Refine
Document Type
- Master's Thesis (4) (remove)
Language
- English (4) (remove)
Has Fulltext
- yes (4)
Is part of the Bibliography
- no (4)
Keywords
- AI (1)
- Automatisierungstechnik (1)
- Computer Vision (1)
- Deep learning (1)
- Deutschland (1)
- Diffusion Models (1)
- E-Learning (1)
- Ethernet APL (1)
- GAN (1)
- Generative Adversarial Network (1)
In the last years generative models have gained large public attention due to their high level of quality in generated images. In short, generative models learn a distribution from a finite number of samples and are able then to generate infinite other samples. This can be applied to image data. In the past generative models have not been able to generate realistic images, but nowadays the results are almost indistinguishable from real images.
This work provides a comparative study of three generative models: Variational Autoencoder (VAE), Generative Adversarial Network (GAN) and Diffusion Models (DM). The goal is not to provide a definitive ranking indicating which one of them is the best, but to qualitatively and where possible quantitively decide which model is good with respect to a given criterion. Such criteria include realism, generalization and diversity, sampling, training difficulty, parameter efficiency, interpolating and inpainting capabilities, semantic editing as well as implementation difficulty. After a brief introduction of how each model works on the inside, they are compared against each other. The provided images help to see the differences among the models with respect to each criterion.
To give a short outlook on the results of the comparison of the three models, DMs generate most realistic images. They seem to generalize best and have a high variation among the generated images. However, they are based on an iterative process, which makes them the slowest of the three models in terms of sample generation time. On the other hand, GANs and VAEs generate their samples using one single forward-pass. The images generated by GANs are comparable to the DM and the images from VAEs are blurry, which makes them less desirable in comparison to GANs or DMs. However, both the VAE and the GAN, stand out from the DMs with respect to the interpolations and semantic editing, as they have a latent space, which makes space-walks possible and the changes are not as chaotic as in the case of DMs. Furthermore, concept-vectors can be found, which transform a given image along a given feature while leaving other features and structures mostly unchanged, which is difficult to archive with DMs.
As noted by Roman poet Virgil already more than 2,000 years ago: “The greatest wealth is health.”. Without health, there is no happiness, no peace, and no success according to the Reflections Recovery Center from Arizona, United States (USA, U.S.). The goal of the Healthy People 2020-project (HP2020), which is led by the Office of Disease Prevention and Health Promotion (ODPHP), was to “promote quality life, healthy development, and health behaviors across all life stages” among the U.S. population. HP2020 measures progress by using so-called Leading Health Indicators (LHI), reliable data sources, baseline values as well as targets for LHI-individual improvements for every measurable objective to be achieved by 2020 and each following decade. In the further course, these values were compared to student populations from the U.S., Germany, and Poland. The goal of this master's thesis was to obtain more data on international health, particularly among student populations. For the statistical analysis, data were obtained from an online survey that was distributed to students in at least one university in each of the three countries. In total, data from 380 students were analyzed in terms of HP2020 goal attainment. To determine if statistically significant differences were present, the z-test was used. The biggest differences emerged on the following topics: access to healthcare, environmental quality, obesity as well as reproductive and sexual health.
Pathologists need to identify abnormal changes in tissue. With the developing digitalization, the used tissue slides are stored digitally. This enables pathologists to annotate the region of interest with the support of software tools. PathoLearn is a web-based learning platform explicitly developed for the teacher-student scenario, where the goal is that students learn to identify potential abnormal changes. Artificial intelligence (AI) and machine learning (ML) have become very important in medicine. Many health sectors already utilize AI and ML. This will only increase in the future, also in the field of pathology. Therefore, it is important to teach students the fundamentals and concepts of AI and ML early in their studies. Additionally, creating and training AI generally requires knowledge of programming and technical details. This thesis evaluates how this boundary can be overcome by comparing existing end-to-end AI platforms and teaching tools for AI. It was shown that a visual programming editor offers a fitting abstraction for creating neural networks without programming. This was extended with real-time collaboration to enable students to work in groups. Additionally, an automatic training feature was implemented, removing the necessity to know technical details about training neural networks.
The trend towards the use of Ethernet in automation networks is ongoing. Due to its high flexibility, speed, and bandwidth, Ethernet nowadays is not only widely used in homes and offices worldwide but finding its way into industrial applications. Especially in automation processes, where many field devices send data in relative short time spans, the requirements for a safe and fast data transfer are high. This makes the use of industrial Ethernet essential. A new hardware-layer, specifically tailored for industrial applications, has been introduced in the form of Ethernet-APL (‘Advanced Physical Layer’). Ethernet-APL is based on the Ethernet standard and implements a two-wire Ethernet-based communication for field devices and provides data and power over a two-wire cable. The operation in areas with potentially explosive atmosphere is also possible. This enables a modular, fast, and transparent Ethernet network structure throughout the entire plant. However, by integrating Ethernet-APL into the field, industrial networks in the future will face the challenge of operating at varying datarates at different locations in the network, resulting in a ‘mixed link speed’ network. This can lead to limitations in packet-throughput and consequently to potential packet loss of system relevant data, which must be avoided. Therefore, the purpose of this thesis is to investigate the potential of packet loss in ‘mixed link speed’ networks.