e não apresentou correlação com a distribuição da chuva interna. Os dados automatizados da chuva interna e escoamento pelo tronco foram distribuídos em 60 eventos, a chuva externa foi 1.303 mm, a chuva interna e escoamento pelo tronco corresponderam a 75% (981 mm) e 2 % (19 mm) da chuva externa. A média da evapotranspiração potencial da série calculada foi de 1,09 mm dia-1. O erro relativo por evento dos modelos de interceptação variou de 2 a 400% para o modelo de Rutter e de 0 a 350% para o modelo de Gash. O modelo de Gash subestimou a I acumulada enquanto que o modelo de Rutter a superestimou. A simulação com o modelo de Rutter foi a que obteve os menores valores do erro relativo (4%) em relação a I acumulada. Os dois modelos tiveram limitações para representar o processo de interceptação em eventos extremos, possivelmente devido aos erros associados à estrutura dos modelos assim como ao conjunto de dados medidos e estimados.Abstract :The interception is the process by which the precipitation that falls on the ground surface is retained, it is redistributed to the ground or evaporate later. This process is important especially in forested areas since it influences recycling of moisture from the air and also the amount of water that effectively reaches the ground. Most of the interception studies in Brazilian Forests were carried out in the Amazon and Atlantic Forest, but few have been conducted in the Atlantic Forest Mixed Ombrophilous Forest type. The objective of this work was to understand and estimate the process of interception of rain in an experimental catchment covered by Mixed Ombrophilous Forest through monitoring and modeling. The interception loss will be estimated by measuring rainfall (P), throughfall (Tf) and stemflow (Sf). The monitoring consists of three rain gauges and automatic rain gauge installed outside the basin for rainfall monitoring. Nine gauges and one trough-type collectors coupled to a rain gauge for throughfall monitoring. The stemflow monitoring was conducted in ten trees connected to a storage container and four trees connected to a rain gauge. The canopy cover fraction was estimated using photographs taken of the canopy. Hydrometeorological data were acquired with a meteorological station of Rio Feio. The Rutter and Gash models were used to model the interception process. The total rainfall monitoring of the complete 3 years was 5.309 mm, from 321 rainfall events and in 489 days there was no rainfall. The throughfall ranged from 61 to 83% the rainfall, which showed the throughfall heterogeneity. The stemflow ranged from 0,1 to 22,9 liters, averaging 2,92 liters and the volumes tend to increase with the diameters at breast height and crown area increases. The canopy cover fraction ranged from 45 to 94 %, averaging 80%. The calculation of the interception loss using automated date rail and stemflow were distributed in 60 events. The rainfall (P) was 1303 mm, Tf and Sf corresponded to 75% (981 mm) and 2% (19 mm) of P. The mean potential evapotranspiration was 1.09 mm day-1. The relative error in the models for the events ranged from 2 to 400% for the Rutter model and 0 to 350% for the Gash model. The Gash model underestimated I while the simulation of the Rutter model overestimated. The simulation with the Rutter model was the one with the lowest values of relative error (4%) for I accumulated. Both models have limitations to identify the process of interception
Building a generalisable ML pipeline at ING
No full textAdvances in data science have caused an increase in the use of Artificial Intelligence (AI), specifically Machine Learning (ML), throughout various fields. Not only in research but in the industry as well, has ML been receiving increasing amounts of interest. Many companies rely on ML models to increase the efficiency of existing processes or offer new services and products. The industry, however, is facing several additional challenges compared to the academic context. One of those challenges is applying the Development Operations (DevOps) model to an ML application, also referred to as MLOps. This thesis sets out to find the specific challenges that practitioners encounter while operationalising ML models. To do so, we perform a single-case case study on an ML pipeline built by the Trade & Communication Surveillance team at the ING bank. This case study consists of conducting a set of interviews and performing a manual code inspection of the pipeline. The team faces challenges ranging from having insufficient time for operationalising each ML project individually to operating in the highlyregulated fintech context. Their pipeline is able to deploy a single ML model but it does not generalise well to other projects. We present the first version of an application that mitigates these challenges. The application is able to deploy ML models to the development environment at ING and can be operated by data scientists to reduce the effort of operationalising an ML model. Computer Science | Software Technolog
'Project smells' - Experiences in Analysing the Software Quality of ML Projects with mllint
No full textMachine Learning (ML) projects incur novel challenges in their development and productionisation over traditional software applications, though established principles and best practices in ensuring the project's software quality still apply. While using static analysis to catch code smells has been shown to improve software quality attributes, it is only a small piece of the software quality puzzle, especially in the case of ML projects given their additional challenges and lower degree of Software Engineering (SE) experience in the data scientists that develop them. We introduce the novel concept of project smells which consider deficits in project management as a more holistic perspective on software quality in ML projects. An open-source static analysis tool mllint was also implemented to help detect and mitigate these. Our research evaluates this novel concept of project smells in the industrial context of ING, a global bank and large software- and data-intensive organisation. We also investigate the perceived importance of these project smells for proof-of-concept versus production-ready ML projects, as well as the perceived obstructions and benefits to using static analysis tools such as mllint. Our findings indicate a need for context-aware static analysis tools, that fit the needs of the project at its current stage of development, while requiring minimal configuration effort from the user. Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Software EngineeringSoftware Technolog
