Dr.-Ing. Tomás Arias Vergara

Lehrstuhl für Informatik 5 (Mustererkennung)

Research associates

Address

Martensstraße 3
91058 Erlangen
10.134  10

Tomás Arias Vergara

2025

  • Israel ISF-DFG: The effect of social interactions on (dis)honest communication

    (Third Party Funds Single)

    Project leader:
    Term: since September 1, 2025
    Funding source: DFG-Einzelförderung / Sachbeihilfe (EIN-SBH)

    Animals make decisions based on cues that they receive from conspecifics. Acoustic cues, encoded in vocal frequencies, tempo, and syntax, convey honest information on, for example, the age and social status of the singer. Yet numerous studies in birds, anurans, and humans show that in social interactions (e.g., duets and counter-singing), acoustic features shift, compared to solo singing. This raises intriguing questions: How do social interactions affect honesty? Do they prompt a dishonest display of traits or a correction to display honest ones? Here, we propose to elucidate whether and how these social-context-affected changes reflect the honest depiction of individual traits, a topic that remains underexplored. For over 25 years, we have been studying the complex songs of male wild rock hyraxes (Procavia capensis), focusing on solo songs (performed spontaneously) and counter-singing (e.g. induced by other male songs). Male hyrax songs include a challenging sound called the "snort," a harsh sound developed with age, which varies with weight, social status, and hormone levels. While prior research shows that snorts differ between solo and counter-singing, their honesty in reflecting the singer’s traits wasn't assessed. This study aims to evaluate whether counter-singing represents the responder's traits than solo singing and to understand how social factors like weight and status affect these acoustic differences. To this end, we will adapt/implement novel methods originally designed for monitoring orcas (Orcinus orca), including denoising, segmentation, encoding, and generating artificial calls. Hyrax songs will be analyzed using deep learning methods to cluster the songs; thus, enabling automatic identification of distinct singers’ vocalizations (including ones we fail to observe in the field). Furthermore, new songs will be synthesized using generative AI to examine counter-singing behavior in the field. The effect of social context on the honest transmission of individual traits will be measured using acoustic analysis by contrasting the responses with the singer's solo songs. We hypothesize that vocal behavior will change according to both the initiating singer’s (or playback’s) and responder’s individual characteristics. This study is expected to elucidate whether social constraints generate honesty and expand our understanding of the contribution of sociality to the evolution of vocal communication in nature.

2024

  • A multimodal approach for automatic generation of radiology reports using chest X-ray images, clinical free-text, and spoken commands.

    (FAU Funds)

    Project leader:
    Term: January 15, 2024 - January 14, 2025

    Advancements in Artificial Intelligence (AI) methods have enabled thedevelopment of Large Language Models (LLMs) capable of generating informationfrom user instructions and supporting various tasks in education, research,healthcare, and others. AI has also impacted the field of medical imaging withseveral deep learning models capable of achieving expert-level performanceacross different tasks, e.g., detection, segmentation, and assisted clinicaldiagnosis. In addition, open-source Automatic Speech Recognition (ASR) systemscan be incorporated as modules in AI-based systems. This proposed fundedproject aims to combine LLMs, medical imaging, and speech recognition using AImethods to generate high-quality radiology reports from chest X-ray images.

  • Coordinated grid protection based on machine learning methods

    (Third Party Funds Single)

    Project leader: , ,
    Term: July 1, 2024 - June 30, 2027
    Acronym: Netzschutz-KI
    Funding source: DFG-Einzelförderung / Sachbeihilfe (EIN-SBH)

2017

  • Training Network on Automatic Processing of PAthological Speech

    (Third Party Funds Group – Overall project)

    Term: November 1, 2017 - October 31, 2021
    Acronym: TAPAS
    Funding source: Innovative Training Networks (ITN)
    URL: https://www.tapas-etn-eu.org/

