Understanding mutational processes active in cancer at the single-cell level is essential for characterizing intra-tumor heterogeneity. Previous studies extracted these processes, called mutational signatures, and the known signatures can be found in the Catalogue of Somatic Mutations in Cancer (COSMIC) database. These signatures were derived based on bulk sequencing data of thousands of whole genomes. This study proposes and applies a systematic method to compare single-cell-derived de novo mutational signatures to the COSMIC signatures. Using two single-cell cancer datasets (breast and neck cancer), two stable signatures were extracted per dataset. Within each dataset, the de novo signatures were extremely similar (cosine similarity > 0.96), suggesting uniform mutational processes within individual tumors. No direct one-to-one matches were found between de novo and COSMIC signatures. However, the de novo signatures can be interpreted as combinations of known mutational processes. These results demonstrate the feasibility of extracting de novo signatures based on single-cell data, while also highlighting limitations due to possible overfitting. Future work should include simulation experiments, analysis of additional tumors, and evaluation of alternative signature extraction methods.

Tumor heterogeneity complicates mutational signature analysis at the single-cell level, where sparse catalogues and uneven mutation burdens can destabilise exposure estimates. This study quantifies the robustness of fitted mutational signatures in single-cell RNA-seq data from 688 breast-cancer cells. Known COSMIC v3.4 SBS96 signatures were assigned with SigProfilerAssignment and the input data was systematically perturbed by randomly deleting 5%, 10%, 20% and 40% of mutations, repeating each perturbation twenty times. Robustness was assessed with four complementary metrics: (i) persistence of each signature in the dataset, (ii) stability of the number of cells containing each signature, (iii) mean relative error of persignature exposures, and (iv) per-cell cosine similarity between original and perturbed exposure vectors. Six signatures (SBS1, 5, 12, 26, 40c and 54) were consistently recovered, even after 40% deletion, demonstrating that core biological signals may survive substantial data loss. Nevertheless, higher deletion levels triggered progressive overfitting: the number of additional signatures rose from three at 5% deletion to eighteen at 40%. Exposures seemed to shift between highly similar signature pairs (e.g., SBS12 and SBS26, SBS5 and SBS40c), and merging such pairs halved the mean relative error. Signature SBS54, detected in only eight cells and suspected to be artefactual, showed the poorest stability. Across cells, robustness scaled positively with the number of mutations per cell (ρ ≈ 0.38 to 0.59) and negatively with entropy of the exposure vectors (ρ ≈ −0.27 to –0.53), indicating that abundant or signature-dominated catalogues resist perturbation, whereas sparse or evenly distributed ones are more fragile. Together, our results indicate that while some signatures and cells can survive substantial data loss, signature exposures in sparse single-cell catalogues must be interpreted with caution.

The field of oncology has greatly benefited due to the study of mutational signatures, pat terns of mutations that appear within the cancer genome. Previous research has focused its resources on utilizing various mathematical models to uncover and understand these mutational signatures by looking at the genetic information of aggregated cells, typically sequenced from a tumor biopsy, which is referred to as bulk data. However, recent developments in sequencing techniques have provided us with the possibility of investigating the genetic information at the single cell level rather than bulk. Thus, in this paper, we utilized machine learning-based tools to examine the effect of performing signature fitting at the single-cell level in relation to pseudo-bulk. We found that single cells have a higher degree of expression by contrast to the pseudo bulk, having the capability to identify a higher number of active mutational signatures. We also saw some single cells achieving better accuracy in the reconstruction of their mutational profile, by comparison to the pseudo-bulk. We identified that the heterogeneity across the single cells could be explained by a small number of clusters, which can potentially elucidate the active signatures found at the level of the pseudo-bulk sample. Finally, we found that some pseudo-bulk samples generated from subpopulations of cells unexpectedly deviate from the single cells which created them. From the findings, we believe that the study of active mutational signatures at the level of single cells has the potential to enlarge our understanding of cancer by providing us a more in-depth view of this disease. However, further research should be undergone in order to either augment or refute these findings, mainly due to the limiting factor of a relatively small number of mutations characterizing our data, together with the absence of a ground-truth for the bulk data.

