Influence of speaker de‐identification in depression detection

López-Otero, Paula; Magariños, Carmen; Docío-Fernández, Laura; Rodriguez-Banga, Eduardo; Erro, Daniel; Garcı́a-Mateo, Carmen

doi:10.1049/iet-spr.2016.0731

Cited by 8 publications

(4 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With a focus on the signi cance of model interpretability, a hybrid model attained an impressive 99.65% accuracy, demonstrating the potential of arti cial intelligence to transform agricultural practices and improve food security. A study by Lopez-Otero et al (2017) [17] examined how speaker deidenti cation affected depression diagnosis, showing how voice alteration affects automated systems' accuracy. Their research indicates that, in voice-based depression detection systems, there should be a careful balance struck between privacy and diagnostic e cacy.…”

Section: Related Workmentioning

confidence: 99%

Enhancing Depression Predictive Models: a Comparative Study of Hybrid Ai, Machine Learning and Deep Learning Techniques

Kanchapogu,

Mohanty

2024

Preprint

View full text Add to dashboard Cite

In the quest to enhance predictive models for depression, this study introduces a novel comparative analysis of machine learning (ML) and deep learning (DL) techniques, further innovating with the development of hybrid AI models. Leveraging a dataset comprising 2,000 participants, enriched with demographic, socio-economic, behavioral, and clinical variables, including pre- and post-treatment Montgomery-Åsberg Depression Rating Scale (MADRS) scores, we embarked on a comprehensive exploration of factors influencing depression outcomes. Through meticulous data collection, we harmonized diverse variables ranging from basic demographic details to intricate clinical outcomes, paired with a rigorous feature selection process employing Recursive Feature Elimination (RFE) and Principal Component Analysis (PCA). Our analytical journey was underpinned by a robust hyper parameter tuning phase, ensuring the optimization of each model's predictive capacity. The study's core contribution lies in its exhaustive comparison of standalone ML and DL models against our crafted hybrid AI models, revealing a marked superiority of the latter in predicting depression with unparalleled accuracy. The hybrid models, through their synergetic integration of ML and DL methodologies, demonstrated a profound ability to navigate the complexity of depression's multifactorial nature, achieving a perfect prediction accuracy rate in our tests. Our findings advocate for a paradigm shift in predictive analytics for depression, underscoring the potential of hybrid AI models in transcending the limitations of traditional standalone approaches. This research not only paves the way for more nuanced and effective predictive tools in mental health care but also sets a benchmark for future studies at the intersection of biology, psychology, and artificial intelligence. The implications of this work are vast, offering a beacon of hope for personalized and preemptive mental health interventions.

show abstract

Section: Related Workmentioning

confidence: 99%

Enhancing Depression Predictive Models: a Comparative Study of Hybrid Ai, Machine Learning and Deep Learning Techniques

Kanchapogu,

Mohanty

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Papers (Count and Sources) Voice 22 ([2], [5], [4], [20], [74], [64], [53], [66], [42], [61], [33], [62], [54], [24], [63], [13], [32], [65], [10], [29], [85]) Face 10 ( [55], [48], [17], [19], [57], [18], [49], [7], [35], [36]) Gait 8 ( [3], [25], [28], [31], [45], [78], [79], [80]) Brain Activity 2 ([90], [44])…”

Section: Traitmentioning

confidence: 99%

A False Sense of Privacy: Towards a Reliable Evaluation Methodology for the Anonymization of Biometric Data

Hanisch,

Todt,

Patino

et al. 2024

PoPETs

View full text Add to dashboard Cite

Biometric data contains distinctive human traits such as facial features or gait patterns. The use of biometric data permits an individuation so exact that the data is utilized effectively in identification and authentication systems. But for this same reason, privacy protections become indispensably necessary. Privacy protection is extensively afforded by the technique of anonymization. Anonymization techniques protect sensitive personal data from biometrics by obfuscating or removing information that allows linking records to the generating individuals, to achieve high levels of anonymity. However, our understanding and possibility to develop effective anonymization relies, in equal parts, on the effectiveness of the methods employed to evaluate anonymization performance. In this paper, we assess the state-of-the-art methods used to evaluate the performance of anonymization techniques for facial images and for gait patterns. We demonstrate that the state-of-the-art evaluation methods have serious and frequent shortcomings. In particular, we find that the underlying assumptions of the state-of-the-art are quite unwarranted. State-of-the-art methods generally assume a difficult recognition scenario and thus a weak adversary. However, that assumption causes state-of-the-art evaluations to grossly overestimate the performance of the anonymization. Therefore, we propose a strong adversary which is aware of the anonymization in place. This adversary model implements an appropriate measure of anonymization performance. We improve the selection process for the evaluation dataset, and we reduce the numbers of identities contained in the dataset while ensuring that these identities remain easily distinguishable from one another. Our novel evaluation methodology surpasses the state-of-the-art because we measure worst-case performance and so deliver a highly reliable evaluation of biometric anonymization techniques.

show abstract

“…Glackin et al [96] propose symbolic encoding with an acoustic model on the client's side, then, the data are sent encrypted to the server. Lopez et al [97] claim that de-identification with frequency warping and amplitude scaling for depressed speech yields "promising deidentification results at the expense of a slight degradation of depression detection". Encryption for privacy-preserving paralinguistic mining is presented in [98], combined with Support Vector Machines, for emotion recognition, and in [99], combined with neural networks, for health-related tasks.…”

Section: Anonymisation and De-anonymisation Of Speech And Text Datamentioning

confidence: 99%

Ethics and Good Practice in Computational Paralinguistics

Batliner

Hantke

Schuller

2022

IEEE Trans. Affective Comput.

View full text Add to dashboard Cite

With the advent of 'heavy Artificial Intelligence' -big data, deep learning, and ubiquitous use of the internet, ethical considerations are widely dealt with in public discussions and governmental bodies. Within Computational Paralinguistics with its manifold topics and possible applications (modelling of long-term, medium-term, and short-term traits and states such as personality, emotion, or speech pathology), we have not yet seen that many contributions. In this article, we try to set the scene by (1) giving a short overview of ethics and privacy, (2) describing the field of Computational Paralinguistics, its history and exemplary use cases, as well as (de-)anonymisation and peculiarities of speech and text data, and (3) proposing rules for good practice in the field, such as choosing the right performance measure, and accounting for representativity and interpretability.

show abstract

Influence of speaker de‐identification in depression detection

Cited by 8 publications

References 31 publications

Enhancing Depression Predictive Models: a Comparative Study of Hybrid Ai, Machine Learning and Deep Learning Techniques

Enhancing Depression Predictive Models: a Comparative Study of Hybrid Ai, Machine Learning and Deep Learning Techniques

A False Sense of Privacy: Towards a Reliable Evaluation Methodology for the Anonymization of Biometric Data

Ethics and Good Practice in Computational Paralinguistics

Contact Info

Product

Resources

About