The crucial role of information retrieval (IR) is highlighted by its presence across a wide range of tasks, such as web search and fact-checking, and domains, including finance and healthcare. Effective and efficient IR systems are critical for finding relevant information from vast amounts of data. Traditional sparse retrieval methods such as BM25 are efficient but often fail to capture the context, while more recent dense retrieval models are highly inefficient in terms of resources and latency.
In our research, we evaluate multiple Transformer-based models to understand the impact of the semantic re-ranking phase within interpolation-based re-ranking, using the FAST-FORWARD indexes framework, a retrieve-and-re-rank approach which combines the benefits of both lexical and semantic matching. We focused on identifying specific scenarios in which particular models excel in terms of ranking performance or latency, aiming to provide model recommendations tailored to different settings. Our evaluations reveal that no single model outperforms others across all datasets. We hypothesise that the main factors influencing encoder performance are the datasets used for finetuning and the method employed for computing the contextualised vector embedding. However, ablations studies would be beneficial for validating these observations.

Interpolation-based re-ranking emerged to make dense retrieval possible in low-latency applications such as web engine search. However, to this day there is no clear winner among the different ranking approaches. Moreover, missing document scores in hybrid retrieval have not been investigated in detail. This paper aims to address this by comparing the interpolation-based re-ranking with dense and hybrid retrieval approaches in terms of ranking performance and latency. It goes to show that while interpolation-based re-ranking has a notable latency advantage, hybrid retrieval achieves best performance in ranking metrics for an in-domain setting. Additionally, differences in missing document score techniques are slight with zero imputation emerging on top.

Ad-hoc retrieval involves ranking a list of documents from a large collection based on their relevance to a given input query. These retrieval systems often show poorer performances when handling longer and more complex queries. This paper aims to explore methods of improving retrieval effectiveness on these types of queries across different information retrieval (IR) tasks, within the context of Fast-Forward indexes. An analysis is conducted to determine the actual impact of query length and complexity. Interestingly, the hypothesis that longer queries are more challenging does not hold true for all cases, and in some datasets the opposite is true. To improve the performance of long and complex queries, two approaches are explored: utilising multiple dense models during the re-ranking stage instead of the traditional single model and reducing the queries via large language models. The use of multiple dense models for re-ranking proves to be effective, with two models providing the best balance between performance and ranking quality. Utilising LLM's for query reduction achieves performance similar to the original queries but fails to improve their ranking scores.

In the past years, data has become increasingly important to more and more domains, leading to more efficient decision-making. As the amount of collected data grows, there is an increased need for tools that help with various Information Retrieval (IR) tasks. One of the most widespread IR tasks is ad-hoc retrieval which, for a given search query, returns a list of relevant documents from a large corpus ordered by their relevance. Initial Ad-hoc retrieval models were based on term matching, which could not overcome vocabulary mismatch. On one hand, initial strategies aiming to overcome semantic limitations were adopting query expansion, augmenting the initial search query with new terms to capture more relevant documents. On the other hand, newer strategies rely on Natural Language Processing (NLP) for ranking documents by semantic similarity. One such example is retrieve-and-re-rank models, which retrieve documents by their lexical similarity and re-rank the retrieved documents based on semantic similarities, by making use of NLP embedding models. This research focuses on analysing the performance of combining RM3, a pseudo-relevance feedback query expansion strategy, with the semantic re-ranking model TCT-ColBERT. This model is compared with the lexical retrieval model BM25 which serves as a benchmark, as well as with its components RM3 and TCT-ColBERT. Results indicate that on certain tasks, the model performs better (up to 7%), while on other tasks it performs worse (up to 3%).

This paper is an analysis of the performance and logic behind different query expansion models. Query expansion and pseudo relevance feedback are techniques for adding more terms to a query based on the results of an initial query and the data in the body of documents. Four different query expansion models that are provided in the pyterrier python library and its extensions have been analysed, namely Bo1, KL, RM3, and Axiomatic query expansion. It was found that Axiomatic query expansion often does not perform any query expansion, and when it does, has no increase in performance. Bo1 and KL, although different in exact logic, have similar results most of the time. The most significant difference is the execution time, with Bo1 being faster with larger datasets and KL being faster with many documents on smaller datasets. Lastly, RM3 while not having a dominant performance has a lot of potential for good results with the right combination or parameters.

