A5046091617- Landslides and related hazards
- Tree Root and Stability Studies
- Geothermal Energy Systems and Applications
- Soil and Unsaturated Flow
- Reinforcement Learning in Robotics
- Cryospheric studies and observations
- Climate change and permafrost
- Machine Learning and Data Classification
- Geotechnical Engineering and Soil Stabilization
- Adversarial Robustness in Machine Learning
- Advanced Control Systems Optimization
- Mycorrhizal Fungi and Plant Interactions
- Data Stream Mining Techniques
- Flood Risk Assessment and Management
- Soil Management and Crop Yield
- Meteorological Phenomena and Simulations
- Scientific Computing and Data Management
- Heat Transfer and Boiling Studies
- Spacecraft and Cryogenic Technologies
- CO2 Sequestration and Geologic Interactions
- Geotechnical Engineering and Analysis
- Gaussian Processes and Bayesian Inference
- Atmospheric and Environmental Gas Dynamics
- Hydraulic Fracturing and Reservoir Analysis
- Aerospace and Aviation Technology
RWTH Aachen University
2022-2025
Istanbul University
2022-2024
Istanbul Medipol University
2022-2024
Newcastle University
2019-2022
University of Göttingen
2021
Stanford University
2021
Istanbul Technical University
2018-2020
ETH Zurich
2018-2020
Swiss Federal Institute for Forest, Snow and Landscape Research
2018-2019
Soils with roots or root-like inclusions have often been tested in direct shear to quantify the effects of vegetation on strength soil, and turn, stability slopes. However, a straightforward evaluation root reinforcement is challenging due complex nature roots, dependency soil behaviour many factors. An Inclinable Large-scale Direct Shear Apparatus (ILDSA) was built study shearing root-permeated soils. Planted specimens, consisting two different sets species, were prepared moraine, sampled...
Deterioration of earthworks and the resultant implications for serviceability increased occurrence failures have a significant negative impact on transport networks both in UK internationally. There is evidence field that deterioration processes are occurring over life an asset, comprising cracking loss suction. These weather-driven occur absence mechanical loads can lead to failure many years after construction. To demonstrate progressive performance clay fill due purely environmentally...
Abstract The Reusability-targeted Enriched Sea Ice Core Database (RESICE) combines data and metadata from 287 sea ice cores. database enables reuse scenarios such as the validation of physics-based models training data-driven algorithms. RESICE is enriched in two ways. First, originating 138 sources including 107 sets repositories Zenodo, Australian Antarctic Data Center Pangaea. Second, contains additional automatically generated tailored to specific scenarios. checked for plausibility...
Shallow landslides, either on bare or vegetated slopes, can be triggered after a rainfall event due to loss of suction. An extensive laboratory programme was performed in this study assess the relationships between water content, plant-induced suction, root biomass and shear strength parameters. Root-permeated soils, planted with combinations different species, were tested an inclinable large-scale direct apparatus. The effects mycorrhizal fungi also investigated inoculated specimens....
Abstract Machine learning has grown in popularity the past few years for susceptibility and hazard mapping tasks. Necessary steps generation of a or map are repeatedly implemented new studies. We present Random Forest classifier-based landslide framework to facilitate future studies using machine learning. The framework, as piece software, follows FAIR paradigm, hence is set up transparent, reproducible modularly extensible workflow. It contains pre-implemented from conceptualisation...
Online solvers for partially observable Markov decision processes have difficulty scaling to problems with large action spaces. Monte Carlo tree search progressive widening attempts improve by sampling from the space construct a policy tree. The performance of is dependent upon policy, often requiring problem-specific samplers. In this work, we present general method efficient based on Bayesian optimization. proposed uses Gaussian process model belief over action-value function and selects...
Dispersed space heating alone accounts for 40% of UK energy use and 20% carbon dioxide (CO 2 ) emissions. Tackling building cooling demands is therefore critical to achieve net-zero ambitions in the UK. The most energy-efficient way reduce emissions through ground-source heat pumps district technology. However, capital costs are often high, sometimes prohibitively so. To investment costs, it proposed buried infrastructure as sources stores thermal energy. Barriers this innovative approach...
The current study was designed to assess the response treatment, as well clinical and survival outcomes, across different breast cancer subtypes in patients who underwent neoadjuvant chemotherapy (NAC). From 2014 2019, a total of 139 were histologically confirmed have cancer, NAC, subsequently received axillary surgery, retrospectively included this study. rates pathological complete (pCR) NAC significantly higher for HER2-positive triple-negative than luminal A HER2-negative (p = 0.013)....
Both above- and below-ground biomass of plants provide additional strength to soil through different mechanisms, which help increase the stability a vegetated slope. Nonetheless, shallow landslides on steep slopes covered with vegetation still occur, often being triggered above groundwater table, due loss suction subsequent rainfall. Therefore, it is essential know what extent enhances slope stability, quantify contribution shear determine factors safety. Results large-scale direct...