    There are an increasing number of people across Europe with debilitating speech pathologies (e.g., due to stroke, Parkinson's, etc). These groups face communication problems that can lead to social exclusion. They are now being further marginalised by a new wave of speech technology that is increasingly woven into everyday life but which is not robust to atypical speech. TAPAS is a Horizon 2020 Marie Skłodowska-Curie Actions Innovative Training Network European Training Network (MSCA-ITN-ETN) project that aims to transform the well being of these people.
    The TAPAS work programme targets three key research problems:
    (a) Detection: We will develop speech processing techniques for early detection of conditions that impact on speech production. The outcomes will be cheap and non-invasive diagnostic tools that provide early warning of the onset of progressive conditions such as Alzheimer's and Parkinson's.
    (b) Therapy: We will use newly-emerging speech processing techniques to produce automated speech therapy tools. These tools will make therapy more accessible and more individually targeted. Better therapy can increase the chances of recovering intelligible speech after traumatic events such a stroke or oral surgery.
    (c) Assisted Living: We will re-design current speech technology so that it works well for people with speech impairments and also helps in making informed clinical choices. People with speech impairments often have other co-occurring conditions making them reliant on carers. Speech-driven tools for assisted-living are a way to allow such people to live more independently.
    TAPAS adopts an inter-disciplinary and multi-sectorial approach. The consortium includes clinical practitioners, academic researchers and industrial partners, with expertise spanning speech engineering, linguistics and clinical science. All members have expertise in some element of pathological speech. This rich network will train a new generation of 15 researchers, equipping them with the skills and resources necessary for lasting success.

2026

Conference Contributions

Unpublished Publications

2025

Journal Articles

Conference Contributions

2024

Journal Articles

Conference Contributions

2023

Authored Books

Journal Articles

Conference Contributions

2022

Authored Books

Journal Articles

Conference Contributions

2021

Journal Articles

Conference Contributions

2020

Journal Articles

Conference Contributions

2019

Book Contributions

Conference Contributions

2018

Journal Articles

Conference Contributions

2017

Authored Books

Conference Contributions

2016

Conference Contributions

2022

  • : Junior researcher status (Ministry of Science, Technology, and Innovation (Colombia)) – 2022
  • : Summa cum laude (Doctoral thesis) (Friedrich-Alexander Universität Erlangen-Nürnberg (Germany) & Universidad de Antioquia (Colombia)) – 2022
  • : GI-Dissertation price nominee (Friedrich-Alexander Universität Erlangen-Nürnberg) – 2022

2018

  • : Early Stage Researcher under Marie Sklodowska-Curie grant (European Union’s Horizon 2020) – 2018

2017

  • : National PhD scholarship program (Colciencias (Colombia)) – 2017
  • : Distinction to Master thesis (Universidad de Antioquia (Colombia)) – 2017

2015

  • : Young researchers and innovators scholarship (Colciencias (Colombia)) – 2015

Current Theses & Projects

Title Type Student Period Status
Text-based Cross-Lingual Emotion Recogntion using Natural Language Processing Methods BA thesis Dila S. Celikkol Apr 2026 – Oct 2026 running
Adaptive Hybrid Deep Learning modal for Personalized Electric Vehicle Energy Consumption Prediction with Continuous Learning MA thesis Neel Thakkar Mar 2026 running
Large Language Models for Surgical Workflow Monitoring and Summarization MA thesis Daoqi Jin running
Transfer learning Based Forecasting Of Heat Pump Energy Consumption Across Multiple Time Horizons MA thesis Dhruvil Kalubhai Kalathiya Aug 2025 running

