16,142 research outputs found
Gonzalez, Ml
No full textCentro Asturiano membership record of Ml. Gonzalez; Socio Number: 15520.https://digitalcommons.usf.edu/asturiano_membership/3322/thumbnail.jp
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
Get PDFThis 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
Dataset for "The State of the ML-universe: 10 Years of Artificial Intelligence & Machine Learning Software Development on GitHub"
No full textSupplementary data to "The State of the ML-universe: 10 Years of Artificial Intelligence & Machine Learning Software Development on GitHub" accepted for publication at MSR 2020.
The data included in this package were used to conduct analyses to characterize the AI & ML software development community hosted on GitHub. Please read the paper for a full understanding of what data was collected and how it was used.
Questions and comments can be directed to Danielle Gonzalez [email protected]</p
Informe Final Semillero de Investigación Automatización de Procesos en ML(APML) 2024-2
No full textInforme de gestion del semillero de investigacion APML durante el semestre 2024-2 en el cual se utilizarón herramientas de ML para la investigación, se desarrollaron pipelines para la automatización del entorno de desarrollo, se realizó la automatizacion de procesos en machine learning y se condujeron experimentos, además se interactuó con bases de datos y se crearon aplicaciones web. Como resultado del trabajo del semillero se realizo un proyecto de investigación con datos sobre el abandono de clientes bancarios. El proyecto esta implementado como un archivo .ipynb y se encuentra disponioble en el aula en Moodle del semillero.2024-
ml-struct-bio/cryodrgn: v3.1.0-b: interactive filtering
No full text<p>We have introduced a number of small fixes and feature updates since our last release <code>v3.0.1-beta</code>:</p>
<ul>
<li>creating a new interactive command-line interface <code>cryodrgn filter</code> as an alternative to the buggy interface in the Jupyter filtering notebook (https://github.com/ml-struct-bio/cryodrgn/issues/323)</li>
<li>making <code>cryodrgn analyze</code> produce a plot of the learning curve (https://github.com/ml-struct-bio/cryodrgn/issues/304)</li>
<li>adding cell in <code>cryoDRGN_filtering</code> jupyter notebook returned by <code>cryodrgn analyze</code> for filtering by UMAP/PC values (https://github.com/ml-struct-bio/cryodrgn/pull/313)</li>
<li>fixing bugs with deprecated signatures in plotting functions (https://github.com/ml-struct-bio/cryodrgn/issues/322) and numpy dependency versioning (https://github.com/ml-struct-bio/cryodrgn/issues/318)</li>
</ul>
Axon-glia interactions regulate ECM patterning in the postnatal rat olfactory bulb
No full textIt has been suggested that an inhibitory ECM containing chondroitin-6- sulfate proteoglycan (C-6S-PG) and tenascin (TN), which appears homogeneously in the core of the OB following afferent fiber arrival, helps position ingrowing olfactory axons in the prospective glomerular layer (GL) (Gonzalez and Silver, 1992; Gonzalez et al., 1993). Later, a similar ECM associated with astrocytes envelopes axonal glomeruli in rings, suggesting that axons may control the precise ECM patterning. The question remains whether formation of the matrix ring pattern around each axonal glomerulus is an intrinsic property of the matrix- producing cells or a response to developing axons. To determine if the organization of glial associated matrix in the OB was dependent on the presence of axons, we studied the effect of unilateral injection of a neurotoxin into the olfactory epithelium of postnatal rats. Using olfactory marker protein (OMP), beta-tubulin (TUJ1) antibodies, and Nissl staining, we found that at 5 and 10 d following neurotoxin administration the number of glomeruli decreased by an average of 77.0% in the injected side. At the same time, we observed that the TN/C-6S-PG rings and periglomerular cells were present only around the remaining small number of glomeruli. Elsewhere, ECM expression and the periglomerular cell configuration were more disorganized in the GL. The pattern of glial fibrillary acidic protein (GFAP) did not change significantly. We found that OMP staining, beta-tubulin immunoreactivity, and periglomerular cells reformed in a glomerular- like pattern as the olfactory axons reformed by 20 d. As the glomeruli- shaped collection of axon terminals reappeared, TN/C-6S-PG immunoreactivity also reoccurred in rings around the new axon bundles. Again, at this later stage, the expression of GFAP was similar in both sides. In our previous study (Gonzalez et al., 1993), we suggested that the initial gross positioning of glomeruli may be controlled by the overall positioning of TN/C-6S-PG. In the present study, we suggest that the formation of TN/C-6S-PG in the precise ring pattern around glomeruli appears to be dependent upon the presence of bundled olfactory axons. Various mechanisms are discussed that may explain the dynamic change in ECM expression that occurs inside the glomerulus after the neurotoxin treatment.</jats:p
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
Low-complexity soft ML detection for generalized spatial modulation
Get PDF[EN] Generalized Spatial Modulation (GSM) is a recent Multiple-Input Multiple-Output (MIMO) scheme, which achieves high spectral and energy efficiencies. Specifically, soft-output detectors have a key role in achiev-ing the highest coding gain when an error-correcting code (ECC) is used. Nowadays, soft-output Maxi-mum Likelihood (ML) detection in MIMO-GSM systems leads to a computational complexity that is un-feasible for real applications; however, it is important to develop low-complexity decoding algorithms that provide a reasonable computational simulation time in order to make a performance benchmark available in MIMO-GSM systems. This paper presents three algorithms that achieve ML performance. In the first algorithm, different strategies are implemented, such as a preprocessing sorting step in order to avoid an exhaustive search. In addition, clipping of the extrinsic log-likelihood ratios (LLRs) can be incor-porating to this algorithm to give a lower cost version. The other two proposed algorithms can only be used with clipping and the results show a significant saving in computational cost. Furthermore clipping allows a wide-trade-off between performance and complexity by only adjusting the clipping parameter.Acknowledgements This work has been partially supported by Spanish Ministry of Science, Innovation and Universities and by European Union through grant RTI2018-098085-BC41 (MCUI/AEI/FEDER) , by GVASimarro, MA.; García Mollá, VM.; Martínez Zaldívar, FJ.; Gonzalez, A. (2022). Low-complexity soft ML detection for generalized spatial modulation. Signal Processing. 196:1-12. https://doi.org/10.1016/j.sigpro.2022.108509S11219
- …
