Sökresultat

Filtyp

Din sökning på "*" gav 550233 sökträffar

Optimizing L2-regularization for Binary Classification Tasks

An Artificial Neural Network (ANN) is a type of machine learning algorithm with widespread usage. When training an ANN, there is a risk that it gets overtrained and cannot solve the task for new data. Methods to prevent this, such as L2-regularization, introduce hyperparameters that are time-consuming to optimize. In this thesis, I investigate a hypothesis which postulates how the optimal L2-regul

Transforming Clinical Decision Support with AI

Machine learning, alongside other branches of AI have entered an era of rapid growth in several industries including healthcare, creating vast opportunities to advance several aspects of clinical practice. Specific to clinical decision support, despite the immense potential benefits of incorporating machine learning into clinical decision support systems, also known as machine learning enabled cli

Gunshot Detection from Audio Streams in Portable Devices

Machine learning and artificial neural networks can be used to classify or detect specific sound events in audio signals. Gunshot detection is one use case for such networks and can be used to help law enforcement by alerting officers or triggering camera recordings. However, artificial neural networks with a high performance usually require large amounts of computational power, meaning that they

Material flow improvements - An investigation of the picking process at Scania

Problem description: At Scania in Oskarshamn, there is a phenomenon referred to as external picking. External picking is in the context of this report defined as the process where stock-keeping units, not intended for the main production lines, are picked at one of Scania’s Oskarshamn facilities. These items are intended for various organizations within Scania, but also external business partners.

En mosasaurie (Reptilia, Mosasauridae) av paleocensk ålder?

Mosasaurierna utgör en grupp storvuxna ödlor som dominerade positionen som toppredatorer i haven under sen krittid. Det anses allmänt att dessa marina reptiler, tillsammans med de sista landlevande dinosaurierna, dog ut under det massutdöende som markerar övergången mellan krita och paleogen. Nyligen påträffades en del av ett skalltak tillhörande en mosasaurie i ett löst kalk- och flintblock på enMosasaurs are a group of marine reptiles that are considered to have been the apex predators of the oceans during the Late Cretaceous. It is generally assumed that they, along with the last non-avian dinosaurs, died out during the mass extinction that marks the end of the Cretaceous. Recently, a mosasaur skull roof was discovered in a loose boulder on a beach in the vicinity of the town of Hanstho

Assessment of hip deformities during childhood: a combined approach of Machine Learning and Finite Element Analysis

Kan simuleringsmodeller förbättra diagnostik av förändringar i höftens anatomi hos barn? Anatomiska förändringar i höftleden kan drabba barn i olika åldrar. En följd av detta kan vara att belastningen i höften ändras, vilket kan öka risken för artros. Att undersöka och behandla förändringar i höftens anatomi hos barn är viktigt för att förhindra komplikationer senare i livet. Höftleden är en viOsteoarthritis is a degenerative disease that will cause pain and stiffness in the joint if not treated. One cause of early osteoarthritis in the hip joint is pediatric hip deformities. Hip deformities can occur in different stages of childhood and cause the bones to be misshapen or the hip joint to be dislocated. This thesis aimed to build subject-specific computer models of the hip joint to fi

Accelerating MCR-ALS decomposition of hyperspectral images using k-means clustering

Att analysera livets biokemi, kemin bakom biologiska processer, är viktigt för att få en förståelse av livets mekanismer, speciellt för små längdskalor. Biokemin hos olika typer av celler skiljer sig åt, vilket kan användas för att t.ex skilja mellan elakartade och icke-elakartade tumörer. Vilka ämnen som finns i en cell kan bli identifierade via deras respektive absorptionspektrum. Ett absorptionA hyperspectral image may be decomposed into component spectra and their distri- bution in the image to simplify analysis by revealing underlying patterns and reducing the dimensionality of the image; this may be achieved by the algorithm MCR-ALS. However, the algorithm is time consuming, but could be accelerated by a data re- duction. Data reduction can be done by using a clustering method. In th

Structure-based design of new antiviral substances against the Venezuelan equine encephalitis virus

The Venezuelan equine encephalitis virus (VEEV) is a highly infectious alphavirus responsible for severe outbreaks and epidemics in the Americas. The nonstructural protein 2 (nsP2) cysteine protease plays a fundamental role in the VEEV replication process, and is therefore, an attractive drug target for structure-based drug design using x-ray crystallography. Generating protein crystals of suffi

On The Run: How Fast Runaway Stars Can Escape From Their Home Cluster?