The aim of this study was to investigate the predictive value PLR and NLR as an indicator pathological complete response (pCR) in patients with breast cancer after NACT. One hundred thirty-nine early or LABC candidates NACT were retrospectively analyzed. prognostic significance In addition, indicators pCR also evaluated. obtained 48.9% patients. Significant difference detected between PLR, tumor grade, clinical lymph node status molecular subgroup. higher rate significantly achieved for low...
Mechanical contribution of roots to the shear strength saturated soils is not only function vegetation in favour slope stability. Plants also cause an increase due plant-induced suction, resulting from partial saturation. Although understanding, quantifying and modelling shearing behaviour directly related volume change characteristics soil, dilative rooted has been much interest. Large-scale suction-monitored direct tests were performed on specimens planted with a mixture species different...
This paper describes a novel finite-state conditional switching structure that enables autonomous recovery for large envelope of loss-of-control conditions. The control architecture mimics human pilot actions through numerous discrete states are activated individually as the proceeds, based on guidelines inspired by Federal Aviation Administration's suggestions piloted upset recovery. conditions between optimized using interpolation algorithms extend while decreasing average time. presented...
Online solvers for partially observable Markov decision processes have difficulty scaling to problems with large action spaces. This paper proposes a method called PA-POMCPOW sample subset of the space that provides varying mixtures exploitation and exploration inclusion in search tree. The proposed first evaluates according score function is linear combination expected reward information gain. actions highest are then added tree during expansion. Experiments show able outperform existing...
Epithelial Ovarian cancer (EOC) is the most lethal gynecologic worldwide. Carboplatin (CP) main chemotherapeutic agent in treatment of ovarian cancer. However, development a hypersensitivity reaction (HSR) 10% to 15% patients with EOC an important limiting factor for clinical use CP. Herein, we aimed investigate efficacy and safety CP-desensitization (CP-D) therapy recurrent EOC. Forty-seven cases treated at Istanbul University Oncology Institute were retrospectively analyzed between...
Both Green Infrastructure and Ground Source Heat Exchangers provide opportunities to significantly improve the resilience sustainability of our built environment. This work explores thermo-hydrological response a vegetated Sustainable Drainage System under physically simulated heat injection conditions using soil column 1800 mm diameter 950 height in heavily-instrumented lysimeter. A range field testing scenarios (thermal load cycling) were applied natural, external ambient conditions. Soil...
Abstract Machine learning has grown in popularity the past few years for susceptibility and hazard mapping tasks. Necessary steps generation of a or map are repeatedly implemented new studies. We present Random Forest classifier-based landslide framework to facilitate future studies using machine learning. The framework, as piece software, follows FAIR paradigm, hence is set up transparent, reproducible modularly extensible workflow. It contains pre-implemented from conceptualisation...
Despite having a higher mortality risk than conventional chemotherapeutics, high-dose chemotherapy (HDCT) has the potential to be curative in relapsed/refractory germ-cell tumors. Therefore, selecting best patient group for this treatment is critical. This study aimed determine factors that affect survival our GCT cohort who received HDCT and autologous stem-cell transplantation. Between September 2010 2020, we included 44 male patients with treated plus The patients' demographic features,...
The pursuit of reproducibility in research has long been emphasized. It is even more critical geohazards and practice, where model-based decision-making needs to be transparent for trustworthy applications. However, enabling process-based or machine learning workflows requires time, energy, sometimes manual operations unavailable resources. Moreover, the diversity modern compute environments, both hardware software, significantly hinders path reproducibility. While many researchers focus on...
Non-clear cell renal carcinoma (non-ccRCC) is a highly heterogeneous disease group, accounting for approximately 25% of all RCC cases. Due to its rarity and especially heterogeneity, phase III trial data limited treatment options generally follow those clear RCC. In the literature, there exists number studies with sunitinib, cabozantinib, everolimus, but on efficacy pazopanib are limited. Our aim in this study was compare multicenter retrospective cohort non-ccRCC patients. included patients...
The efficiency of shallow ground heat exchangers is controlled by the thermal conductivity soil they are placed in, which predominantly a function type, mineralogy, density and water content. As such systems at depths where hydrological regime highly fluctuant due to rainfall evaporation, it important have soil-specific information on how changes with There numerous models available in literature, most not able define full range dryout curve (TCDC) single equation. This study suggests...
Introduction: Increasing complexity and capacity of computational physics-based landslide run-out modelling yielded highly efficient model-based decision support tools, e.g. susceptibility or maps, geohazard risk assessments. A reliable, robust reproducible development such tools requires a thorough quantification uncertainties, which are present in every step workflow from input data, as topography release zone, to framework used, numerical error. Methodology: Well-established methods...