Important Dates

What is MART?


MART (Micro-activity Retrieval Task)
is an NTCIR-15 workshop data challenge task. The NTCIR15-MART pilot task aims to motivate the development of a first generation of techniques for high-precision micro-activity detection and retrieval of micro-activities of daily living, to support identification and retrieval of activities that occur over short time-scales, such as minutes, rather than the long-duration event segmentation tasks of the past work. Participating researchers will develop and benchmark approaches to retrieve micro-activities from rich time-aligned multi-modal sensor data. The provided data will be gathered by equipping the individual to gather a detailed lifelog of activities as an individual progresses through a protocol of real-world activities (e.g. using a computer, solving a problem, eating, daydreaming, etc). The chosen sensors will capture a lifelog camera data stream, biosignal activity (EOG, HR), and computer accesses to record interaction with digital artefacts. This task will present a new challenging set of micro-activity-based topics. Groups will be ranked in decreasing order of micro-activity retrieval accuracy (via mAP). Prior to the NTCIR-15 December conference participating organisations will be expected to submit a paper outlining their approach which they will then present at the MART workshop in Tokyo, Japan this coming December 2020.

Important: Although this is a collaborative evaluation where the performance of participating organisation's strategies will be ranked in terms of mAP (mean average precision), it is expected that many good approaches that may perform sub optimally to others in terms of mAP alone may offer other advantages in terms of speed, insight generation, interpretability and/or cross-subject applicability. Papers exploring such aspects of the data and tasks set are highly encouraged as part of the Insights task.

There will be two sub-tasks in this pilot phase of the MART task:

  • A) Retrieval task: a (novel) retrieval task that can be undertaken in an interactive or automatic manner and
  • B) Insights task: an insights task where participants (using the dataset) can explore interesting use cases of the dataset.
Further information on the dataset: DATA.


How to Participate?
In order to gain access to the dataset and participate in the MART task you should consult the NTCIR-15 page detailing registration information for task participation:
http://research.nii.ac.jp/ntcir/ntcir-15/howto.html
After registering for the task please contact graham.healy@dcu.ie to gain access to the dataset and baseline system.


Who is organising this?
Graham Healy (* Lead Contact / Lead Organiser) Dublin City University, Ireland

Cathal Gurrin Dublin City University, Ireland