Completed Theses & Projects

Title Type Student Period Status
Automatic Prediction of German Regional Accents MA thesis Veronika Stengl finished
Investigating the Influence of Different Motion Sensors for Detecting Parkinson’s Disease MA thesis Mohammad Hamza Sep 2025 finished
Reduction of die trials via machine learning approaches MA thesis Tai Hoang Nguyen Sep 2025 – Mar 2026 finished
Large Language Models for Modified Frenchay Dysarthria Assessment Reports from Parkinson’s Speech: Model Choice and Prompting Effects MA thesis Zixuan Chai Sep 2025 – Mar 2026 finished
Sequence-Based Deep Learning for Endovascular Device Segmentation in Interventional X-ray Imaging MA thesis Sleiman Sharara Sep 2025 – Mar 2026 finished
Parkinson’s Disease Classification from Smartwatch Inertial Measurement Unit (IMU) Signals Across Structured Motor Tasks MA thesis Emin Mammadov Jul 2025 finished
Analysis of Speech Production Assessment of Cochlear Implant Users MA thesis Tejashree Dhawle Jul 2025 finished
PaiChat: A Visual – Language Assistant for Histopathology MA thesis Bhavanikbhai Kanani Mar 2025 – Dec 2025 finished
Pathological Voice Analysis with Selective State Space Models MA thesis Lucca Baumgärtner Jun 2025 – Dec 2025 finished
Interpretable Vision Transformers with Attention Maps for Phonological Precision Assessment from MRI Project finished
Heart sound detection using audio fingerprint MA thesis Shayan Alvandnyia May 2025 – Dec 2025 finished
Automatic Assessment of Parkinson’s Disease Using Audio and Text Analyses MA thesis Zhipeng Peng Mar 2025 – Sep 2025 finished
Removing age bias in the context of pathological speech MA thesis Yuhan Gao Mar 2025 – Sep 2025 finished
Influence of Demographic Parameters in Radar-Based Blood Pressure Estimation MA thesis Felix Tobias Büppelmann Dec 2024 – May 2025 finished
Deep Learning-Based Classification of Skin Diseases: A Comparative Analysis of CNN and Transformer Architectures Project Sleiman Sharara finished
Influence of Age in Neural Embeddings to Analyze Voice Disorders of Parkinson’s Disease Patients Project Zixuan Chai finished
Generative Modeling for Glottal Signals Synthesis MA thesis Su Wu Jan 2025 – Jul 2025 finished
Enhancing Lithium-Ion Battery Safety MA thesis Youssef Bouraha Dec 2024 – Jun 2025 finished
Generation of Region-guided Clinical Text Reports from Chest X-Ray Images Using LLMs MA thesis Mohammad Hasan Dec 2024 – Jun 2025 finished
Stammering Identification using Large Language Models MA thesis Aagam Sunilbhai Shah Nov 2024 – Apr 2025 finished
Annotation by Speech in Radiology MA thesis Jan Geier finished
Investigating Liquidity Forecasting with Point-Based and Probabilistic Models to Enhance Financial Business Operations MA thesis Ram Saran Kakumanu Oct 2024 – Mar 2025 finished
Enhancing SBOM Creation with Large Language Models MA thesis Gaurav Bhalala Nov 2024 – May 2025 finished
Signal-Specific Fault Detection in Controller Area Network using Deep Learning MA thesis Vamsi Krishna Chalampalem Nov 2024 – May 2025 finished
Knowledge Distillation of Large Language Models for Automotive HMI Applications MA thesis Aravind Ryali Nov 2024 – May 2025 finished
Automatic Speech Recognition at Phoneme and Word-Level To Analyze Parkinson’s Disease BA thesis Malena Grimm Piquer Nov 2024 – Apr 2025 finished
Speech-Based Classification of Parkinson’s Disease Under Acoustic Variability MA thesis Anisha Bhandare Aug 2024 – Feb 2025 finished
Large Language Models for Knowledge Management in Engineering Projects MA thesis Xinyuan Tu Oct 2024 – Apr 2025 finished
Identification of failure detection patterns in log files of Computer Tomography systems MA thesis Aishwarya Tandel Oct 2024 – Apr 2025 finished
TSI Challenge Summer 2024: Heat & Water Demand Forecasting Project Apr 2024 – Aug 2024 finished
Text Generation in Alzheimer’s Disease MA thesis Mahmoud Alimizel Jul 2024 – Jan 2025 finished
Improving Text Summarization through Guided Decoding of Language Models MA thesis Jannick Gluch Jul 2024 – Jan 2025 finished
Spoken Language Identification for Hearing Aids MA thesis Mahmoud G. A. Sanad Feb 2024 – Aug 2024 finished
Understanding Odor Descriptors through Advanced NLP Models and Semantic Scores MA thesis Fatma Mami Feb 2024 – Aug 2024 finished
Generation of Clinical Text Reports from Chest X-Ray Images Project Md Hasan Feb 2024 – Aug 2024 finished
Cross-Dataset Phonological Speech Analysis of Children with Cleft Lip and Palate MA thesis Marta López-Brea García Dec 2023 – Jun 2024 finished
Automatic recognition of bavarian dialects Project Veronika Stengl Nov 2023 – Apr 2024 finished
Large Language Model for Generation of Structured Medical Report from X-ray Transcriptions MA thesis Uttam Asodariya Sep 2023 – Mar 2024 finished
Natural Language Text Generation for Symbolic Descriptions Using Language Models MA thesis Deepak Parappagoudar Aug 2023 – Dec 2023 finished
Development of a deep learning approach to detect faulty axial bearing components after assembly using acoustic signals MA thesis Gracia Apfelthaler Aug 2023 – Feb 2024 finished
Edge-AI: Self-sensing backpressure estimation in piezoelectric micropumps using machine learning methods on a limited hardware MA thesis Mohammadhossien Sheikhsarraf May 2023 – Nov 2023 finished
CoachLea: An Android Application to evaluate the progress of speaking and hearing abilities of children with Cochlear Implant BA thesis Paula Schäfer Jun 2021 – Nov 2021 finished
CITA: An Android-based Application to Evaluate the Speech of Cochlear Implant Users BA thesis Christoph Popp Jul 2020 – Dec 2020 finished