CEUR workshop proceedings最新文献

The proposed method focuses on speaker disentanglement in the context of depression detection from speech signals. Previous approaches require patient/speaker labels, encounter instability due to loss maximization, and introduce unnecessary parameters for adversarial domain prediction. In contrast, the proposed unsupervised approach reduces cosine similarity between latent spaces of depression and pre-trained speaker classification models. This method outperforms baseline models, matches or exceeds adversarial methods in performance, and does so without relying on speaker labels or introducing additional model parameters, leading to a reduction in model complexity. The higher the speaker de-identification score (DeID), the better the depression detection system is in masking a patient's identity thereby enhancing the privacy attributes of depression detection systems. On the DAIC-WOZ dataset with ComparE16 features and an LSTM-only model, our method achieves an F1-Score of 0.776 and a DeID score of 92.87%, outperforming its adversarial counterpart which has an F1Score of 0.762 and 68.37% DeID, respectively. Furthermore, we demonstrate that speaker-disentanglement methods are complementary to text-based approaches, and a score-level fusion with a Word2vec-based depression detection model further enhances the overall performance to an F1-Score of 0.830.

Addressing the critical shortage of mental health resources for effective screening, diagnosis, and treatment remains a significant challenge. This scarcity underscores the need for innovative solutions, particularly in enhancing the accessibility and efficacy of therapeutic support. Embodied agents with advanced interactive capabilities emerge as a promising and cost-effective supplement to traditional caregiving methods. Crucial to these agents' effectiveness is their ability to simulate non-verbal behaviors, like backchannels, that are pivotal in establishing rapport and understanding in therapeutic contexts but remain under-explored. To improve the rapport-building capabilities of embodied agents we annotated backchannel smiles in videos of intimate face-to-face conversations over topics such as mental health, illness, and relationships. We hypothesized that both speaker and listener behaviors affect the duration and intensity of backchannel smiles. Using cues from speech prosody and language along with the demographics of the speaker and listener, we found them to contain significant predictors of the intensity of backchannel smiles. Based on our findings, we introduce backchannel smile production in embodied agents as a generation problem. Our attention-based generative model suggests that listener information offers performance improvements over the baseline speaker-centric generation approach. Conditioned generation using the significant predictors of smile intensity provides statistically significant improvements in empirical measures of generation quality. Our user study by transferring generated smiles to an embodied agent suggests that agent with backchannel smiles is perceived to be more human-like and is an attractive alternative for non-personal conversations over agent without backchannel smiles.

ImmPort, the world's largest repository of immunology data, includes many vaccine immune response datasets. ImmPort maps the metadata of these studies to ontology and database schema. As of February 28, 2023, our ImmPort data analysis identified 6.258 immune exposures using 47 vaccines in 4,607 human subjects, and 324 cohort studies from the ImmPort. We hypothesized that an integrative ontological representation of the data from these studies would enhance our understanding and analysis of these ImmPort vaccine studies, and with ontological classification and tools such as VIGET, we could further study the effects of different conditions such as vaccine types and host biological sex on the vaccine response gene expression profiles. Our Vaccine Ontology (VO) analysis classified these 37 vaccines into bacterial, viral, and protozoan vaccine types with different vaccine properties. The ImmPort metadata types were modeled with the Vaccine Investigation Ontology (VIO). Our new ontology-based pipeline extracted vaccine response data from the ImmPort database, annotated them based on ontology, obtained corresponding gene expression data from the GEO, and performed consistent omics data analysis. Our use case found gene profiles shared and differed from live and killed inactivated influenza vaccines. Furthermore, our Omics data analysis using the VIGET tool found that female and male human subjects have differential host responses for influenza vaccines. For example, our study showed much stronger early female responses to influenza vaccination than males, and males was able to show active immune responses at a later stage. Interestingly, the female (but not male) human subject group also showed significantly enriched neutrophil degranulation at Day 3 after influenza vaccination; however, males (but not females) displayed significantly enriched neutrophil degranulation at Day 14 after influenza vaccination. These mechanisms have been used to find differences between the gene lists and pathways of host responses to different vaccines conditional to different factors including vaccine types and host biological sex. Moreover, this framework can be expanded to other vaccines and vaccine categories easily.

Advances in computational behavior analysis via artificial intelligence (AI) promise to improve mental healthcare services by providing clinicians with tools to assist diagnosis or measurement of treatment outcomes. This potential has spurred an increasing number of studies in which automated pipelines predict diagnoses of mental health conditions. However, a fundamental question remains unanswered: How do the predictions of the AI algorithms correspond and compare with the predictions of humans? This is a critical question if AI technology is to be used as an assistive tool, because the utility of an AI algorithm would be negligible if it provides little information beyond what clinicians can readily infer. In this paper, we compare the performance of 19 human raters (8 autism experts and 11 non-experts) and that of an AI algorithm in terms of predicting autism diagnosis from short (3-minute) videos of N = 42 participants in a naturalistic conversation. Results show that the AI algorithm achieves an average accuracy of 80.5%, which is comparable to that of clinicians with expertise in autism (83.1%) and clinical research staff without specialized expertise (78.3%). Critically, diagnoses that were inaccurately predicted by most humans (experts and non-experts, alike) were typically correctly predicted by AI. Our results highlight the potential of AI as an assistive tool that can augment clinician diagnostic decision-making.

Adequately representing kinship relations is crucial for a variety of medical and biomedical applications. Several kinship ontologies have been proposed but none of them have been designed thus far in line with the Basic Formal Ontology. In this paper, we propose a novel kinship ontology that exhibits the following characteristics: (1) it is fully axiomatized in First Order Logic following the rules governing predicate formation as proposed in BFO2020-FOL, (2) it is modularized in 6 separate files written in the Common Logic Interface Format (CLIF) each one of which can be imported based on specific needs, (3) it provides bridging axioms to and from SNOMED CT, and (4) it contains an extra module with axioms which would not be literally true when phrased naively but are crafted in such a way that they highlight the unusual kinship relations they represent and can be used to generate alerts on possible data entry mistakes. We describe design considerations and challenges encountered.

Sparse decision trees are one of the most common forms of interpretable models. While recent advances have produced algorithms that fully optimize sparse decision trees for prediction, that work does not address policy design, because the algorithms cannot handle weighted data samples. Specifically, they rely on the discreteness of the loss function, which means that real-valued weights cannot be directly used. For example, none of the existing techniques produce policies that incorporate inverse propensity weighting on individual data points. We present three algorithms for efficient sparse weighted decision tree optimization. The first approach directly optimizes the weighted loss function; however, it tends to be computationally inefficient for large datasets. Our second approach, which scales more efficiently, transforms weights to integer values and uses data duplication to transform the weighted decision tree optimization problem into an unweighted (but larger) counterpart. Our third algorithm, which scales to much larger datasets, uses a randomized procedure that samples each data point with a probability proportional to its weight. We present theoretical bounds on the error of the two fast methods and show experimentally that these methods can be two orders of magnitude faster than the direct optimization of the weighted loss, without losing significant accuracy.

Ontologies have emerged to become critical to support data and knowledge representation, standardization, integration, and analysis. The SARS-CoV-2 pandemic led to the rapid proliferation of COVID-19 data, as well as the development of many COVID-19 ontologies. In the interest of supporting data interoperability, we initiated a community-based effort to harmonize COVID-19 ontologies. Our effort involves the collaborative discussion among developers of seven COVID-19 related ontologies, and the merging of four ontologies. This effort demonstrates the feasibility of harmonizing these ontologies in an interoperable framework to support integrative representation and analysis of COVID-19 related data and knowledge.