Episodic memory (EM) -- the capacity to recall past experiences situated in time and context -- is a critical component of intelligent behavior. Although several cognitive architectures (CAs) have incorporated mechanisms inspired by episodic memory, implementations vary widely in structure, mechanisms, and integration with other cognitive functions. While prior work has reviewed episodic memory across a range of architectures in a high-level manner, detailed, structured comparisons among specific systems remain lacking. This study presents a focused comparative analysis of modeling episodic memory in two contrasting cognitive architectures: Soar, a symbolic, rule-based, general-purpose system, and Xapagy, a system designed specifically for narrative reasoning, relying on direct, unprocessed recordings of autobiographical events. By analyzing the representations, structures, and mechanisms of episodic memory in these two systems, this study highlights important design trade-offs and distinct assumptions about the role of episodic memory in cognition and its modeling approaches in CAs.

Understanding the mutational processes active within cancer cells is essential to improve diagnosis and treatment strategies. This study investigates whether the activity levels of these processes, quantified as mutational signature exposures, can be predicted from single-cell gene expression data. Two regression-based learning paradigms are compared: regular independent modelling, where the different models of each mutational signature selects its own regularisation parameter and set of genes, and multitask modelling, where the different models agree on a set of genes to be used for the prediction of each signature, and the regularisation parameter is shared. We evaluate their predictive performance and interpretability using biologically informed metrics. Furthermore, we assess the models’ robustness on unseen data by simulating real-world shifts through clustering-based data splits. Our results show that while both models achieve reasonable predictive accuracy, independently trained models offer greater flexibility and interpretability by identifying signature-specific genes and regularisation strengths. These findings suggest that gene expression carries meaningful information about a cell’s mutational history and that signature-specific modelling may offer better biological insight into tumour heterogeneity.

Heuristic strategies are an integral part of consumer decision-making. Heuristics serve as mental shortcuts that reduce cognitive effort, simplifying consumer decisions. To go from qualitative insights into these heuristics to quantitative data, a cognitive architecture must represent these heuristic strategies to understand consumer behavior better. This study will focus on the cognitive architectures ACT-R and CLARION since there is an interesting distinction in how they structure symbolic (explicit) and subsymbolic (implicit) cognition, influencing how they represent heuristic behavior. Currently, there is no systematic overview and comparison of how ACT-R and CLARION represent heuristics relevant to consumer decision-making. This paper aims to fill this knowledge gap by performing a systematic literature review on papers that contain an ACT-R or CLARION representation of heuristics relevant to consumer decision-making. The review uses four databases for the literature search: Scopus, Web of Science, IEEE Xplore, and ACM Digital Library. In total, 58 records have been screened, and 12 records have been included in the review. The review shows that ACT-R’s strength relies on representing heuristics by sequentially executing rule-based heuristics, while CLARION focuses on representing similarity- based heuristics by using bottom-up activation from its implicit layers. The results show a pattern in which the architectural structure mainly determines which heuristic strategies have been represented.

Understanding the relationship between mutational processes and gene expression patterns is essential for gaining insights into tumor heterogeneity. In this study, we analyze single-cell RNA sequencing data from a breast cancer tumor to investigate associations between mutational signature exposures and gene expression profiles. We propose a scoring method that integrates principal component loadings, canonical correlation analysis (CCA) loadings, and signature contributions to quantify gene-signature associations. Enrichment analysis of the top-ranking genes reveals consistent involvement of extracellular matrix (ECM) receptor interaction, focal adhesion, and immunerelated pathways across multiple mutational signatures. These findings suggest that different mutational processes converge on pathways involved in cell adhesion, invasion, and immune modulation. Our approach demonstrates the utility of multivariate statistical methods combined with enrichment analysis to explore the transcriptional consequences of mutational processes in cancer at the single-cell level.

This paper surveys nine studies that implement aspects of moral reasoning within cognitive architectures (CAs) or CA-inspired frameworks. Its primary aim is to assess the viability of this approach for future research and to clarify the state of the domain. Two research paradigms emerge: (1) modeling human moral reasoning and (2) constructing artificial moral agents. Despite this distinction, all studies face similar challenges: fragmented reuse (each employs a different architecture), limited pre-programmed behaviors, and the absence of standardized benchmarks or metrics. Researchers remain optimistic about the explainability of their systems' behaviors and inner workings, yet often they acknowledge significant scalability and validation hurdles. Overall, CAs currently support only small-scale experiments; substantial further research – both empirical and into the theoretical basis of the field – is needed before these systems can attain real-world relevance.