This research explores the landscape of dataset generation through the lens of Probabilistic Principal Component Analysis (PPCA) and β-Conditional Variational Auto-encoder (β-CVAE) models. We conduct a comparative analysis of their respective capabilities in reproducing datasets that mirror the distribution of the original data that comes from a hypothetical pathway kinetic model based on an E.coli strain using varied parameter settings falling within a specified range. The requirement of significant prior investment in acquiring accurate details about the distinct mechanisms governing each reaction and its parameters for the construction of these kinetic models push us to find alternative ways to generate data that guide metabolic engineering processes. This paper tries to find a viable option through compression algorithms that reduce dimensionality. The PPCA model demonstrates commendable fidelity in capturing overarching patterns, though areas for refinement in reproducing specific data points are identified. In contrast, the β-CVAE model exhibits higher fidelity, precision, and consistency, positioning it as a robust choice for data generation tasks. This study was constrained by both time and the specificity of the model architectures and the dataset. These limitations underscore the imperative for continual exploration and refinement within the dynamic landscape of generative modeling. Opportunities could be found in the refinement of both VAE, CVAE and β-CVAE models utilizing varied hyperparameters alongside different architectures, to increase applicability across diverse datasets within the realm of metabolic engineering.

This study investigates the application of generative models for synthetic data generation in pathway optimization experiments within the field of metabolic engineering. Conditional Variational Autoencoders (CVAEs) use neural networks and latent variable distributions to generate new, plausible data samples. We adapt this model by conditioning the training process on the target flux to acquire increased performance.

Additionally, a baseline model, namely Probabilistic Principal Component Analysis (PPCA), was selected for a comparative analysis to generate the underlying latent space to test the hypothesis that a type of Variational Autoencoder (VAE) can be used to learn a reduced-dimensional latent space for configurations of a kinetic pathway model. A dataset comprising 5000 hypothetical configurations of a kinetic pathway model was utilized to extract relationships between elements of a kinetic pathway.

The results indicate that PPCA can model the underlying distribution of the dataset when the latent space is large enough. However, the traditional CVAE might struggle to capture the underlying distribution, resulting in an entangled latent space. The study suggests that an implementation of $\beta$-CVAE could lead to a better balance between parts of the objective function during training, offering improved prospects for generating cost-efficient kinetic pathways for combinatorial pathway optimization experiments.

This research investigates the application of Generative Adversarial Networks (GANs) and probabilistic Principal Component Analysis (PPCA) in generating synthetic data for pathway optimization in metabolic engineering. The study aims to compare the performance of these generative models, addressing key questions regarding their utilization, the quality of generated data compared to experimental data, and overall efficiency. The dataset comprises 5000 parameter configurations of kinetic models that simulate a hypothetical pathway. Constructing kinetic models traditionally involves obtaining complex scientific knowledge, a process that may be alleviated through a data-driven approach. Results indicate that both models, tried with different sizes of latent space, demonstrate good performance in modeling the underlying latent space of the data. However, GANs with the right set of parameters exhibit a better performance, evidenced by lower KL divergence and superior visual structure in the generated data. The findings highlight the potential of GANs to outperform probabilistic PCA, offering valuable insights for more cost-effective and streamlined strain optimization in metabolic engineering. Overall, this research advocates for further investigation of GANs capabilities in metabolic engineering as a potentially powerful tool for synthetic data generation.

Images possess the ability to convey a wide range of emotions, and extracting affective information from images is crucial for affect prediction systems. This process can be achieved through the application of machine learning algorithms. Categorical Emotion States (CES) and Dimensional Emotion Space (DES) are two typical models used for representing emotions. Moreover, the development of a mapping schema between these representations can contribute to benefit the research in AI and psychology. Consequently, this study focuses on investigating the feasibility of translating emotions from DES to CES. To accomplish this goal, relevant databases are identified and combined as the training data, and machine learning models, namely Naive Bayes, K-nearest Neighbors, and Decision Tree are employed to perform the classification task. The results indicate the superior performance of the K-nearest Neighbors classifier, exhibiting higher mean accuracy (60%) and low standard deviation (0.004) among all implemented classifiers. Overall, the translation from DES to CES offers several advantages, including a simplified and interpretable representation of emotions, as well as the provision of a common language for discussing and expressing emotions.

