Search results

We describe how to utilize the possibility of differentiating through arbitrary Julia codeto perform tasks such as controller optimization. The user specifies a cost function, forexample, the integrated squared error between output and reference, and constraints, suchas a maximum acceptable value of the sensitivity function. Julia performs the integrationand calculates the sensitivities of the cos

Savannas are inherently “disturbed” ecosystems, but the regularly recurring disruptions play such a fundamental ecological role (Scholes and Archer, 1997) that “episodic events” rather than “disturbance” may the more apt terminology. From an atmospheric perspective, fire is the most significant of these episodic events. Fires shape community species composition; tree to grass ratio and nutrient re

Data intense defined software is becoming more and more prevalent, especially with the advent of machine learning and artificial intelligence. With data intense systems comes both challenges – to continue to collect and maintain quality – and opportunities – open innovation by sharing with others.To understand challenges and opportunities with ODC, we ran 5 focus groups (4 in Lund and 1 in Kista)

Changes in the characteristics of cyclone activity (frequency, depth and size) in the Arctic are analyzed based on simulations with state-of-the-art regional climate models (RCMs) from the Arctic-CORDEX initiative and global climate models (GCMs) from CMIP5 under the Representative Concentration Pathway (RCP) 8.5 scenario. Most of RCMs show an increase of cyclone frequency in winter (DJF) and a de

The importance of user-centric design methods in the design of programming tools is now well accepted. These methods depend on creating a feedback loop between the designers and their users, providing data about developers, their needs and behaviour gathered through various means. These include controlled experiments, field observations, as well as analytical frameworks. However, whilst there have

We propose a new structure for the variational auto-encoders (VAEs) prior, with the weakly informative multivariate Student’s t-distribution. In the proposed model all distribution parameters are trained, thereby allowing for a more robust approximation of the underlying data distribution. We used Fashion-MNIST data in two experiments to compare the proposed VAEs with the standard Gaussian priors.

This paper presents an object-based, global climatology (1979-2014) of strong large-scale ocean evaporation (SLOE) and its associated climatic properties. SLOE is diagnosed using an ''atmospheric moisture uptake efficiency'' criterion related to the ratio of surface evaporation and integrated water vapor content in the near-surface atmosphere. The chosen Eulerian identification procedure focuses o

Fifth generation cellular systems will be deployed in the microwave and millimeter–wave (mmWave) frequency bands (i.e., between 0.5–100 GHz). Propagation characteristics at these bands have a fundamental impact on each aspect of the cellular architecture, ranging from equipment design to real-time performance in the field. While we have a reasonable understanding of the propagation characteristics

The recent explosion in size and complexity of datasets and the increased availability of computational resources has led us to what is sometimes called the big data era. In many big data fields, mathematical optimization has over the last decade emerged as a vital tool in extracting information from the data sets and creating predictors for unseen data. The large dimension of these data sets and

We present an accurate analysis of phase noise in ring oscillators, where each stage in the ring is a CMOS inverter loaded by a capacitance. Closed-form phase noise expressions are obtained in both 1/f2 and 1/f3 offset frequency regions, displaying an excellent agreement with numerical simulations using either ideal Verilog-A MOS models (allowing theoretical predictions to be tested with exactly t

In this thesis, we draw inspiration from both classical system identification and modern machine learning in order to solve estimation problems for real-world, physical systems. The main approach to estimation and learning adopted is optimization based. Concepts such as regularization will be utilized for encoding of prior knowledge and basis-function expansions will be used to add nonlinear model

Programming languages are ever evolving, with new languages being invented to solve new problems, and old languages being extended to solve old problems in new ways. With the continued evolution of programming languages, and with new and improved static program analyses, we need flexible systems for building our static analyses and compilers. Declarative programming with Reference Attribute Gramma

Friction stir welding (FSW) is a solid-state welding process where metals are joined without melting. Heat is generated by friction between a rotating non-consumable tool and the work-piece material, and by mechanical deformation of the material. The process, invented in 1991, provides several benefits over arc welding and other fusion processes: No filler material has to be added, low energy cons

In this chapter, we discuss the open challenges in building self-aware computing systems that are still being faced by the research and development community. The challenges can be theoretical, technical, computational, or even sociological. First, we highlight the challenges associated with each of the earlier parts of the book and summarize on respective future research directions. We then offer

Knowing whether a software feature will be completed in its planned iteration can help with release planning decisions. However, existing research has focused on predictions of only low-level software tasks, like bug fixes. In this paper, we describe a mixed-method empirical study on three large IBM projects. We investigated the types of iteration changes that occur. We show that up to 54% of high

In networked control systems it is desirable to have efficient wireless communication (saving energy and bandwidth) while still ensuring good control performance. By abandoning periodic sampling, communication can be made more efficient by sampling and updating the control signal only "when required" based on the system’s behaviour. This is the concept of event-based control. In this work we consi

Nowadays, we live in a society with billions of devices that are interconnected and interact together to improve the quality of our lives. The management and processing of information and knowledge have by now become our main resources, and the fundamental factors of economic and social development, and it is achieved through Big Data Frameworks (BDFs). The amount of such data is becoming larger e

The “wooded-meadows system” is a multifunctional use of vegetation resources widespread in Europesince the Neolithic, and well documented in the Ligurian Apennines (NW Italy) between the MiddleAges and the first half of the 19th century. The management of wooded-meadows included: collectionof fallen and dead branches in spring, later used for fuel; mowing and grazing in summer; collection ofsecond

Datalog has emerged as a powerful tool for expressing staticprogram analyses. Program analysis researchers have builtnontrivial code bases in Datalog, but tool support for workingwith Datalog itself has been lacking. In this paper, weintroduce MetaDL, a language extension to Datalog thatenables source-level Datalog program analysis within Datalog.We describe several program analyses implemented in