Personal Bibliography

[2026-01-15 Thu]

This paper introduces a human-centered stress management system that integrates uncertainty-aware physiological stress detection with spoken dialogue-based interaction. Stress levels are inferred from various physiological signals using random forest and convolutional neural network classifiers with uncertainty quantification achieved using entropy-based and Monte Carlo dropout, respectively. Depending on the sytem’s confidence, either a stress-reducing intervention such as guided breathing is offered or the user is asked to confirm the prediction. The research is relevant to researchers and practitioners in physiological computing, human-robot interaction and assistive technologies, particularly those looking to deploy such a system in high-pressure environments.

@InProceedings{sheth-and-others--2026--uncertainty-aware-stress-detection,
  author =       {Sheth, Prachi and Schneider, Jordan and Hassan,
                  Teena},
  title =        {Integrating Uncertainty-Aware Stress Detection with
                  Spoken Dialogue-Based Interaction for Human-Centered
                  Stress Management},
  booktitle =    {Proceedings of the {AHFE} International Conference
                  on Intelligent Human Systems Integration},
  volume =       200,
  year =         2026,
  publisher =    {{AHFE} Open Access},
  doi =          {10.54941/ahfe1007120}
}

[2025-10-12 Sun]

This paper presents a doctoral research project using a design science research approach to develop a physiologically adaptive cognitive assistive system to support individuals with mild cognitive impairment in sheltered workshops.

@InProceedings{schneider--2025--physiologically-adaptive-cas,
  author =       {Schneider, Jordan},
  title =        {Towards Intelligent Adaption in Cognitive Assistance
                  Systems through Physiological Computing},
  booktitle =    {Proceedings of the 27th International Conference on
                  Multimodal Interaction},
  year =         2025,
  series =       {ICMI '25},
  pages =        {749--753},
  address =      {New York, NY, USA},
  publisher =    {ACM Press},
  doi =          {10.1145/3716553.3750828}
}

[2025-05-23 Fri]

This whitepaper presents guidelines for handling research data when researching and developing pro-adaptive cognitive assistive technologies (pro-CAT), drawn from the experience gained during the project Center for Assistive Technologies Rhein-Ruhr (Zentrum Assistive Technologien Rhein-Ruhr, ZAT). Systems that fall under the term pro-CAT are those that intelligently adapt the assistance they provide to the needs of the user (i.e., participant in the context of collecting research data). Common bilingual (English and German) terminology and a classification for research data are introduced and methods and tools for validating data in CSV, XML, audio and video formats are proposed. This would be a useful starting point for interdisciplinary teams working with a mixture of qualitative and quantitative data in human-computer interaction scenarios to gain an overview of best practices for handling these types of research data.

@Misc{ferger-and-others--2025--zat-research-data-guidelines-for-pro-cat,
  author =       {Ferger, Anne and Grashof, Robin and Krause, André
                  Frank and Schmidt, Thomas and Schneider, Jordan and
                  Yavuz, Sinan},
  title =        {{ZAT}: {Guidelines} for Research Data in Research
                  and Development Processes of Pro-adaptive Cognitive
                  Assistive Technologies (Pro-{CAT})},
  year =         2025,
  month =        apr,
  publisher =    {Zenodo},
  version =      {1.0.0},
  doi =          {10.5281/zenodo.15187761}
}

[2024-10-05 Sat]

In this mini-review, an overview of explainable artificial intelligence (XAI) is provided and papers using XAI techniques with physiobehavioural signals in practical applications are reviewed. The deficiencies of the current literature with respect to explanations for machine learning models are discussed and used to derive the DEEP Principles, which stands for define the audience of explanations, engage stakeholders to obtain feedback, embed domain knowledge into the model and prove the validity of explanations. The authors argue that future work should aim to involve the stakeholders more when designing explanations and rigorously evaluate these explanations against objective metrics. This paper might interest researchers who are concerned about the explainability of models trained on time-series data.

@InProceedings{schneider-and-others--2024--time-for-an-explanation,
  author =       {Schneider, Jordan and Cheruvalath, Swathy Satheesan
                  and Hassan, Teena},
  title =        {Time for an Explanation: {A} Mini-Review of
                  Explainable Physio-Behavioural Time-Series
                  Classification},
  booktitle =    {Companion of the 2024 on {ACM} International Joint
                  Conference on Pervasive and Ubiquitous Computing},
  year =         2024,
  series =       {UbiComp '24},
  location =     {Melbourne VIC, Australia},
  pages =        {885--889},
  address =      {New York, NY, USA},
  publisher =    {ACM Press},
  doi =          {10.1145/3675094.3679001}
}