The emergence of Language Language Models (LLMs)-based agents represents a significant advancement in artificial intelligence (AI), offering new possibilities for complex problem-solving and interaction within a virtual environment. Our work is based on the Voyager paper [1], which is a state-of-the-art LLM-based agent for Minecraft. However, this system suffers from some significant limitations, such as its reliance on closed-source LLMs and lack of social awareness. Indeed, current open-source LLMs often fail to match closed-source ones in the agent setting, leaving research reliant on third-party closed-source technology [1] [2]. This gap highlights the need for alternative strategies to enhance LLM performance without the high costs associated with fine-tuning. To address these challenges, we propose the Collaborative Voyager, a new architecture designed to enable agent collaboration and social awareness using open-source LLMs. Inspired by the social intelligence hypothesis, which suggests that intelligence emerges from social interactions, we propose collaboration as an alternative learning paradigm for LLMs. This alternative learning paradigm could potentially supplement the expensive fine-tuning currently needed to bridge the performance gap between open-source and closed-source models in the agent setting [2]. Our approach involves developing a framework that allows agents to communicate, understand, and learn from each other, enabling them to correct errors and adapt to new tasks dynamically. By using a memory module, our agent is able to remember interactions and learn from them in order to accomplish a task that it was previously unable to do on its own. Through various experiments, we demonstrate that collaboration significantly enhances the performance of LLM agents in both task completion and adaptability, addressing issues like hallucinations. This study provides insights into developing more sophisticated, adaptable AI systems capable of dynamic interactions and problem-solving. These findings have potential applications extending beyond Minecraft and virtual environments to fields such as robotics, where collaboration and social awareness are crucial.

Emotional datasets for automatic affect prediction usually employ raters to annotate emotions or verify the annotations. To ensure the reliability of these raters some use interrater agreement measures, to verify the degree to which annotators agree with each other on what they rate. This systematic review explores what kind of interrater agreement measures are used in emotional speech corpora. The affective states, the affect representation schemes, and the collection method of the datasets as well as the popularity of these measures were investigated. Scopus, IEEE Xplore, Web of Science, and ACM digital library were used to extract papers that describe the creation of datasets. 45 papers were included in the review. The review concludes that the interrater agreement measures used, are highly dependent on the collection method for speech and the affect representation schemes. It was found that there is no standardized way to measure interrater agreement. Datasets that use actors to record emulated emotions mostly use recognition rate as their interrater agreement measures. Datasets that use a dimensional representation scheme often compute the mean agreement of the raters and the standard deviation of that measure to check the interrater agreement. Datasets that do not use actors nor are dimensional use a plethora of different measures such as probabilistic computing of agreement, or majority agreement measures, but a large amount use no measures at all.

With the rise in the number of human-computer interactions, the need for systems that can accurately infer and respond to users' emotions becomes increasingly important. One can achieve this by examining audio-visual signals, aiming to identify the underlying emotions from an individual's gestures, auditory cues, and surroundings. Such automatic affect prediction systems depend heavily on labeled datasets. However, the subjective nature of emotion interpretation often introduces uncertainty, making it challenging to create reliable and high-quality datasets. To combat this issue, researchers have tried employing multiple raters to judge the affective state of a person, while employing interrater agreement measures to monitor uncertainty. To this moment, it is still unclear to what extend does reaching a good level of interrater agreement impact the performance of audio-visual Automatic Affect Prediction models.
As a first step towards understanding the potential influences on performance, this paper conducts a systematic literature review to investigate the objective annotation procedure in audio-visual databases. The survey extracted relevant literature from 4 scientific databases: Scopus, IEEE Xplore, Web of Science and ACM Digital Library. The results are aggregated from 55 papers and presented by following the PRISMA guidelines. They indicate that most databases use multiple annotators, a little more than half measure interrater agreement, and most train the raters to increase the uniformity of the labels.

