Wykaz obszarów badawczych związanych z tagiem Machine-learning:
| # | Obszar badawczy | Dziedzina naukowa |
|---|---|---|
| 1 |
see https://www.gagolewski.com/
|
|
| 2 |
Use of information technology for proactive detection (identification and exposure) of disinformation and misinformation content - automated methods for flagging disinformation content, tracking propagation channels, detecting original sources of spreading false information. I am primarily interested in projects enabling the combination of engineering sciences with social sciences in preventing and combating disinformation, e.g. by applying methods of Generative Adversarial Networks and Large Language Models, in conjunction with psychosocial analysis of the problem. It is important for me to focus on issues related to the impact of disinformation on crisis situations, amplification of public sentiment, manipulation of public opinion (public health crises, natural disasters, warfare using asymmetric and hybrid methods).
|
|
| 3 |
1)Methods of detection and classification of heart abnormalities using deep learning techniques - concerns techniques and algorithms for processing ECG signals for automatic detection and classification of heart abnormalities using deep machine learning methods. It is planned to use a variety of deep convolutional neural networks for automatic analysis of the structure (waveform) of ECG signals, as well as mechanisms for detecting anomalies (deviations) and searching for information based on autoencoder neural networks. The work will be carried out in cooperation with the Department of Cardiology-Intensive Therapy and Internal Diseases of the Medical University of Karol Marcinkowski in Poznań. 2)Methods of detection and classification of heart abnormalities using heart rate variability (HRV) parameters and machine learning techniques - concerns techniques and algorithms for processing ECG signals for automatic detection and classification of heart abnormalities using machine learning methods. It is planned to use a wide variety of heart rate variability (HRV) parameters as well as new original heart rate asymmetry (HRA) parameters. The work will be carried out in cooperation with the Department of Cardiology-Intensive Therapy and Internal Diseases of the Medical University of Karol Marcinkowski in Poznań. 3)Passive radar for space object detection using signals recorded by antennas of the international network of radio telescopes LOFAR - The research area concerns the techniques and algorithms for processing signals recorded by the antennas of the network of radio telescopes LOFAR (Low-Frequency Array for radio astronomy) in order to use them for passive radiolocation of space objects: satellites in low orbits and so-called space debris. The considered system of passive radiolocation does not require the construction of dedicated transmitters, but uses the existing so-called illuminators of opportunity, e.g. FM, DAB + or TV DVB-T transmitters. After reflecting the signals from the objects, they are received by the antennas of the LOFAR system. The research in this area, carried out in cooperation with the Space Research Center of the Polish Academy of Sciences, is pioneering on a global scale. Three LOFAR stations are located in Poland. A single LOFAR station consists of many antennas creating a large radio telescope that can receive relatively weak signals. 4) Methods and algorithms for signal processing in passive radar for small unmanned aerial vehicles (drones) - concerns techniques and algorithms for signal processing dedicated to the passive radiolocation of small unmanned aerial vehicles (drones). The work will be carried out in cooperation with the Faculty of Power and Aeronautical Engineering of the Warsaw University of Technology at the airport in Sieraków near Przasnysz, recently purchased by the Warsaw University of Technology, where the Area Monitoring Laboratory with four antenna stations was built. Problems to be solved within the research area are related to detection of small flying objects with the use of specific features of signals reflected from the considered objects, estimation of their parameters, tracking, as well as classification of detected objects, in particular, the research is to focus on the possibility of distinguishing small drones from birds. 5) Optimization of methods and algorithms of people identification based on the EEG signal with the use of machine learning techniques - concerns the methods and algorithms of people identification based on the EEG signal with the use of machine learning techniques. It is planned to use the approach based both on the spectral features of the EEG signal in its individual bands, as well as the analysis of the EEG signal itself using the so-called deep learning techniques with convolutional neural networks. The work will include the selection and optimization of EEG signal parameters and classifiers used to identify people, the number of sessions necessary to train classifiers, the minimum number of electrodes used for identification, as well as the development of headbands/caps dedicated to collecting the EEG signal for the application under consideration, and testing the developed solutions under the conditions similar to their practical implementation. The research will be carried out in cooperation with the Nencki Institute of Experimental Biology of the Polish Academy of Sciences.
|
|
| 4 |
Operations and Maintenance system in 6G Mobile Networks: analysis of Machine Learning impact in the network management as a whole (multi-agent environment)
|
|
| 5 |
Research in the area of supply chain management (SC): flexibility and resilience of SC to internal phenomena and global disruption (natural disasters, wars, diseases, sanctions), management and shaping of SC, supply, distribution, risk management, development and/or the use of decision support tools in SC based on the machine learning, artificial intelligence, simulation.
Research in the field of city logistics: planning a cargo distribution system, developing innovative solutions for the last mile logistics, infrastructure designing for environmentally-friendly vehicles, including methods of locating charging stations for electric vehicles, developing and/or using decision support tools in urban logistics based on the machine learning, artificial intelligence, simulation.
|
|
| 6 |
computer vision (SLAM, visual search), machine learning (deep learning, generative models, continual learning), representation learning (binary descriptors).
|
|
| 7 |
Machine learning for social and sensor data This issue is dedicated to candidates with knowledge of machine learning methods, the ability to acquire and process data and those interested in both development in this area and the application of these methods to social issues. The goal will be to develop methods for modeling transport behaviors using data from various sources and machine learning methods, including methods dedicated to data streams and transfer learning methods. In particular, it is planned to use data streams describing the actual availability of various means of communication and their fusion, e.g. with survey data. This research issue is directly related to the ongoing international CoMobility research grant including research groups focused on social research and air pollution modelling.