Metoderna för att testa olika hypoteser har genom åren utvecklats drastiskt. Fortfarande finns många experiment som utförs på det klassiska viset i en laborationssal, med fysiska hjälpmedel och mätverktyg för att få fram en slutsats. Dock har den omfattande tekniska utvecklingar breddat möjligheten för vilka hypoteser som är testbara. Genom simuleringar kan tidigare omöjliga teser kontrolleras ochGalaxies contain a wide variety of star clusters with different properties. Each cluster is a possible source of runaway stars, which can be progenitors of supernovae in other parts of the galaxy than their home cluster. Considering runaway stars in galaxy simulations has shown to increase the galactic feedback. This project investigates how the initial condition of a star cluster affects the velo

Will

Culture is in danger. Globalization provides an extraordinary opportunity for people from different countries and areas communicate with each other, which leads to culture intersection. Different kinds of culture are dying with the impact, remaining pity. Within those culture fading away, one of Chinese street food culture called ‘sugar painting’ or ‘sugar man’ attracts me the most. It was popular

Konsten att styra diskret - En kvalitativ studie om samspelet mellan styrning och internal branding

Examensarbetets titel: Konsten att styra diskret - En kvalitativ studie om samspelet mellan styrning och internal branding Seminariedatum: 2 juni 2022 Kurs: FEKH49 - Examensarbete inom organisation - Kandidatnivå. 15 hp. Författare: Isa Lilja Järnström, Claudia Lyckler och Hanna Westerberg Handledare: Tony Huzzard Nyckelord: Resultatstyrning, normativ styrning, internal branding, callcenter, anstäThesis title: The Art of Controlling Discreetly - A qualitative study of the interplay between control and internal branding Seminar date: June 2, 2022 Course: FEKH49 - Bachelor’s Degree Project in Organization. Undergraduate Level. 15 credits. Authors: Isa Lilja Järnström, Claudia Lyckler and Hanna Westerberg Supervisor: Tony Huzzard Key words: Output control, normative control, internal branding

Improving a Background Model for Tracking and Classification of Objects in LiDAR 3D Point Clouds

This thesis studied methods of improving a background model for a data processing pipeline of LiDAR point clouds. For this, two main approaches were evaluated. The first was to implement and compare three different models for detecting ground in a point cloud. These were based on more classical modeling approaches. The second was to use Deep Learning for semantic segmentation of point clouds and t

Dwarf Galaxies in the Galactic Potential as probes of its internal structure

Mörk materia, som det ofta kallas, utgör den största delen av materian i Universum. Då den inte interagerar med synligt ljus, är det omöjligt att observera dess närvaro direkt. Istället får man tillta indirekta metoder, där man kollar på hur objekt som är synliga påverkas av den mörka materiens närvaro. Ett exempel på en sådan metod är att studera egenskaper hos så kallade stellar streams, eller `Satellite galaxies on bound orbits around the Milky Way (MW) are constantly subject to tidal forces from the MW. If the tidal forces are large enough, they may strip the outermost stars of the satellite, that in turn distribute themselves in long elongated stellar streams. These stellar streams are good probes of the Galactic potential due to the constraints they place on the structure of the pote

Vulnerabilities and Capacities in Menstrual Hygiene Management (MHM) and Menstrual Health: A Gendered Perspective

Menstrual hygiene management (MHM) is essential for those who menstruate because they have different needs during disasters and crises. Current MHM is focused on immediate needs instead of longer-term vulnerabilities and capacities. The purpose of this thesis is to examine how gendered vulnerabilities and capacities are considered in MHM policies and practice. The aim is to bring more focus to the

Cybersecurity engagement in a remote work environment

The recent spread of COVID-19 pandemic encouraged organisations to facilitate a remote work environment for their employees. This work environment would most likely continue in the upcoming years. At the same time, this context has been followed with an increase in cyberthreats that could affect the optimal processes of organisations. Because of this, cybersecurity awareness must be cultivated to

Predicting Acute Kidney Injury In Sepsis Patients: A Machine Learning Approach

Kan maskininlärning hjälpa oss förutspå njurskada i sepsispatienter? Årligen drabbas omkring 40000 människor i Sverige av sepsis och närmare 20% av dessa avlider som följd av sjukdomen. Trots detta är sepsis ett relativt okänt sjukdomstillstånd för allmänheten och även inom akademin är det ett relativt ostuderat område. Således finns det ett stort behov av en fördjupad förståelse av sepsis. SepsiSepsis-induced Acute Kidney Injury (AKI) is a common cause of death for sepsis patients, yet it is poorly understood by the medical community. Using Tandem Mass Spectrometry (MS/MS), the blood plasma of 135 sepsis patients has been analyzed to measure protein levels. In this thesis we will present an exploratory application of machine learning on this data, with the intent of producing a model wh