Human-computer interaction has long been the focus of technological evolution; however, in order for this type of system to reach its peak potential, machines must recognize that humans are constantly influenced by emotions. Text affective content analysis models are one attempt to integrate human psychology into computers, trying to detect the emotion transmitted by written input. There are numerous approaches to implementing such systems, with supervised learning still popular. The challenge of creating textual affective datasets is not in the availability of records, as humanity has reached a peak in data production, especially text, but in ensuring the consistency of the annotations provided by humans when included in the process. This study conducts a systematic literature review focused on providing details of published corpora. The annotation process, as well as any trends in how it evolved, will be examined to obtain dataset particularities. The ultimate intent is to lay the groundwork for an ample study aimed at analyzing the relationship between interrater agreement levels and performance scores of models trained on these datasets. The relevant literature was extracted from 3 search engines: Scopus, IEEE Xplore, and Web of Science, with a focus on manually labeled written records that are not part of multimodal systems, resulting in an analysis of 41 datasets. According to the aggregations, when humans are recruited to perform this task, researchers are more likely to use multiple annotators and calculate the degree of agreement between them to ensure the data's reliability before using it. The conducting researchers are inclined to either train these people before the procedure or tailor the set of labels to potentially increase the uniformity of ratings. As a result, this paper highlights the variety of annotation process characteristics and points towards standardizing this task.

This study explores the influence of interrater agreement measures and affect representation schemes in automatic affect prediction systems using physiological signals. These systems often use supervised learning and require unambiguous and objective labeling, a challenge when multiple human annotators are involved, which can affect model performance.
The research involved the first part of a two-stage process: systematically reviewing datasets and their characteristics concerning interrater agreement on the affective interpretation of physiological signals. This stage established a reliable foundation for the second step: a future analysis of model performance reported in technical papers utilizing these datasets. The main takeaways were that the number of raters varies significantly over datasets and the complexity introduced by combining affect representation schemes can negatively affect interrater agreement.

Understanding how users retrospectively evaluate their interactions with adaptive intelligent systems is crucial to improving their behaviours during interactions. Prior work has shown the potential to predict retrospective evaluations based on different real-time aspects of conversations, such as verbal cues and non-verbal behaviours. However, the relationship between how one retrospectively evaluates and the real-time evaluations in the moment of conversations remains unclear. This study investigates the relationship between real-time evaluations of a situation, using the Situational Interdependence Scale (SIS) framework, and its retrospective evaluations. We investigate the presence of the peak-end rule and a complex relationship that could be modelled using Long Short-Term Memory (LSTM) for each SIS dimension using the PACO dataset. Due to the absence of ground truth for real-time SIS evaluations, we also present a methodologically sound technical approach to utilize a Large Language Model (LLM) to estimate values for each SIS dimension for each spoken utterance in conversations. Analysis of the experiments revealed the absence of both the peak-end rule and an LSTM-modelled relationship across all dimensions of SIS. However, both types of models at least predict better than the average of the estimated real-time evaluation. This may be largely due to the inaccuracy of the estimated real-time SIS evaluations and the limited LLM’s capability of labelling real-time SIS in conversational data. Future works may focus on improving the annotation of real-time SIS evaluations through human annotation or human-supervised few-shot learning of LLM, using other modalities in combinations with verbal content, and exploring other predictive models.

Automatic speech recognition (ASR) is increasingly used in daily applications, such as voice-activated virtual assistants like Siri and Alexa, real-time transcription for meetings and lectures, and voice commands for smart home devices. However, studies show that even state-of-the-art (SotA) ASR systems do not recognize the speech of everyone equally well.

To the best of my knowledge, this paper, for the first time, evaluates the performance of Meta's SotA ASR system, Massively Multilingual Speech (MMS), on Dutch native and non-native speech. Using the Jasmin Corpus dataset, which includes a diverse set of both native and non-native Dutch speakers, this study uses metrics such as word error rate (WER), character error rate (CER), and word information lost (WIL) to assess performance. Additionally, the same methodology is applied to the same data using OpenAI's ASR system, Whisper, to provide a comparative analysis.

The paper analyzes WER, CER, and WIL error metrics, processing time, and investigates the best-suited beam size for Whisper. It also lists out the types of errors made in terms of deletions, insertions, and substitutions made by each model across different age groups of Dutch speakers.