Supervisory Control and Data Acquisition (SCADA) systems are sometimes exposed on the public Internet. It is possible to quickly and efficiently identify such exposed services. They are commonly part of critical infrastructure, so they need to be protected against cyber attacks. In the past, researchers have scanned the Internet to detect such systems. However, such data may be biased due to honeypots set up by other researchers, which are fake hosts mimicking real industrial systems in order to detect malicious attacks.
In this paper, we develop a methodology to discover SCADA systems, classify them as real or honeypots, and analyse the metadata collected from them. We show that a large part of all exposed SCADA services are in fact likely to be honeypots, and we find correlations between independent honeypot-related indicators.

Studies in Music Affect Content Analysis use varying emotion schemes to represent the states induced when listening to music. However, there are limited studies that explore the translation between these representation schemes. This paper explores the feasibility of using machine learning models to translate from a dimensional scheme of Valence, Energy and Tension, to a categorical emotion of Anger, Fear, Happy, Sad, or Tender, specifically in the context of musical stimuli. Additionally, this paper considers how the close proximity of certain emotions in the dimensional space, such as Fear and Anger, negatively influence the performance of translation models. This paper reflects on past studies and presents new results concluding feasible translations of music affect content, moreover, providing suggestions for future analysis.

The objective of this report is to establish and present a machine learning model that effectively translates affect representation from emotional attributes such as arousal (passive versus active) and valence (negative versus positive) to dominance (weak versus strong). In the pursuit of this goal, various research questions are addressed. The paper outlines the process of dataset selection, ensuring appropriateness for the problem at hand. Subsequently, a comprehensive investigation into suitable evaluation methods for the developed model is conducted, providing well-reasoned justifications for the chosen approach. An additional research question focuses on assessing different machine learning approaches to determine the optimal performer. The motivation behind this translation lies in the recognition of the interdependence between these affect attributes, supported by both theoretical underpinnings and practical evidence. This contrasts with previous studies that have treated these dimensions as independent descriptors for representing emotions.

This paper explores algorithms to optimize networked caching, where requests for files can be handled by a local cache instead of a remote server. Caches work collaboratively to prevent redundant caching, and each new batch of file requests is used to update the entire network. Data is cached in an integral manner, meaning that only discrete files or chunks can be stored, not fractions of them. Bipartite networks are studied, though the proposed model supports arbitrary network topologies. The approach is based upon an Online Mirror Descent (OMD) policy, which has been shown to have sublinear regret in single cache scenarios.

The way that surgeons currently use surgical navigation technology impacts their hand-eye coordination and their ability to view images and data critically. To tackle this issue Augmented Reality goggles, built from the HoloLens 2, have been developed specifically for the purpose of aiding surgeons during surgery and it is called: The HoloNav. Optical reflective spheres are commonly used to track surgical instruments. This research aims to find out if grayscale cameras could contribute to locating these spheres more accurately with the HoloNav. This is done by finding the spheres on the infrared images, reducing the search space on the grayscale images and finding the spheres on the reduced search spaces of the grayscale images. By triangulating those retrieved coordinates and comparing them with the optically tracked coordinates a conclusion can be drawn about the accuracy. The accuracy measured in this research is between 0.66 mm and 10.64 mm (mostly between 1.8 mm and 6.4 mm) depending on the frame. This is quite accurate and similar to the results from related work. With better image quality or different input conditions the accuracy could be improved even more.

Surgical navigation is a tool that surgeons rely on everyday to perform accurate surgeries all over the world. However, this technology requires good hand-eye coordination and a high level of concentration. HoloNav is a project that inquires to see if using the HoloLens and augmented reality can replace the current surgical navigation methods. To do so, the HoloLens must be able to identify the patient and the location of the surgery instruments, which uses optical reflective spheres. This study focuses on using the grayscale cameras of the HoloLens and a deep learning algorithm YOLOv5 to test if it is possible to precisely detect optical reflective spheres. 3 models were trained with two different data sets, where the results show that the model trained on a data set would perform well on the validation set. However, they would perform far worse when exposed to a data set it was not trained on.

