SURVEY – THE SEVEN DEADLY SINS OF AI IN MEDICINE

/in , /by
SURVEY – THE SEVEN DEADLY SINS OF AI IN MEDICINE

We invite you to participate in a crucial survey examining the key ethical challenges in medical AI.

Your insights are invaluable in shaping a future where AI enhances healthcare responsibly.

This survey will take only approx 4 minutes of your time, yet it addresses issues of paramount importance that affect us all.

https://ec.europa.eu/eusurvey/runner/seven-sins-of-medical-ai

Thank you very much,

Heimo MUELLER, Vimla L. PATEL, Edward H. SHORTLIFFE, Andreas HOLZINGER

Enhancing trust in automated 3D point cloud data interpretation through explainable counterfactuals

/in , , /by

Our most recent paper introduces a novel framework for augmenting explainability in the interpretation of point cloud data by fusing expert knowledge with counterfactual reasoning. Given the complexity and voluminous nature of point cloud datasets, derived predominantly from LiDAR and 3D scanning technologies, achieving interpretability remains a significant challenge, particularly in smart cities, smart agriculture, and smart forestry. This research posits that integrating expert knowledge with counterfactual explanations – speculative scenarios illustrating how altering input data points could lead to different outcomes – can significantly reduce the opacity of deep learning models processing point cloud data. The proposed optimization-driven framework utilizes expert-informed ad-hoc perturbation techniques to generate meaningful counterfactual scenarios when employing state-of-the-art deep learning architectures. Read the paper here:  https://doi.org/10.1016/j.inffus.2025.103032   and get an overview by listening to this podcast 🙂

 

Graph Neural Networks with the Human-in-the-Loop > Trustworthy AI

/in , , , /by

In our Nature Scientific Reports paper we introduce a novel framework – our last deliverable to the FeatureCloud project – that integrates federated learning with Graph Neural Networks (GNNs) to classify diseases, incorporating Human-in-the-Loop methodologies. This advanced framework innovatively employs collaborative voting mechanisms on subgraphs within a Protein-Protein Interaction (PPI) network, situated in a federated ensemble-based deep learning context. This methodological approach marks a significant stride in the development of explainable and privacy-aware Artificial Intelligence, significantly contributing to the progression of personalized digital medicine in a responsible and transparent manner. Read the article here https://doi.org/10.1038/s41598-024-72748-7 and get an overview by listening to this podcast:

 

Your Human-AI Co-Existence (5 Minutes Survey)

/in , /by
The Future of AI starts with You
If we want to achieve a more sustainable and AI-friendly future, we need to start with your individual participation.

Your responses are anonymous and your personal data will not be recorded.

How can you participate?

By filling out this five minute-long and anonymous survey, you can help us in making AI technology more accessible and understandable:

https://forms.gle/DTHmeD9v6XbwXeFn9

Why is that important?

Establishing adaptable and interpretable AI machinery is crucial for individuals and governments to catch up with the speed of technology. Key is not promoting solely development-friendly AI and regulatory overseeing frameworks, but rather working on transparency and readability of technologies through insightful guidelines so that participation for the individual is made possible. This includes the topic of informational self-determination through open legislation frameworks, policies, and ethical guidelines. Both the collective and individual aspect are important for AI technology progression, but a future towards sustainable-friendly AI as an enabler of the 17 sustainable development goals (SDGs) and targets rather than an inhibitor starts with open participation and constant confrontation of the individual with AI technology. One global example that affects us all is ongoing climate change [2], and here we need AI – and the workhorse machine learning (ML) – to contribute to what is clearly the greatest challenge facing humanity. Each and every one of us can contribute to the global challenges of climate change, and we want to explore how AI can help us do that.

[1] Vinuesa, R., Azizpour, H., Leite, I., Balaam, M., Dignum, V., Domisch, S., Felländer, A., Langhans, S. D., Tegmark, M. & Fuso Nerini, F. 2020. The role of artificial intelligence in achieving the Sustainable Development Goals. Nature communications, 11, (233), 1–10, https://doi.org/10.1038/s41467-019-14108-y

[2] Rolnick, D., Donti, P.L., Kaack, L.H., Kochanski, K., Lacoste, A., Sankaran, K., Ross, A.S., Milojevic-Dupont, N., Jaques, N. & Waldman-Brown, A. (2022). Tackling climate change with machine learning. ACM Computing Surveys (CSUR), 55, (2), 1–96, https://doi.org/10.1145/3485128

[3] This page is: https://human-centered.ai/2023/09/21/human-ai-5-minutes-survey

 

Thank you very much:

Andreas HOLZINGER, Heimo MUELLER, Jianlong ZHOU, Fang CHEN

Human Centered AI Lab Austria and Human-Centered AI Lab Australia

 