Automatic Speech Recognition (ASR) systems have become increasingly important for society, yet their performance varies significantly across different diverse speaker groups. With a significant non-native population in the Netherlands, it is crucial that ASR systems accurately recognize diverse speech. Commercial state-of-the-art ASR systems are yet under-explored in their performance on Dutch diverse speech. This study evaluates the performance of two recently developed and affordable ASR systems, Google Chirp and OpenAI's Whisper, on speech from native and non-native Dutch teenagers. This research evaluates the recognition accuracy of these ASR systems and identifies common transcription errors. The results show slightly worse performance compared to previous research on non-native speech, and Whisper performing generally better than Google Chirp on the speaker groups.

Automatic Speech Recognition (ASR) systems are found in many places and are used by many people. Some groups of people, superficially older Dutch adults, are recognized less well by these systems. Given the aging population of the Netherlands, it would be beneficial to have ASR systems be more inclusive to allow for more independence of the older adults. By conducting tests on the ASR systems of Google and Microsoft, making use of the JASMIN dataset, I compared the two using word-error-rate (WER), word-information-lost (WIL) and character-error-rate (CER). Results show Microsoft outperforming Google with an average word error rate of 19.6\% compared to 27.35\%. However, Google is less biased on the topics of gender and age. Microsoft was slightly less biased in regards to region, but only by a small margin. Overall, the most notable findings from both systems are a small bias toward female speakers, and a strong bias against speakers from the southern regions of the Netherlands. These findings highlight the need for more inclusive ASR systems, enhancing the independence of older adults.

Automatic Speech Recognition (ASR) technology is becoming more and more useful in everyday life, therefor also requiring higher accuracy across all different user demographics. This study compares the performance of Google's and Microsoft's ASR systems on native Dutch child and teenager speech using the JASMIN-CGN dataset as ASR for children presents unique challenges due to their shorter vocal tracts and irregular speech patterns. This research evaluates each system's performance based on Word Error Rate (WER) and Character Error Rate (CER), highlighting the differences between gender, age, and dialect regions. The results indicate that while Microsoft's ASR consistently outperforms Google's in terms of WER, Google demonstrates slightly higher precision in terms of CER. Therefor Microsoft is considered the better overall performing system but depending on one's needs, such as precision, Google would be the more favorable one.

Existing content-based image retrieval models work well for natural photos, but not for images of architectural floor plans.
Previous work on floor plan retrieval has focused on graph-based methods, rather than image-based floor plans.
Training a CNN-based representation learning framework on segmented floor plan images with standard image augmentations does not result in semantically meaningful retrievals.
This work shows that a CNN-based representation learning model can learn features for retrieving floor plans that have similar graphs given the right training signal. Two methods were investigated here: GeomPerturb, a data augmentation that perturbs the underlying geometry of a floor plan, and a weakly supervised method with labels based on the graph edit distance between a pair of floor plans. The results show that while GeomPerturb learns representations that are correlated with the floor plan graph, training with GED labels leads to better retrievals both in terms of the floor plan graph and with respect to room shapes.

In the Netherlands, there is a shortage of primary school teachers, due to this shortage, teachers often do not have a lot of one-on-one time with the students. A social robot could be the solution to creating more one-on-“one” time with the students. In addition to relieving some workload for the teachers, a social robot could assist in gaining insight into personal values. Gaining insight into personal values can contribute to the learning experience by increasing motivation. The value awareness of the children is the learning objective as value awareness is thought to increase the child’s competence and motivation. One of the options to gain insight into contextual values is Socratic Questioning. These questions are created to further the learning experience and are expected to increase value awareness. This thesis investigates how a social robot can assist children in attaining learning goals, reflecting on these goals, and exploring the decisions and values underlying their learning experiences related to neurodiversity. The effectiveness of the robot in achieving learning objectives, encouraging reflection on decisions and gaining insight into personal moral values is shown. To gain insight into personal values, two methods of Socratic Questioning have been compared; Clarification and Implications & Consequences. The participating children of this thesis play an educational game where moral dilemmas are presented. A social robot assists in the game and asks the Socratic Questions. The explanation to the Clarification has proven to be the most effective method out of the two to gain insight into moral values. Moreover, the interaction effect between question type and time shows that the impact of the question type on insight into values varies with the time a question is asked. Qualitative observations show the participants’ anticipation of the robot’s questions and show instances where responses are different from the self-reported value importance. Overall, this research sheds light on the potential of Socratic Questioning and its uses for understanding individual values, calling for further exploration in this area.