In order to be able to use the Microsoft HoloLens for surgical navigation purposes, performing good patient alignment is of utmost importance. This paper will discuss how this patient alignment can be done using different point cloud registration algorithms.
A lot of research is being conducted on point cloud registration algorithms. However, most research assumes that the point clouds to be aligned are almost identical, while patient alignment aims at aligning a very detailed pre-operative scan with a very sparse point cloud obtained by the surgeon using an optical marker.
In order to get around this problem, the HoloLens' depth camera is used to obtain a detailed point cloud so that registration algorithms can be used. Then the performance of different point cloud registration algorithms is tested on this depth sensor data to see whether using the HoloLens' depth sensor is a viable option for patient alignment.
From the results, it appears that algorithmic approaches for performing patient alignment are feasible, but the performance of these algorithms is very dependent on the quality of the input data.

Surface Registration is a registration problem that handles the registration of two similar surfaces. In most research that utilises Deep Learning (DL) models to handle surface registration two theories are investigated; the first being whether surfaces sampled from the same origin can be registered together, and the second theory being whether the models can register Point Clouds with low overlapping data for utilisation in Simultaneous Localisation and Mapping (SLAM) applications. However, the surface registration to be utilised in the HoloNav Augmented Reality (AR) navigation system will utilise Point Clouds sampled from different origins with a high overlap ratio. This research, therefore, aims to determine the viability of DL methods for surface registration in HoloNav data. To determine the viability, rotation and translation errors in the match were used, with the aforementioned metrics later being evaluated manually with the utilisation of a visualiser. The results indicate that the models can generalise on the navigator data for an initial Euler angle difference of 45 degrees, but due to the difference in sampling density on the utilised point clouds can not provide accurate matches. Therefore, the utilisation of DL models can be considered to be viable if the navigator data has a sampling density similar to the pre-operative model.

Patient and instrument tracking are fundamental parts of surgical navigation systems. Traditional surgical navigation systems rely on stationary cameras for tracking and stationary screens for presenting information. An increased mental load is exerted by surgeons as they switch focus between the surgical site and the presented data. By projecting the required data on the patients anatomy, augmented reality surgical navigation would allow surgeons to navigate seamlessly during surgical procedures. In this paper, an augmented reality tracking algorithm for the Hololens 2 will be presented and tested against state of the art optical tracking. The algorithm is based on the use of reflective markers, which are extensively used in medical applications. An accuracy of 1.04mm was found in optimal conditions using the presented method. The presented tracking method is less accurate and less consistent than state of the art optical tracking methods with a minimum median accuracy of 25mm.

Learning to Rank is the application of Machine Learning in order to create and optimize ranking functions. Most Learning to Rank methods follow a listwise approach and optimize a listwise loss function which closely resembles the same metric used in the evaluation. Popular listwise loss functions such as nDCG, AP and nRBP do not have consistent bounds across topics and do not account for instance-difficulty. As a result, the loss score does not solely reflect the performance of the model but also depends on the instance properties. During training, each instance is assumed to be equally informative, while in reality, this informativeness might depend on the difficulty of the instance. In this thesis, we propose four bounding methods which utilize some notion of instance-difficulty to produce difficulty-aware losses. Experimental results showed that, in most cases, optimizing a bounded variant of nDCG, AP or nRBP results in a consistent but marginal increase in the overall performance. More interestingly, we found that optimizing a bounded variant of nRBP and AP may increase the nDCG@k score, increasing the recommendation utility. Overall, our results show promising results for user/query-wise metric bounding in Learning to Rank, especially when applied to nRBP.

In group recommendation, a key question is how preferences from individuals should be obtained and then aggregated into a group outcome. Collecting individual preferences can be done through implicit or explicit means, but there is insufficient research available on what option is optimal. For comparing different possible aggregation strategies, much of the existing research in the field departs from existing preference data (e.g.\ ratings), but considers synthetically created groups, rather than real groups. This study describes two experiments focusing on these issues. The first compared historical listening data with explicitly provided data and showed them to be similar. The second experiment compares different aggregation strategies in a more ecologically valid setting. More specifically, it considers playlist creation scenarios for groups, in which participants were asked to create their own, real-life groups. While the playlists as a whole did not show any significant differences in performance, inspecting the ratings of the tracks did show some differences. Finally, acoustic similarity was explored using the data obtained in the experiments. Some early results were observed, but more research is needed in this area.