|
|
| 8 |
The objective of the PhD student will be to train a neural network based on a family of impedance frequency characteristics measured for magnetic nanocrystalline rings. This network will enable the mapping of the measured frequency characteristics with lumped element electric circuit (LEEC) model. The intention is to replace the modified-Pade method, as published in the article CEM2014 with a new method based on neutral network (the CEM2014 paper: Szewczyk, J. Pawłowski, K. Kutorasiński, S. Burow, S. Tenbohlen, W. Piasecki, "Identification of a rational function in the s-domain describing the frequency characteristics of a magnetic material," Proceedings of the 9th International Conference on Computation in Electromagnetics (CEM), 31st March - 1st April 2014, London, UK). More information, see: https://wutwaw-my.sharepoint.com/:w:/g/personal/marcin_szewczyk_pw_edu_pl/Ea6m_83KcehBo-oi2NuWHb8BlwiEg6mxS307fjKB0lp3GA?e=h45tAF
|
|
| 9 |
Self-supervised learning (SSL) methods have shown significant potential in improving the sample efficiency of deep learning and in providing a starting point for the learning of downstream tasks. Contrastive SSL methods have become a standard pre-training approach for a range of domains such as NLP and vision. Initial results suggest that non-contrastive SSL has been able to narrow the gap to the performance levels of contrastive methods, while obviating the need for explicit construction of negative samples. The aim of this research project is two-fold: 1) investigate the extent to which contrastive and non-contrastive methods can be used in novel SSL architectures, 2) examine whether the alignment of representations in SSL can be achieved by alternative methods, such as enforcing the ability to predict an input representation from the representations of similar inputs.
|
|
| 10 |
A prevalent theme present in the contemporary representation learning approaches is to pre-train large foundation models on huge datasets. Such approaches utilize static datasets constructed at a particular point in time, which contrasts with the constantly changing and expanding nature of data available on the internet. The proposed research will explore a new paradigm where the training dataset is constructed on the fly by querying the internet, enabling efficient adaptation of representation learning models to selected target tasks. The aims of this research project include 1) design methods to query relevant training data and use it to adapt the representation learning model in a continuous manner, 2) make progress towards building self-supervised methods that given a description of a task, autonomously formulate their learning curricula, query the internet for relevant training data, and use it to iteratively optimize the model.
|
|
| 11 |
Abstract Visual Reasoning (AVR) comprises problems that resemble those appearing in human IQ tests. For example, Raven's Progressive Matrices present a set of images arranged in a 3x3 grid with a missing panel in the bottom-right corner. The test-taker has to discover relations governing 2D shapes (and their attributes) located in the images to select an answer, from a provided set of options, that best completes the matrix. In general, AVR tasks focus on fundamental cognitive abilities such as analogy-making, conceptual abstraction, or extrapolation, which makes advancements delivered by this research applicable to diverse areas, extending well beyond the investigated tasks. In this research we plan to verify the abilities of Large Language Models (LLMs) and Large Vision Models (LVMs) to solve AVR tasks, both synthetic and representing real-world images.
|
|
| 12 |
“Machine learning-based diagnosis of spine injuries using computed tomography” Computed tomography (CT) is a crucial imaging technique in medical diagnosis and is the preferred modality for assessing spinal trauma. However, the large volume of image data generated during tomographic examinations presents significant challenges for image analysis and diagnosis. Artificial intelligence (AI) offers the potential to enhance the speed and accuracy of diagnosis in such cases. This research aims to explore the application of machine learning (ML) methods and deep neural networks (DNNs) for the automated detection of traumatic vertebral body injuries. The study will focus on classifying vertebral fractures, distinguishing between traumatic and non-traumatic cases. The learning dataset will be constructed using trauma examination records from a clinical hospital in Warsaw. These records will consist of X-ray spine tomography studies in DICOM format, annotated by experienced radiologists. The AO Spine Classification system for thoracolumbar injuries will serve as the framework for categorizing spinal fractures. To address the computational challenges posed by the high dimensionality of tomographic data (i.e., the large number of voxels per examination), methods for reducing data size will be employed. Various deep neural network architectures will be evaluated to determine their efficacy and performance in fracture classification. Furthermore, the interpretability and explainability of the developed ML-based approach will be analyzed using tools and techniques from Explainable AI (XAI). Efforts will be made to validate the reliability of the recommendations generated by the ML models, ensuring they align with clinical expertise and established diagnostic standards.
|
|
| 13 |
“Deep neural networks for image reconstruction in electrical capacitance tomography” Electrical capacitance tomography (ECT) enables the visualization of the spatial distribution of an object’s electrical permittivity. Image reconstruction in ECT presents a significant challenge as it is an ill-posed and ill-conditioned inverse problem. Advanced nonlinear algorithms, such as the Levenberg-Marquardt method, are iterative and computationally intensive, primarily due to the repeated calculation of the Jacobian matrix. Deep neural networks (DNNs) have emerged as a promising alternative for image reconstruction. This work will explore deep network architectures that not only match but significantly outperform classical methods in reconstruction quality. Synthetic data will be employed for supervised learning, and the performance of DNNs will be tested against real measurements. Application-specific training datasets will be analyzed, and their size will be expanded using data augmentation techniques. A comparative evaluation of various training datasets will also be conducted. To optimize the network, different loss functions and solvers will be utilized. Reconstructed images will be generated using DNNs and compared against a baseline provided by the Levenberg-Marquardt algorithm. Results will be assessed using selected image quality metrics. It is anticipated that deep networks will enhance the spatial resolution of ECT scanners. The insights gained from selecting deep networks, analyzing their architectures, and refining training strategies can potentially be applied to solve inverse problems in other fields.
|