Health Data Science today needs Machine Learning of Tomorrow.

SUCCESSFUL MACHINE LEARNING …

needs a concerted international effort without boundaries, supporting collaborative and integrative research between experts from seven fields: ❶ data science, ❷ algorithms, ❸ network science, ❹ graphs/topology, ❺ time/entropy, ❻ visualization, and ❼ privacy, data protection, safety and security.

The goal of our scientific community is to design, develop, test and evaluate algorithms which can learn from data and improve with experience over time and can be used for making predictions. Machine learning (ML) is the most growing field in computer science and health is amongst the greatest application challenges. The grand vision is in automatic machine learning algorithms and recent advances e.g. in speech recognition, recommender systems, or autonomous vehicles show impressive results. Automatic approaches greatly benefit from big data with many training sets. However, the application of such automatic machine learning (aML) approaches in the complex health domain, where we are confronted with complex, heterogeneous, high-dimensional, probabilistic, uncertain, incomplete, noisy data and sometimes with small data, or rare events, seems elusive in the near future. A good example are Gaussian processes, where aML (e.g. standard kernel machines) struggle on function extrapolation problems which are trivial for human learners. Here interactive machine learning (iML) may be of help, which can be de fined as “algorithms that can interact with agents and can optimize their learning behaviour through these interactions, where the agents can also be human”. This human-in-the-loop can be bene ficial in solving computationally hard problems, particularly with a doctor-in-the-loop, e.g. in subspace clustering, protein folding, or k-Anonymization, where human expertise can help to reduce an exponential search space through heuristic selection of samples. Therefore, what would otherwise be an NP-hard problem, reduces greatly in complexity through an human agent involved in the ML pipeline.

integrative interactive machine learning human-in-the-loop

The knowledge discovery pipeline needs a concerted cross-disciplinary effort of diverse experts

Consequently, a synergistic combination of methodologies and approaches of two areas offer ideal conditions towards unraveling these challenges and to foster new, efficient and user-friendly algorithms and tools: Human-Computer Interaction (HCI) and Knowledge Discovery/Data Mining (KDD), with the goal of supporting human intelligence with computational intelligence.

The cross-domain integration and appraisal of different fields shall provide an atmosphere to foster different perspectives and opinions; it will offer a platform for novel ideas and a fresh look on the methodologies to support crazy ideas – and at the end of the day to put these ideas into Business. The initial idea can be read [here],<springerlink>.

The mission of the international expert network HCI-KDD is to bring together diverse researchers sharing a common vision and to stimulate crazy ideas and a fresh look to methodologies from other disciplines without any boundaries, encouraging multi-disciplinary work, to bundle synergies, to participate in joint project proposals for getting funding on various levels, inclusive travel funds, international student exchange and promoting young and early-stage researchers.

The expert network HCI-KDD organizes special sessions at least once a year, see e.g. [1st – Graz], [2nd – Macau], [3rd – Maribor], [4th – Regensburg], [5th Lisbon], [6th Warszawa], [7th Banff], [8th London], [9th Salzburg], and two upcoming in [10a Reggio di Calabria][10b Reggio di Calabria]
The review templates for these events can be found here:
[REVIEW_TEMPLATE_XXXX] (word doc, 140kB)
[REVIEW_TEMPLATE_XXXX] (pdf, 102 kB)

Some recent example outputs of our concerted effort can be seen here:

Springer Lecture Notes in Artificial Intelligence LNAI 9605

Springer Lecture Notes in Computer Science LNCS 8700

Springer Lecture Notes in Computer Science LNCS 8401

Springer Lecture Notes in Computer Science LNCS 7947

Springer Lecture Notes in Computer Science LNCS 7058

Springer Lecture Notes in Computer Science LNCS 6389

concerted effort of the HCI-KDD international expert network

Integrative Machine Learning needs a concerted effort

International Scientific Committee:

MED = Medical Doctor (“doctor-in-the-loop”); IND = Industry Member; ESR = Early Stage Researcher, e.g. PhD-Student)