Your Views on ChatGPT in Applications (3 Minutes Survey)

/in , /by

The current development in Large Language Models is good for the machine learning community because it demonstrates the state of the art in statistical learning in an easy to understand way. For example, ChatGPT can fluently answer questions from users. It produces human-like texts with a seemingly logical connection between different sections. According to recent reports, individuals have already used ChatGPT extensively to formulate university essays, write scientific articles with references, debug computer programme code, compose music, write poetry, submit restaurant reviews, create advertising copy and solve exams, co-author magazine articles, and much, much more.
Despite the apparent benefits of ChatGPT, many human users have various ethical concerns about misinformation, transparency, privacy and security, bias, abuse, loss of jobs, lack of originality, over-dependence and even massive job loss.

In our survey, we want to know your views and concerns about ChatGPT so that we can summarise recommendations to users when they use ChatGPT in applications.

Your responses are anonymous and your personal data will not be recorded.

Please take part in our 3 minutes survey:

https://forms.gle/cYMzDyTT7UUP9wRi7

Thank you very much:

Jianlong ZHOU, Heimo MUELLER, Andreas HOLZINGER, Fang CHEN

Human-Centered AI Lab Australia and Human Centered AI Lab Austria

Usability Evaluation of Interactive XAI platform for Graph Neural Networks

/in /by

Lack of trust in artificial intelligence (AI) models in medicine is still the key blockage for the use of AI in clinical decision support systems (CDSS). Although AI models are already performing excellently in medicine, their black-box nature entails that it is often impossible to understand why a particular decision was made. In the field of explainable AI (XAI), many good tools have already been developed to “explain” to a human expert, for example, which input features influenced a prediction. However, in the clinical domain, it is essential that these explanations lead to some degree of causal understanding by a clinician in the context of a specific application. For this reason, we have developed an interactive XAI platform that allows the domain expert to ask manual counterfactual (“what-if”) questions. CLARUS allows the expert to observe how changes based on their questions affect the AI decision and the corresponding XAI explanation [1].

Please help us now with a usability evaluation and spend 10 minutes and go through this:

https://survey.medunigraz.at/index.php/368984?lang=en

and please fill out all fields (please include all feedback you think is necessary into the boxes),

please note that there will be TWO Windows open: one is the application and one is the questionaire,
so maybe it is better to open them in two separate windows for your convienience,

thank you very much

[1] https://doi.org/10.1101/2022.11.21.517358

Human-Centered AI for smart farming BOKU Tulln March, 9, 2023

/in , , /by

On March, 9, 2023 at the BOKU Tulln, we were guest speakers at the traditional Schlumberger lectures. For us a wonderful opportunity to show what Human-Centered AI can do for smart farming. Thanks to the organizers Michaela Griesser and Astrid Forneck from the Department of Crop Sciences (DNW) lead by Hans-Peter Kaul. Looking forward to help to discover the causality of berry shrivel (Traubenwelke) with methods from deep geometric learning for knowledge discovery from point cloud data.

Inaugural Lecture Andreas Holzinger Human-Centered AI

/in , /by

The inaugural lecture of Andreas Holzinger on Monday, Nov, 7, 2022, 18:00 on Human-Centered AI is open to the public – you are cordially welcome

Read more

Open Postdoc Position “Artificial intelligence for smart forest operations”

/in /by

We continue to build up our HCAI-Lab in an absolutely cool environment with exciting Artificial Intelligence topics.

Cyber-physical systems, robotics, sensor technology, data management in general, and methods of artificial intelligence (Al) and machine learning (ML) with applications to smart farm and forest operations are of increasing interest.

We seek an postdoctoral research associate in AI/machinelearning for Forest Operations, Reference Code 184

Please note the following required qualifications:

  • Doctorate degree / PhD in Computer Science/Informatics or equivalent
  • Language skills: German and English
  • Evidence of a very active publication record is required
  • Extensive teaching experience in Al/machine learning and data science
  • Experience in developing deep learning models to solve complex problems

please apply here:

https://euraxess.ec.europa.eu/jobs/838860

Note: We regret that we cannot reimburse applicants travel and lodging expenses incurred as part of the selection and hiring process.

We constantly seek to increase the number of female faculty members. Therefore qualified women are strongly encouraged to apply. In case of equal qualification, female candidates will be given preference unless reasons specific to an individual male candidate tilt the balance in his favour. People with disabilities and appropriate qualifications are specifically encouraged to apply.

Cross Domain Machine Learning and Knowledge Extraction Conference

/in /by

Great Cross Domain Machine Learning and Knowledge Extraction Conference in Vienna

Read more