Audiomobiles, Sculptures and Conundrums
No full textRoberto Gerhard was a pioneer of electronic music in England creating a number of substantial concert, theatre and radio works from as early as 1954. Gerhard’s electronic music is one of the richest repositories for understanding the development of the composer’s late compositional technique. Apart from the Symphony no.3, ‘Collages’, none of Gerhard’s electronic music is published. This paper will discuss aspects of Gerhard’s electronic music, focusing on Audiomobiles (1958-59) and Sculptures (1963)
Music for classical guitar by South African composers : a historical survey, notes on selected works and a general catalogue
No full textIncludes abstract.Includes bibliographical references (leaves 296-309).This is the first comprehensive investigation of music for, or including, the classical guitar by South African composers. The focus of this research has been, firstly, to uncover as much of the repertoire as possible, and, secondly, to collate, study, catalogue and report on the information. A brief historical survey of the guitar in South Africa provides the context within which this study was conducted. The primary sources of quantitative data collection were through the archival catalogues of the South African Music Rights Organisation and through personal contact with guitarists, composers and guitar teachers. Other sources consulted were publishers, broadcasting corporations, recording companies, libraries and the internet. The body of the dissertation comprises biographical sketches, background notes, analyses and technical notes on 17 selected solo and chamber works dating from 1947 to 2007 by some of South Africa's most prominent composers and guitaristcomposers. The repertoire ranges in style from the traditional and ethnically inspired to the experimental and abstract. As this is an empirical survey, each selected entry includes details on instrumentation, duration, level of difficulty, number of pages, scordatura, commissions or requests, sources or publishers, premières and recordings. A biography of each composer is provided as well as background notes which offer an overview of the selected work. The notes discuss historical, cultural, musical and extra-musical influences, and frequently include references to interview material. The commentaries on the selected works, with musical examples, include an analytical component describing structure, form, stylistic and compositional elements, while the technical observations include performance suggestions and a grading for each work
Assessing Robustness of ML-Based Program Analysis Tools using Metamorphic Program Transformations
No full textMetamorphic testing is a well-established testing technique that has been successfully applied in various domains, including testing deep learning models to assess their robustness against data noise or malicious input. Currently, metamorphic testing approaches for machine learning (ML) models focused on image processing and object recognition tasks. Hence, these approaches cannot be ap- plied to ML targeting program analysis tasks. In this paper, we extend metamorphic testing approaches for ML models targeting software programs. We present Lampion, a novel testing frame- work that applies (semantics preserving) metamorphic transforma- tions on the test datasets. Lampion produces new code snippets equivalent to the original test set but different in their identifiers or syntactic structure. We evaluate Lampion against CodeBERT, a state-of-the-art ML model for Code-To-Text tasks that creates Javadoc summaries for given Java methods. Our results show that simple transformations significantly impact the target model be- havior, providing additional information on the models reasoning apart from the classic performance metric.Software EngineeringSoftware Technolog
Systematic reviews with language restrictions and no author contact have lower overall credibility: a methodology study
No full textZhen Wang,1–3 Juan P Brito,4 Apostolos Tsapas,5 Marcio L Griebeler,4 Fares Alahdab,1,3 Mohammad Hassan Murad,1,3,61Robert D and Patricia E Kern Center for the Science of Health Care Delivery, 2Division of Health Care Policy and Research, Department of Health Sciences Research, 3Knowledge and Evaluation Research Unit, 4Division of Endocrinology, Diabetes, Metabolism, and Nutrition, Mayo Clinic, Rochester, MN, USA; 5Aristotle University of Thessaloniki, Thessaloniki, Greece; 6Division of Preventive, Occupational and Aerospace Medicine, Mayo Clinic, Rochester, MN, USABackground: High-quality systematic reviews (SRs) require rigorous approaches to identify, appraise, select, and synthesize research evidence relevant to a specific question. In this study, we evaluated the association between two steps in the conduct of an SR – restricting the search to English, and author contact for missing data – and the overall credibility of a SR.Methods: All SRs cited by the Endocrine Society's Clinical Practice Guidelines published from October 2006 through January 2012 were included. The main outcome was the overall A Measurement Tool to Assess Systematic Reviews (AMSTAR) score, as a surrogate of SR credibility. Nonparametric Kruskal–Wallis tests and multivariable linear regression models were used to investigate the association between language restriction, author contact for missing data, and the overall AMSTAR score.Results: In all, 69 SRs were included in the analysis. Only 31 SRs (45%) reported searching non-English literature, with an average AMSTAR score of 7.90 (standard deviation [SD] =1.64). SRs that reported language restriction received significantly lower AMSTAR scores (mean =5.25, SD =2.32) (P<0.001). Only 30 SRs (43%) reported contacting authors for missing data, and these received, on average, 2.59 more AMSTAR points (SD =1.95) than those who did not (P<0.001). In multivariable analyses, AMSTAR score was significantly correlated with language restriction (beta =-1.31, 95% confidence interval [CI]: -2.62, -0.01, P=0.05) and author contact for missing data (beta =2.16, 95% CI: 0.91, 3.41, P=0.001). However, after adjusting for compliance with reporting guidelines, language restriction was no longer significantly associated with the AMSTAR score.Conclusion: Fewer than half of the SRs conducted to support the clinical practice guidelines we examined reported contacting study authors or searched non–English literature. SRs that did not conduct these two steps had lower quality scores, suggesting the importance of these two steps for overall SR credibility.Keywords: evidence-based medicine, research design, validity, quality of evidenc
- …
